Post links to your external data sources here before the deadline specified in the rules. Once it has been posted, you do not need to post it again.
", "messages": [ { "id": 758820, "postDate": "2020-02-28T07:35:42.460Z", "content": "\n\nAs external data might be very helpful to obtain better scores on public/private leaderboards how are you going to validate solutions' compliance with the rules?\nThere is a possibility to cheat by using non allowed external data (or not posted here before merge deadline). I'm strongly against that but we are not in a perfect world and cheaters will always exist. \nIdeally winning solutions must be reproduced end-to-end: retrain, test - check that scores are quite similar to public/private leaderboards. This will ensure that winners used only allowed datasets.
", "rawMarkdown": "@addisonhoward @juliaelliott\n\nAs external data might be very helpful to obtain better scores on public/private leaderboards how are you going to validate solutions' compliance with the rules?\nThere is a possibility to cheat by using non allowed external data (or not posted here before merge deadline). I'm strongly against that but we are not in a perfect world and cheaters will always exist. \nIdeally winning solutions must be reproduced end-to-end: retrain, test - check that scores are quite similar to public/private leaderboards. This will ensure that winners used only allowed datasets.", "votes": 25, "replies": [ { "id": 758877, "postDate": "2020-02-28T09:27:16.747Z", "content": "Also it would be nice to have a pinned post from organizers summarizing the approved datasets from all the comments here. Like \"Ok guys, the merge deadline is a thing now, here are the datasets that we approve: <...>. All the others, including <...> are not compatible with the license rule after the review.
\n\nAs for now some dataset links posted here can be just misleading and not all of them obviously got organizers attention to reply something like \"notice that this one will not be allowed\".
", "rawMarkdown": "Also it would be nice to have a pinned post from organizers summarizing the approved datasets from all the comments here. Like \"Ok guys, the merge deadline is a thing now, here are the datasets that we approve: <...>. All the others, including <...> are not compatible with the license rule after the review.\n\nAs for now some dataset links posted here can be just misleading and not all of them obviously got organizers attention to reply something like \"notice that this one will not be allowed\".", "votes": 17 }, { "id": 758910, "postDate": "2020-02-28T10:25:07.640Z", "content": "I agree with both. It is still unclear which external data is allowed and what isn't, since it seems like many of the declared datasets are under non-commercial licences and thus not allowed.
\n\nI would also like to know that are we allowed to use our self-collected videos from Youtube, if we post links to these videos here?
", "rawMarkdown": "I agree with both. It is still unclear which external data is allowed and what isn't, since it seems like many of the declared datasets are under non-commercial licences and thus not allowed.\n\nI would also like to know that are we allowed to use our self-collected videos from Youtube, if we post links to these videos here?", "votes": 9 }, { "id": 759060, "postDate": "2020-02-28T14:19:53.500Z", "content": "\"Ideally winning solutions must be reproduced end-to-end: retrain, test - check that scores are quite similar to public/private leaderboards.\"
\n\nAlthough I agree with the concern and repeatibility is really important; I am not sure if repeatibility is always guaranteed for all training frameworks(maybe it is; i am an ML noob so i might be mistaken).\nIf at the end the scores are really close; this validation step can create a lot of confusion. What if one gets 3% worse results during retrain/test and the difference between 1st and 6th is less than 3%? does that mean the top one cheated? Also there is ensembling; so things will get really confusing. You have to reproduce all the single models in similar quality and the ensembled model should also produce the same good results.
\nIt is like \"if you win you have to prove you are not guilty\".
I agree the concern should be addressed but not sure retraining / reproducing the test results in a satisfactory way is always easily doable.
", "rawMarkdown": "*\"Ideally winning solutions must be reproduced end-to-end: retrain, test - check that scores are quite similar to public/private leaderboards.\"*\n\nAlthough I agree with the concern and repeatibility is really important; I am not sure if repeatibility is always guaranteed for all training frameworks(maybe it is; i am an ML noob so i might be mistaken).\nIf at the end the scores are really close; this validation step can create a lot of confusion. What if one gets 3% worse results during retrain/test and the difference between 1st and 6th is less than 3%? does that mean the top one cheated? Also there is ensembling; so things will get really confusing. You have to reproduce all the single models in similar quality and the ensembled model should also produce the same good results. \nIt is like \"if you win you have to prove you are not guilty\". \n\nI agree the concern should be addressed but not sure retraining / reproducing the test results in a satisfactory way is always easily doable." }, { "id": 759744, "postDate": "2020-02-29T12:12:03.100Z", "content": "+1 I still don't understand if we can use or not these datasets hosted on Kaggle:\nhttps://www.kaggle.com/tunguz/70000-real-faces-1\nhttps://www.kaggle.com/tunguz/70000-real-faces-2\nhttps://www.kaggle.com/tunguz/70000-real-faces-3\nhttps://www.kaggle.com/tunguz/70000-real-faces-4\nhttps://www.kaggle.com/tunguz/70000-real-faces-5\nhttps://www.kaggle.com/tunguz/70000-real-faces-6
\n\n@juliaelliott Data is hosted on Kaggle, but it does not mean we can use it, correct? We have to find out underlying license but is someone at Kaggle will review and give a green light? It requires a bit of legal skills to be sure.
", "rawMarkdown": "+1 I still don't understand if we can use or not these datasets hosted on Kaggle:\nhttps://www.kaggle.com/tunguz/70000-real-faces-1\nhttps://www.kaggle.com/tunguz/70000-real-faces-2\nhttps://www.kaggle.com/tunguz/70000-real-faces-3\nhttps://www.kaggle.com/tunguz/70000-real-faces-4\nhttps://www.kaggle.com/tunguz/70000-real-faces-5\nhttps://www.kaggle.com/tunguz/70000-real-faces-6\n\n@juliaelliott Data is hosted on Kaggle, but it does not mean we can use it, correct? We have to find out underlying license but is someone at Kaggle will review and give a green light? It requires a bit of legal skills to be sure.", "votes": 3 }, { "id": 761672, "postDate": "2020-03-02T20:45:02.410Z", "content": "If your model makes use of any external data that is prohibited by the rules, then you are subject to disqualification by the host upon review of your solution, in particular if you are a prospective winner. As stated many times previously, if there are any restrictions imposed on the dataset's use (including non-commercial use only or restriction on those who have access to the dataset), that is considered in violation of the requirement that the data be \"available to use by all participants of the competition\" and therefore prohibited. This should be quite clear at this point.
", "rawMarkdown": "If your model makes use of any external data that is prohibited by the rules, then you are subject to disqualification by the host upon review of your solution, in particular if you are a prospective winner. As stated many times previously, if there are any restrictions imposed on the dataset's use (including non-commercial use only or restriction on those who have access to the dataset), that is considered in violation of the requirement that the data be \"available to use by all participants of the competition\" and therefore prohibited. This should be quite clear at this point." }, { "id": 761681, "postDate": "2020-03-02T20:57:26.280Z", "content": "@juliaelliott That makes sense. Thank you!
\n\nSo basically we may use data that:\n- is available for commercial usage - this was not clear from the rules\n- and adheres to competition rules
\n\nWhat about datasets collected from youtube videos? Or just youtube videos? \nI think there are a lot of questions regarding them as they actually represent \"organic videos\".
", "rawMarkdown": "@juliaelliott That makes sense. Thank you!\n\nSo basically we may use data that:\n- is available for commercial usage - this was not clear from the rules\n- and adheres to competition rules \n\nWhat about datasets collected from youtube videos? Or just youtube videos? \nI think there are a lot of questions regarding them as they actually represent \"organic videos\". ", "votes": 8 }, { "id": 761735, "postDate": "2020-03-02T22:10:07.373Z", "content": "I think it makes sense, as this is not a commercial use. This is a competition with the goal of improving detection of deep fakes detection, which has prizes for the best models.\nSo if I understood the terms correctly as long as the data is available to all participants at no cost, it is disclosed in this thread and it's allowed to be used for this purpose, then it is ok.
\n\nFor most of the datasets you need to apply to have access to it and mentioned what you are going to use it for. So if you mention the kaggle competition, and they send you the link, then they are allowing you to use it for this competition.
\n\n@juliaelliott Can you just confirm or not, my statements above? This would help answering a lot of the questions present in this thread.
", "rawMarkdown": "I think it makes sense, as this is not a commercial use. This is a competition with the goal of improving detection of deep fakes detection, which has prizes for the best models.\nSo if I understood the terms correctly as long as the data is available to all participants at no cost, it is disclosed in this thread and it's allowed to be used for this purpose, then it is ok.\n\nFor most of the datasets you need to apply to have access to it and mentioned what you are going to use it for. So if you mention the kaggle competition, and they send you the link, then they are allowing you to use it for this competition.\n\n@juliaelliott Can you just confirm or not, my statements above? This would help answering a lot of the questions present in this thread." }, { "id": 761741, "postDate": "2020-03-02T22:17:48.320Z", "content": "Seems like you got it wrong. What @juliaelliott says is that we cannot use datasets if they are for research only.
", "rawMarkdown": "Seems like you got it wrong. What @juliaelliott says is that we **cannot** use datasets if they are for research only. " }, { "id": 761752, "postDate": "2020-03-02T22:29:29.977Z", "content": "@selimsef does that make imagenet pretrained models(including model zoo of many libraries) + 90% of what is posted on this topic unusable?
", "rawMarkdown": "@selimsef does that make imagenet pretrained models(including model zoo of many libraries) + 90% of what is posted on this topic unusable?", "votes": -1 }, { "id": 761756, "postDate": "2020-03-02T22:39:01.407Z", "content": "AFAIK we cannot use ImageNet dataset directly, but we can use public pretrained weights with open license.
", "rawMarkdown": "AFAIK we cannot use ImageNet dataset directly, but we can use public pretrained weights with open license." }, { "id": 761757, "postDate": "2020-03-02T22:40:50.010Z", "content": "@emrebayram Yeah, if Selim is right that would be the case. Would be great if the organizers (@juliaelliott @addisonhoward) would give a final answer regarding this, cause at the moment I'm very confused regarding what we can, and what we cannot use. And I'm betting I'm not the only one
", "rawMarkdown": "@emrebayram Yeah, if Selim is right that would be the case. Would be great if the organizers (@juliaelliott @addisonhoward) would give a final answer regarding this, cause at the moment I'm very confused regarding what we can, and what we cannot use. And I'm betting I'm not the only one ", "votes": 1 } ] }, { "id": 692881, "postDate": "2019-12-11T20:15:32.410Z", "content": "Post links to your external data sources here before the deadline specified in the rules. Once it has been posted, you do not need to post it again.
", "rawMarkdown": "Post links to your external data sources here before the deadline specified in the rules. Once it has been posted, you do not need to post it again.", "votes": 36 }, { "id": 716551, "postDate": "2020-01-11T21:25:02.777Z", "content": "There are several DeepFake datasets avaialble:\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/
\n\nAll of them avaiable by request. Can we use them for developing solution or not? I want to give it a try, but don't want to waste time on them if they are forbidden.
", "rawMarkdown": "There are several DeepFake datasets avaialble:\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/\n\nAll of them avaiable by request. Can we use them for developing solution or not? I want to give it a try, but don't want to waste time on them if they are forbidden.", "votes": 11, "replies": [ { "id": 717128, "postDate": "2020-01-12T19:22:58.937Z", "content": "DeeperForensics-1.0 also : https://github.com/EndlessSora/DeeperForensics-1.0 .
\n\n@juliaelliott @addisonhoward
\nAnd what about CC BY-NC 4.0 licence ? For example @tunguz shared 1 Million Fake Faces dataset.
\nCan we use pre-trained models and/or datasets under this (or similar for non-commercial research purposes only) license and expect to receive prize money?
I'm new to kaggle, so sorry for this possibly dummy question.
", "rawMarkdown": "DeeperForensics-1.0 also : https://github.com/EndlessSora/DeeperForensics-1.0 .\n\n@juliaelliott @addisonhoward \nAnd what about [CC BY-NC 4.0 licence](https://creativecommons.org/licenses/by-nc/4.0/) ? For example @tunguz shared [1 Million Fake Faces](https://www.kaggle.com/tunguz/1-million-fake-faces) dataset. \nCan we use **pre-trained models** and/or **datasets** under this (or similar for ***non-commercial research purposes only***) license and expect to receive prize money?\n\nI'm new to kaggle, so sorry for this possibly dummy question.", "votes": 10, "isDeleted": true }, { "id": 718257, "postDate": "2020-01-14T08:17:22Z", "content": "Is there any way to get DeeperForensics-1.0? As I can see there only \"Coming Soon\" and nothing more.
", "rawMarkdown": "Is there any way to get DeeperForensics-1.0? As I can see there only \"Coming Soon\" and nothing more.", "votes": 3 }, { "id": 721524, "postDate": "2020-01-17T13:03:25.260Z", "content": "So are we allowed to use them or?
", "rawMarkdown": "So are we allowed to use them or?" }, { "id": 721699, "postDate": "2020-01-17T16:12:50.043Z", "content": "@zfturbo I've contacted them to get the publish date at least. However, it is not defined yet.
\n\nTheir response:\nCurrently, I cannot decide when to release the dataset due to some approval process. But I believe it will be soon. Once it is ready, we will make a notification on our GitHub page.
", "rawMarkdown": "@zfturbo I've contacted them to get the publish date at least. However, it is not defined yet.\n\nTheir response:\nCurrently, I cannot decide when to release the dataset due to some approval process. But I believe it will be soon. Once it is ready, we will make a notification on our GitHub page.", "votes": 1 }, { "id": 721707, "postDate": "2020-01-17T16:21:57.233Z", "content": "I think that we are not allowed to use them. But organizers have not answered yet...
\n\nAs for me, I haven't used any external datasets/models yet except ImageNet pretrained models in some experiments. I think that it can be a waste of time.
", "rawMarkdown": "I think that we are not allowed to use them. But organizers have not answered yet...\n\nAs for me, I haven't used any external datasets/models yet except ImageNet pretrained models in some experiments. I think that it can be a waste of time. ", "isDeleted": true }, { "id": 721812, "postDate": "2020-01-17T18:15:44.640Z", "content": "@vladislavleketush and others - Your use of external data should conform to the external data specification in the competition's rules:
\n\n\n\n\nyou will (i) ensure the External Data is available to use by all participants of the competition for purposes of the competition at no cost to the other participants and (ii) post such access to the External Data for the participants to the official competition forum prior to the Entry Deadline.
\n
Therefore, licenses which place restrictions on datasets' use (whether by purpose, affiliation, cost or other restrictive means) would be in violation of this rule.
", "rawMarkdown": "@vladislavleketush and others - Your use of external data should conform to the external data specification in the [competition's rules](https://www.kaggle.com/c/deepfake-detection-challenge/rules): \n> you will (i) ensure the External Data is available to use by all participants of the competition for purposes of the competition at no cost to the other participants and (ii) post such access to the External Data for the participants to the official competition forum prior to the Entry Deadline.\n\nTherefore, licenses which place restrictions on datasets' use (whether by purpose, affiliation, cost or other restrictive means) would be in violation of this rule.", "votes": 8 }, { "id": 728872, "postDate": "2020-01-25T12:17:52.423Z", "content": "March 3, 2020 - Entry deadline. You must accept the competition rules before this date in order to compete. (is it fair sharing external data at end of competition?)\n- we treat these external store are magic features till March 3rd, 2020?
\n\nPlease clarify if i miss understood
", "rawMarkdown": "March 3, 2020 - Entry deadline. You must accept the competition rules before this date in order to compete. (is it fair sharing external data at end of competition?)\n- we treat these external store are magic features till March 3rd, 2020?\n\nPlease clarify if i miss understood", "votes": 1 }, { "id": 730874, "postDate": "2020-01-28T03:36:32.123Z", "content": "@seshurajup - yes, the external data declaration deadline on this forum thread is the entry deadline, March 3, 2020.
", "rawMarkdown": "@seshurajup - yes, the external data declaration deadline on this forum thread is the entry deadline, March 3, 2020.", "votes": 1 } ] }, { "id": 694492, "postDate": "2019-12-13T17:40:14.113Z", "content": "For clarity on the datasets that need to be declared here: You do not need to declare your own self-trained/original work models. But if you are using a dataset or pre-trained model obtained externally which is licensed with the right for you to use, then that dataset needs to be declared.
\n\nSo, for example: if you are using the imagenet dataset and efficientnet to train your model(s), you must declare imagenet and efficientnet on this thread, but you don’t need to make your trained model public.
\n\nYou also do not need to declare things that have already been declared.
", "rawMarkdown": "For clarity on the datasets that need to be declared here: You do not need to declare your own self-trained/original work models. But if you are using a dataset or pre-trained model obtained externally which is licensed with the right for you to use, then that dataset needs to be declared. \n\nSo, for example: if you are using the imagenet dataset and efficientnet to train your model(s), you must declare imagenet and efficientnet on this thread, but you don’t need to make your trained model public. \n\nYou also do not need to declare things that have already been declared.", "votes": 9, "replies": [ { "id": 702696, "postDate": "2019-12-25T02:58:03.390Z", "content": "Just for clarity, would that mean if I wanted to use one of the Keras pre-trained models as part of this competition I would need to declare it here?
\n\nSorry if that is a noob question.
", "rawMarkdown": "Just for clarity, would that mean if I wanted to use one of the Keras pre-trained models as part of this competition I would need to declare it here? \n\nSorry if that is a noob question." }, { "id": 712382, "postDate": "2020-01-07T07:02:49.667Z", "content": "Sorry I'm still a bit confused.
\n\n1) you mentioned use imagenet as external data source. But isn't imagenet larger than 1Gb?\n2) can i create my own deepfake dataset and use it to train my model? do i have to declare the dataset\n3) can i create my own deepfake dataset and use it to PRE-train my model? do i have to just decalre the pretrained model or the original dataset? does the 1 Gb rule apply to the whole dataset or just the pretrained model?
", "rawMarkdown": "Sorry I'm still a bit confused.\n\n1) you mentioned use imagenet as external data source. But isn't imagenet larger than 1Gb?\n2) can i create my own deepfake dataset and use it to train my model? do i have to declare the dataset\n3) can i create my own deepfake dataset and use it to PRE-train my model? do i have to just decalre the pretrained model or the original dataset? does the 1 Gb rule apply to the whole dataset or just the pretrained model?" }, { "id": 713022, "postDate": "2020-01-07T20:21:49.673Z", "content": "@davidbonn Yes, the pre-trained model should be declared.
", "rawMarkdown": "@davidbonn Yes, the pre-trained model should be declared." }, { "id": 713041, "postDate": "2020-01-07T20:38:57.670Z", "content": "@shitaili \n1) The 1GB external data requirement is a systematic constraint on any notebook you submit. So, you can use large external datasets (like imagenet) that exceed 1GB if you train offline, and then upload that trained model into your submission notebook. The total of external datasets uploaded to the submission notebook you commit in Kaggle must not exceed 1GB.\n2) Yes, you can create your own deepfake dataset if you have the right to use and share the videos/dataset that you are using. You must declare the source datasets that are used.\n3) The original dataset. See #1 about the 1GB constraint.
", "rawMarkdown": "@shitaili \n1) The 1GB external data requirement is a systematic constraint on any notebook you submit. So, you can use large external datasets (like imagenet) that exceed 1GB if you train offline, and then upload that trained model into your submission notebook. The total of external datasets uploaded to the submission notebook you commit in Kaggle must not exceed 1GB.\n2) Yes, you can create your own deepfake dataset if you have the right to use and share the videos/dataset that you are using. You must declare the source datasets that are used.\n3) The original dataset. See #1 about the 1GB constraint." }, { "id": 725200, "postDate": "2020-01-21T22:36:26.943Z", "rawMarkdown": "", "isDeleted": true }, { "id": 725241, "postDate": "2020-01-21T23:34:13.180Z", "content": "@mindcool The notebook in Kaggle that you are submitted is not included. But any trained models or other external data that is being used in your submission notebook is included in the 1GB.
", "rawMarkdown": "@mindcool The notebook in Kaggle that you are submitted is not included. But any trained models or other external data that is being used in your submission notebook is included in the 1GB." }, { "id": 753674, "postDate": "2020-02-22T14:35:17.847Z", "content": "Imagenet is non-commercial usage only. Below in this thread, you texted:
\n\n\n\n\n\"I’ve answered the question about BY-NC not being available for use by all (non-commercial use) and therefore violating the requirement that external data be available for use by all participants.\"
\n
and
\n\n\n\n\n\"You raise a fair question. The rule indicates to use by all participants -- Inherent in data licensing terms that constrain the data's use is the restriction that it is not available to be used by all participants. In many cases with research/academic-use licensing, only those in a research profession may gain access to the dataset. Sure, in some cases, everyone may be able to physically download it, but if a license prohibits its use for non-research uses, then it can be interpreted as also prohibiting use by users who are non-researching. So the issue here is one of legally-compliant accessibility to use of the dataset by all.\"
\n
Does it mean, that we can not use the imagenet as well?
", "rawMarkdown": "Imagenet is non-commercial usage only. Below in this thread, you texted:\n> \"I’ve answered the question about BY-NC not being available for use by all (non-commercial use) and therefore violating the requirement that external data be available for use by all participants.\"\n\nand\n> \"You raise a fair question. The rule indicates to use by all participants -- Inherent in data licensing terms that constrain the data's use is the restriction that it is not available to be used by all participants. In many cases with research/academic-use licensing, only those in a research profession may gain access to the dataset. Sure, in some cases, everyone may be able to physically download it, but if a license prohibits its use for non-research uses, then it can be interpreted as also prohibiting use by users who are non-researching. So the issue here is one of legally-compliant accessibility to use of the dataset by all.\"\n\nDoes it mean, that we can not use the imagenet as well? ", "votes": 2 }, { "id": 753681, "postDate": "2020-02-22T14:43:43.273Z", "content": "You're asking just to ask or you've already read this ?
", "rawMarkdown": "You're asking just to ask or you've already read [this](http://image-net.org/download-faq ) ? ", "isDeleted": true }, { "id": 753701, "postDate": "2020-02-22T15:11:10.370Z", "content": "Yes, I've been read this, but I really confused about this. \nPoint 1 for agreement to download Imagenet:
\n\n\n\n\n\n
\n- Researcher shall use the Database only for non-commercial research and educational purposes.
\n
Ok, but we can download it by direct link, is it enough to this contest? How we can get proof, that the pre-trained model was not used original images. Typically, code for pre-trained models does not contain a part for download by direct links.
", "rawMarkdown": "Yes, I've been read this, but I really confused about this. \nPoint 1 for agreement to download Imagenet:\n> 1. Researcher shall use the Database only for non-commercial research and educational purposes.\n\nOk, but we can download it by direct link, is it enough to this contest? How we can get proof, that the pre-trained model was not used original images. Typically, code for pre-trained models does not contain a part for download by direct links.", "votes": 2 }, { "id": 753806, "postDate": "2020-02-22T16:58:51.543Z", "content": "\n\n\nthat we can not use the imagenet as well
\n
ImageNet is available for download on Kaggle.
", "rawMarkdown": "> that we can not use the imagenet as well\n\nImageNet is available for download on Kaggle." }, { "id": 753831, "postDate": "2020-02-22T17:41:18.667Z", "content": "@i7p9h9 I have tried to find answer to the same question. Searched internet and checked similar concerns/questions ( see this and that\n ). None of them have clear answer or no answer at all.
\n\nFirst of all I am not a lawyer.\nBut obviously; there is no clear answer. The reason why it is not clear is there is no court decided for a similar issue.\nImagenet FAQ says the database is for non-commercial; but what is the definition of the database there? Using the images as it is? Using the original images? Or any derived information from these images? Then one would claim that google street view is violating many laws(maybe this will be the case in the future) because they did not get my permission to take a photo of my apartment and use it in a commercial app.
\n\nSo it is hard for any organization to make a clear comment about this. I am sure many commercial apps as of today are using imagenet pretrained models because it is part of many library's model zoo.
\n\nIt is also hard for a judge to decide in my opinion. Because it is a complex topic.\nAs a human being, if I look at an image and learn something from it; can I use this knowledge for commercial gain(for example for teaching)? Probably yes. But probably it is not ok to copy-paste the material as it is.\nIn my opinion we are still in an early era for things like this to settle. It is like 15-20 years ago noone cared about \"personal data\"; but as of today there is GDPR and many other regulations.\nProbably for deep learning 5-15 years from now the regulations will be there about data usage but for now it is just kind of impossible to directly say \"this is ok\" or \"this is not\".
\n\nFor this competition at the moment I think every competitor need to use their own judgement about this.
", "rawMarkdown": "@i7p9h9 I have tried to find answer to the same question. Searched internet and checked similar concerns/questions ( see [this](https://discuss.pytorch.org/t/pre-trained-models-license/38647) and [that](https://discuss.pytorch.org/t/can-i-use-the-pretrained-models-for-commercial-use/54279)\n ). None of them have clear answer or no answer at all.\n\nFirst of all I am not a lawyer.\nBut obviously; there is no clear answer. The reason why it is not clear is there is no court decided for a similar issue.\nImagenet FAQ says the database is for non-commercial; but what is the definition of the database there? Using the images as it is? Using the original images? Or any derived information from these images? Then one would claim that google street view is violating many laws(maybe this will be the case in the future) because they did not get my permission to take a photo of my apartment and use it in a commercial app.\n\nSo it is hard for any organization to make a clear comment about this. I am sure many commercial apps as of today are using imagenet pretrained models because it is part of many library's model zoo.\n\nIt is also hard for a judge to decide in my opinion. Because it is a complex topic.\nAs a human being, if I look at an image and learn something from it; can I use this knowledge for commercial gain(for example for teaching)? Probably yes. But probably it is not ok to copy-paste the material as it is.\nIn my opinion we are still in an early era for things like this to settle. It is like 15-20 years ago noone cared about \"personal data\"; but as of today there is GDPR and many other regulations.\nProbably for deep learning 5-15 years from now the regulations will be there about data usage but for now it is just kind of impossible to directly say \"this is ok\" or \"this is not\".\n\nFor this competition at the moment I think every competitor need to use their own judgement about this.", "votes": 1 }, { "id": 753902, "postDate": "2020-02-22T19:55:05.773Z", "content": "@emrebayram, thank you for the detailed answer. I hope participants will not any have any problem due to using imagenet pre-trained models in this competition.
", "rawMarkdown": "@emrebayram, thank you for the detailed answer. I hope participants will not any have any problem due to using imagenet pre-trained models in this competition.", "votes": 1 } ] }, { "id": 760710, "postDate": "2020-03-01T16:33:10.880Z", "content": "Some people have surfaced this already but let me ask again: it would be nice to have the final list of allowed external resources under the post. It probably requires some effort to create this list and check it but will be very useful to everyone since the deadline is approaching. \nWhat do you think @addisonhoward?
", "rawMarkdown": "Some people have surfaced this already but let me ask again: it would be nice to have the final list of allowed external resources under the post. It probably requires some effort to create this list and check it but will be very useful to everyone since the deadline is approaching. \nWhat do you think @addisonhoward? ", "votes": 7 }, { "id": 706023, "postDate": "2019-12-29T20:24:05.257Z", "content": "@tunguz' datasets for real faces and fake faces.
\n\nReal Faces:\nhttps://www.kaggle.com/tunguz/70000-real-faces-1
\n\nFake Faces:\nhttps://www.kaggle.com/tunguz/1-million-fake-faces\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-2\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-3\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-4\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-5\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-6\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-7
", "rawMarkdown": "@tunguz' datasets for real faces and fake faces.\n\nReal Faces:\nhttps://www.kaggle.com/tunguz/70000-real-faces-1\n \nFake Faces:\nhttps://www.kaggle.com/tunguz/1-million-fake-faces\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-2\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-3\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-4\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-5\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-6\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-7", "votes": 7, "replies": [ { "id": 707297, "postDate": "2019-12-31T15:38:15.370Z", "content": "Also, The Flickr-Faces-HQ (FFHQ) Dataset of 70000 real faces:\nhttps://github.com/NVlabs/ffhq-dataset
", "rawMarkdown": "Also, The Flickr-Faces-HQ (FFHQ) Dataset of 70000 real faces:\nhttps://github.com/NVlabs/ffhq-dataset", "votes": 2 }, { "id": 745157, "postDate": "2020-02-13T15:05:51.690Z", "content": "@carlolepelaars careful the competition rules state that the models should be available for commercial use. 1-million-fake-faces dataset is under \"Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)\" license so probable is not ok. @juliaelliott am I correct?
", "rawMarkdown": "@carlolepelaars careful the competition rules state that the models should be available for commercial use. 1-million-fake-faces dataset is under \"Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)\" license so probable is not ok. @juliaelliott am I correct? ", "votes": 2 }, { "id": 745167, "postDate": "2020-02-13T15:21:32.127Z", "content": "\n\nI've asked the same question. Check this
", "rawMarkdown": "@emrebayram \n\nI've asked the same question. Check [this](https://www.kaggle.com/c/deepfake-detection-challenge/discussion/121203#721812) ", "votes": 2, "isDeleted": true }, { "id": 745173, "postDate": "2020-02-13T15:26:05.667Z", "content": "Ok, thanks for the heads-up guys! I'm not using these datasets in my current modelling so its fine.
", "rawMarkdown": "Ok, thanks for the heads-up guys! I'm not using these datasets in my current modelling so its fine.", "votes": 1 } ] }, { "id": 763235, "postDate": "2020-03-04T09:21:52.830Z", "content": "By the way, a list of permitted external data and allowed pre-trained models (using approved licenses) should be maintained as a \"Kaggle-level\" topic because we have the same questions repeated across competitions.
", "rawMarkdown": "By the way, a list of permitted external data and allowed pre-trained models (using approved licenses) should be maintained as a \"Kaggle-level\" topic because we have the same questions repeated across competitions.", "votes": 6 }, { "id": 701432, "postDate": "2019-12-23T13:37:49.860Z", "content": "```\nUADFV: The UADFV dataset - Xin Yang, Yuezun Li, and Siwei Lyu. Exposing deep fakes\nusing inconsistent head poses. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),\n2019.
\n\nDF-TIMIT: Pavel Korshunov and Sebastien Marcel. Deepfakes: a new ´\nthreat to face recognition? assessment and detection. arXiv\npreprint arXiv:1812.08685, 2018.
\n\nFF-DF: Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Chris- ¨\ntian Riess, Justus Thies, and Matthias Nießner. FaceForensics++: Learning to detect manipulated facial images. In\nICCV, 2019.
\n\nDFD: Nicholas Dufour, Andrew Gully, Per Karlsson, Alexey Victor Vorbyov, Thomas Leung, Jeremiah Childs, and Christoph Bregler. Deepfakes detection dataset by google & jigsaw.\n```
", "rawMarkdown": "```\nUADFV: The UADFV dataset - Xin Yang, Yuezun Li, and Siwei Lyu. Exposing deep fakes\nusing inconsistent head poses. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),\n2019.\n\nDF-TIMIT: Pavel Korshunov and Sebastien Marcel. Deepfakes: a new ´\nthreat to face recognition? assessment and detection. arXiv\npreprint arXiv:1812.08685, 2018.\n\nFF-DF: Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Chris- ¨\ntian Riess, Justus Thies, and Matthias Nießner. FaceForensics++: Learning to detect manipulated facial images. In\nICCV, 2019.\n\nDFD: Nicholas Dufour, Andrew Gully, Per Karlsson, Alexey Victor Vorbyov, Thomas Leung, Jeremiah Childs, and Christoph Bregler. Deepfakes detection dataset by google & jigsaw.\n```", "votes": 5, "replies": [ { "id": 737828, "postDate": "2020-02-05T20:25:30.737Z", "content": "@juliaelliott @addisonhoward That dataset is not publicly available. I've apply few weeks ago and didn't got approved. In other hands the pretrained models are free to download. Can we use them in such case?
", "rawMarkdown": "@juliaelliott @addisonhoward That dataset is not publicly available. I've apply few weeks ago and didn't got approved. In other hands the pretrained models are free to download. Can we use them in such case?" }, { "id": 737830, "postDate": "2020-02-05T20:29:37.170Z", "content": "I just apply again and got approved immediatelly. Apologies for the false alert!
", "rawMarkdown": "I just apply again and got approved immediatelly. Apologies for the false alert!" } ] }, { "id": 694433, "postDate": "2019-12-13T15:35:10.093Z", "content": "Static build of FFMPEG: https://johnvansickle.com/ffmpeg/
", "rawMarkdown": "Static build of FFMPEG: https://johnvansickle.com/ffmpeg/", "votes": 6, "replies": [ { "id": 695069, "postDate": "2019-12-14T14:13:28.150Z", "content": "@juliaelliott Are we allowed to use the Static Build of FFMPEG from a dataset like https://www.kaggle.com/rakibilly/ffmpeg-static-build ?
", "rawMarkdown": "@juliaelliott Are we allowed to use the Static Build of FFMPEG from a dataset like https://www.kaggle.com/rakibilly/ffmpeg-static-build ?", "votes": 4 }, { "id": 713043, "postDate": "2020-01-07T20:39:45.733Z", "content": "@rakibilly I don't see any problem here if it's publicly available.
", "rawMarkdown": "@rakibilly I don't see any problem here if it's publicly available.", "votes": 2 }, { "id": 725528, "postDate": "2020-01-22T08:03:46.497Z", "content": "I'm using it to extract the audio into a wav file from the original mp4 file.
", "rawMarkdown": "I'm using it to extract the audio into a wav file from the original mp4 file.", "votes": 1 } ] }, { "id": 748358, "postDate": "2020-02-17T12:35:31.360Z", "content": "Edit: Nevermind; the metadata has individual license info.
", "rawMarkdown": "Edit: Nevermind; the metadata has individual license info." }, { "id": 750285, "postDate": "2020-02-19T09:03:35.493Z", "content": "@addisonhoward @juliaelliott Could you please clarify the following ?
\n\nThe metadata itself is said to be \"CC-by-NC\" on Nvidia website, so using the metadata to select \"public domain\" pictures are not allowed ?
", "rawMarkdown": "@addisonhoward @juliaelliott Could you please clarify the following ?\n\nThe metadata itself is said to be \"CC-by-NC\" on Nvidia website, so using the metadata to select \"public domain\" pictures are not allowed ? ", "votes": 2 } ] }, { "id": 767591, "postDate": "2020-03-09T21:05:36.780Z", "content": "@juliaelliott In the rules, it says:
\n\n\n\n\nNo using other Kaggle notebooks or utility scripts as inputs to your submission notebook. External data also cannot be externally-referenced (i.e. via BigQuery, Github, external URL). Instead, load external models or datasets directly as an external data source.
\n
I upload my own scripts as a Kaggle dataset so I can import functions from them and not have an excessively long notebook. Is that allowed?
", "rawMarkdown": "@juliaelliott In the rules, it says:\n\n> No using other Kaggle notebooks or utility scripts as inputs to your submission notebook. External data also cannot be externally-referenced (i.e. via BigQuery, Github, external URL). Instead, load external models or datasets directly as an external data source.\n\nI upload my own scripts as a Kaggle dataset so I can import functions from them and not have an excessively long notebook. Is that allowed? ", "votes": 4, "replies": [ { "id": 779756, "postDate": "2020-03-19T16:57:22.937Z", "content": "Yes. This stipulation in the Code Requirements is a systematic one - in that you can't daisy-chain kernels/notebooks into your submission notebook. But you can definitely download those notebooks and upload them as inputs into your submission. Hence the Instead, load external models or datasets directly as an external data source.
InceptionV3 imagenet weights from Keras:\nhttps://keras.io/applications/#inceptionv3
\n\nUFC101 Khurram Soomro, Amir Roshan Zamir and Mubarak Shah, UCF101: A Dataset of 101 Human Action Classes From Videos in The Wild., CRCV-TR-12-01, November, 2012.
\n\nhttp://crcv.ucf.edu/data/UCF101/
", "rawMarkdown": "InceptionV3 imagenet weights from Keras:\nhttps://keras.io/applications/#inceptionv3\n\nUFC101 Khurram Soomro, Amir Roshan Zamir and Mubarak Shah, UCF101: A Dataset of 101 Human Action Classes From Videos in The Wild., CRCV-TR-12-01, November, 2012.\n\nhttp://crcv.ucf.edu/data/UCF101/", "votes": 3, "replies": [ { "id": 727959, "postDate": "2020-01-24T09:04:47.367Z", "content": "I'm going to use this too.
", "rawMarkdown": "I'm going to use this too.", "votes": 1 } ] }, { "id": 716317, "postDate": "2020-01-11T14:43:08.087Z", "content": "facenet-pytorch pre-trained models: https://github.com/timesler/facenet-pytorch\n(MIT license)
", "rawMarkdown": "facenet-pytorch pre-trained models: https://github.com/timesler/facenet-pytorch\n(MIT license)", "votes": 3 }, { "id": 693664, "postDate": "2019-12-12T16:49:41.610Z", "content": "Open Images Dataset: https://storage.googleapis.com/openimages/web/index.html\nObjects 365 Dataset: https://www.objects365.org/overview.html\nCOCO Dataset: http://cocodataset.org/#home
", "rawMarkdown": "Open Images Dataset: https://storage.googleapis.com/openimages/web/index.html\nObjects 365 Dataset: https://www.objects365.org/overview.html\nCOCO Dataset: http://cocodataset.org/#home", "votes": 3 }, { "id": 692907, "postDate": "2019-12-11T20:53:11.190Z", "content": "“External data is allowed up to 1 GB in size. External data must be freely & publicly available, including pre-trained models”
\n\nDoes it count during training time? Or just for predict time?
", "rawMarkdown": "“External data is allowed up to 1 GB in size. External data must be freely & publicly available, including pre-trained models”\n\nDoes it count during training time? Or just for predict time?", "votes": 3, "replies": [ { "id": 692929, "postDate": "2019-12-11T21:16:43.323Z", "content": "Good question! The 1 GB restriction is only on your submission notebook in Kaggle, so since you should be training elsewhere, this only applies to your inference/predictions notebook.
", "rawMarkdown": "Good question! The 1 GB restriction is only on your submission notebook in Kaggle, so since you should be training elsewhere, this only applies to your inference/predictions notebook.", "votes": 8 }, { "id": 693976, "postDate": "2019-12-13T02:19:09.333Z", "content": "To make sure, are we allowed to use private datasets (and larger than 1 GB) during the training phase? Or is the training limited only on the provided dataset?
", "rawMarkdown": "To make sure, are we allowed to use private datasets (and larger than 1 GB) during the training phase? Or is the training limited only on the provided dataset? " }, { "id": 701058, "postDate": "2019-12-23T03:07:14.303Z", "content": "the 1GB limit is only on the data used for final inference. Training can be done on any dataset as long as you mention it here (and have the right to use/mention it).
", "rawMarkdown": "the 1GB limit is only on the data used for final inference. Training can be done on any dataset as long as you mention it here (and have the right to use/mention it).", "votes": 1 }, { "id": 706022, "postDate": "2019-12-29T20:22:53.757Z", "content": "Awesome!
", "rawMarkdown": "Awesome! ", "votes": 1 } ] }, { "id": 762894, "postDate": "2020-03-03T23:15:21.877Z", "content": "https://www.youtube.com/ Creative Commons videos\nhttps://drive.google.com/file/d/1NeyTFAwdJSjxA9ZtdviwdUjdptEVjM_i\nhttps://www.kaggle.com/tunguz/1-million-fake-faces\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-2\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-3\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-4\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-5\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-6\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-7\nhttps://github.com/IDRnD/LCC_FASD
", "rawMarkdown": "https://www.youtube.com/ Creative Commons videos\nhttps://drive.google.com/file/d/1NeyTFAwdJSjxA9ZtdviwdUjdptEVjM_i\nhttps://www.kaggle.com/tunguz/1-million-fake-faces\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-2\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-3\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-4\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-5\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-6\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-7\nhttps://github.com/IDRnD/LCC_FASD", "votes": 2 }, { "id": 762923, "postDate": "2020-03-03T23:58:28.157Z", "content": "FACENET, VGG16 https://github.com/davidsandberg/facenet\nMTCNN\nhttps://github.com/ipazc/mtcnn
", "rawMarkdown": "FACENET, VGG16 https://github.com/davidsandberg/facenet\nMTCNN\nhttps://github.com/ipazc/mtcnn", "votes": 1 }, { "id": 762654, "postDate": "2020-03-03T17:28:41.200Z", "content": "https://github.com/huawei-noah/ghostnet\nhttps://github.com/iamhankai/ghostnet.pytorch\nhttps://github.com/kuan-wang/pytorch-mobilenet-v3
", "rawMarkdown": "https://github.com/huawei-noah/ghostnet\nhttps://github.com/iamhankai/ghostnet.pytorch\nhttps://github.com/kuan-wang/pytorch-mobilenet-v3", "votes": 1 }, { "id": 762339, "postDate": "2020-03-03T12:52:13.157Z", "content": "https://www.thispersondoesnotexist.com/
", "rawMarkdown": "https://www.thispersondoesnotexist.com/", "votes": 1 }, { "id": 762337, "postDate": "2020-03-03T12:49:36.790Z", "content": "All media with permitting CC license from:\nhttps://badoo.com/\nhttps://www.instagram.com/\nhttps://www.youtube.com/\nhttps://www.flickr.com/\nhttps://www.google.com/imghp\nhttps://yandex.ru/images/\nFrom this thread: https://www.kaggle.com/c/sp-society-camera-model-identification/discussion/47235
\n\nIf any source violates the rules, I will not use it. List it here just in case :)
", "rawMarkdown": "All media with permitting CC license from:\nhttps://badoo.com/\nhttps://www.instagram.com/\nhttps://www.youtube.com/\nhttps://www.flickr.com/\nhttps://www.google.com/imghp\nhttps://yandex.ru/images/\nFrom this thread: https://www.kaggle.com/c/sp-society-camera-model-identification/discussion/47235\n\nIf any source violates the rules, I will not use it. List it here just in case :)", "votes": 1 }, { "id": 762257, "postDate": "2020-03-03T11:26:55.257Z", "content": "pytorch vgg16 model\nwheight https://download.pytorch.org/models/vgg16-397923af.pth\npytorch vgg19 model\nvgg19: https://download.pytorch.org/models/vgg19-dcbb9e9d.pth\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\npretrained model: https://drive.google.com/drive/folders/1zvl89AgFAApbH0At-gMuZSeQB_LpNP-M\nhttps://github.com/timesler/facenet-pytorch\nmodel 20180408-102900: https://drive.google.com/uc?export=download&id=12DYdlLesBl3Kk51EtJsyPS8qA7fErWDX\nmodel 20180402-114759: https://drive.google.com/uc?export=download&id=1TDZVEBudGaEd5POR5X4ZsMvdsh1h68T1\nhttps://github.com/mit-han-lab/temporal-shift-module\nTSN ResNet50: https://hanlab.mit.edu/projects/tsm/models/TSM_kinetics_RGB_resnet50_avg_segment5_e50.pth\nTSM ResNet50: https://hanlab.mit.edu/projects/tsm/models/TSM_kinetics_RGB_resnet50_shift8_blockres_avg_segment8_e100_dense.pth\nhttps://github.com/ufoym/imbalanced-dataset-sampler\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch/releases/\nhttps://github.com/NVlabs/ffhq-dataset\nhttps://github.com/deepinsight/insightface/\nhttps://github.com/deepinsight/insightface/#pretrained-models\nhttps://github.com/Sierkinhane/mtcnn-pytorch\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nweight: https://drive.google.com/file/d/1WeXlNYsM6dMP3xQQELI-4gxhwKUQxc3-/view?usp=sharing\nhttps://github.com/lijiannuist/lightDSFD\nweight: https://github.com/lijiannuist/lightDSFD/raw/master/weights/light_DSFD.pth\nhttps://github.com/cydonia999/VGGFace2-pytorch\nhttps://github.com/cydonia999/VGGFace2-pytorch#pretrained-models\nhttps://github.com/albumentations-team/albumentations
", "rawMarkdown": "pytorch vgg16 model\nwheight https://download.pytorch.org/models/vgg16-397923af.pth\npytorch vgg19 model\nvgg19: https://download.pytorch.org/models/vgg19-dcbb9e9d.pth\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\npretrained model: https://drive.google.com/drive/folders/1zvl89AgFAApbH0At-gMuZSeQB_LpNP-M\nhttps://github.com/timesler/facenet-pytorch\nmodel 20180408-102900: https://drive.google.com/uc?export=download&id=12DYdlLesBl3Kk51EtJsyPS8qA7fErWDX\nmodel 20180402-114759: https://drive.google.com/uc?export=download&id=1TDZVEBudGaEd5POR5X4ZsMvdsh1h68T1\nhttps://github.com/mit-han-lab/temporal-shift-module\nTSN ResNet50: https://hanlab.mit.edu/projects/tsm/models/TSM_kinetics_RGB_resnet50_avg_segment5_e50.pth\nTSM ResNet50: https://hanlab.mit.edu/projects/tsm/models/TSM_kinetics_RGB_resnet50_shift8_blockres_avg_segment8_e100_dense.pth\nhttps://github.com/ufoym/imbalanced-dataset-sampler\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch/releases/\nhttps://github.com/NVlabs/ffhq-dataset\nhttps://github.com/deepinsight/insightface/\nhttps://github.com/deepinsight/insightface/#pretrained-models\nhttps://github.com/Sierkinhane/mtcnn-pytorch\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nweight: https://drive.google.com/file/d/1WeXlNYsM6dMP3xQQELI-4gxhwKUQxc3-/view?usp=sharing\nhttps://github.com/lijiannuist/lightDSFD\nweight: https://github.com/lijiannuist/lightDSFD/raw/master/weights/light_DSFD.pth\nhttps://github.com/cydonia999/VGGFace2-pytorch\nhttps://github.com/cydonia999/VGGFace2-pytorch#pretrained-models\nhttps://github.com/albumentations-team/albumentations", "votes": 1 }, { "id": 762097, "postDate": "2020-03-03T07:53:09.040Z", "content": "@juliaelliott \nTo ensure Fairness and Transparency, upon Competition ended and preliminary Private leaderboard score and place has been calculated, will you ask the top 10 Kernels to fully disclose the External dataset they used (together with their kernel code) for Public scrutiny BEFORE the confirmation of any award? This will allowed the Kaggle community to help organizer in catching any form of cheating or overlook.
", "rawMarkdown": "@juliaelliott \nTo ensure Fairness and Transparency, upon Competition ended and preliminary Private leaderboard score and place has been calculated, will you ask the top 10 Kernels to fully disclose the External dataset they used (together with their kernel code) for Public scrutiny BEFORE the confirmation of any award? This will allowed the Kaggle community to help organizer in catching any form of cheating or overlook.", "votes": 1, "replies": [ { "id": 763233, "postDate": "2020-03-04T09:20:43.430Z", "content": "As per the rules of this competition, the winners' licensing terms require that the solution be open sourced, and we can all inspect the external data used by the winners along with their complete source code.
\n\nAlso, I believe that the organizer will reproduce the winning solutions end-to-end: retrain, test, and validate scores. During this process, they will validate all external data and licenses as well.
\n\nPS: We can ensure repeatability of scores by properly setting the seeds for random number generators in numpy, Torch, TF/Keras.
", "rawMarkdown": "As per the rules of this competition, the winners' licensing terms require that the solution be open sourced, and we can all inspect the external data used by the winners along with their complete source code.\n\nAlso, I believe that the organizer will reproduce the winning solutions end-to-end: retrain, test, and validate scores. During this process, they will validate all external data and licenses as well.\n\nPS: We can ensure repeatability of scores by properly setting the seeds for random number generators in numpy, Torch, TF/Keras." } ] }, { "id": 762082, "postDate": "2020-03-03T07:38:13.020Z", "content": "https://github.com/hukkelas/DSFD-Pytorch-Inference\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/sfzhang15/FaceBoxes\nhttps://github.com/TropComplique/FaceBoxes-tensorflow\nhttps://github.com/XiaXuehai/faceboxes\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/peteryuX/retinaface-tf2\nhttps://github.com/OFRIN/Tensorflow_RetinaFace\nhttps://github.com/yeephycho/tensorflow-face-detection\nhttps://v.qq.com/channel/choice\nwww.youtube.com\nwww.youku.com\nwww.tudou.com\nwww.tiktok.com\nwww.douyin.com\nwww.tonton.com.my\nhttp://tv.cctv.com/\nwww.bbc.com\nwww.cnn.com\nwww.foxnews.com\nwww.iqiyi.com\nhttps://github.com/deepmind/kinetics-i3d\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/yaohungt/Gated-Spatio-Temporal-Energy-Graph\nhttps://github.com/dlpbc/keras-kinetics-i3d\nhttps://github.com/huangyangyu/SeqFace\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/ronghuaiyang/arcface-pytorch\nhttps://github.com/happynear/AMSoftmax\nhttps://github.com/foamliu/InsightFace-v2\nhttps://github.com/1996scarlet/ArcFace-Multiplex-Recognition\nhttps://github.com/xiaoboCASIA/SV-X-Softmax\nhttps://github.com/foamliu/InsightFace-PyTorch\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/iloveuu2011/Face-Detection-Library\nhttps://github.com/iloveuu2011/retinaface-tf2\nhttps://github.com/pvskand/DisguiseNet\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/EricZgw/PyramidBox\nhttps://github.com/swghosh/DeepFace\nhttps://github.com/zma-c-137/VarGFaceNet\nhttps://github.com/cvtower/seesawfacenet_pytorch\nhttps://github.com/kk7nc/RMDL\nhttps://github.com/ZhaoJ9014/High-Performance-Face-Recognition\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/rlaengud123/CMC_LRCN\nhttps://github.com/doronharitan/human_activity_recognition_LRCN\nhttps://github.com/piergiaj/representation-flow-cvpr19\nhttps://github.com/piergiaj/evanet-iccv19\nhttps://github.com/piergiaj\nhttps://github.com/piergiaj/mlb-youtube\nhttps://github.com/craston/MARS\nhttps://github.com/clancylian/retinaface\nhttps://github.com/yangfly\nhttps://pan.baidu.com/share/init?surl=P1ypO7VYUbNAezdvLm2m9w\nhttps://github.com/biubug6/Face-Detector-1MB-with-landmark\nhttps://github.com/610265158/DSFD-tensorflow
", "rawMarkdown": "https://github.com/hukkelas/DSFD-Pytorch-Inference\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/sfzhang15/FaceBoxes\nhttps://github.com/TropComplique/FaceBoxes-tensorflow\nhttps://github.com/XiaXuehai/faceboxes\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/peteryuX/retinaface-tf2\nhttps://github.com/OFRIN/Tensorflow_RetinaFace\nhttps://github.com/yeephycho/tensorflow-face-detection\nhttps://v.qq.com/channel/choice\nwww.youtube.com\nwww.youku.com\nwww.tudou.com\nwww.tiktok.com\nwww.douyin.com\nwww.tonton.com.my\nhttp://tv.cctv.com/\nwww.bbc.com\nwww.cnn.com\nwww.foxnews.com\nwww.iqiyi.com\nhttps://github.com/deepmind/kinetics-i3d\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/yaohungt/Gated-Spatio-Temporal-Energy-Graph\nhttps://github.com/dlpbc/keras-kinetics-i3d\nhttps://github.com/huangyangyu/SeqFace\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/ronghuaiyang/arcface-pytorch\nhttps://github.com/happynear/AMSoftmax\nhttps://github.com/foamliu/InsightFace-v2\nhttps://github.com/1996scarlet/ArcFace-Multiplex-Recognition\nhttps://github.com/xiaoboCASIA/SV-X-Softmax\nhttps://github.com/foamliu/InsightFace-PyTorch\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/iloveuu2011/Face-Detection-Library\nhttps://github.com/iloveuu2011/retinaface-tf2\nhttps://github.com/pvskand/DisguiseNet\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/EricZgw/PyramidBox\nhttps://github.com/swghosh/DeepFace\nhttps://github.com/zma-c-137/VarGFaceNet\nhttps://github.com/cvtower/seesawfacenet_pytorch\nhttps://github.com/kk7nc/RMDL\nhttps://github.com/ZhaoJ9014/High-Performance-Face-Recognition\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/rlaengud123/CMC_LRCN\nhttps://github.com/doronharitan/human_activity_recognition_LRCN\nhttps://github.com/piergiaj/representation-flow-cvpr19\nhttps://github.com/piergiaj/evanet-iccv19\nhttps://github.com/piergiaj\nhttps://github.com/piergiaj/mlb-youtube\nhttps://github.com/craston/MARS\nhttps://github.com/clancylian/retinaface\nhttps://github.com/yangfly\nhttps://pan.baidu.com/share/init?surl=P1ypO7VYUbNAezdvLm2m9w\nhttps://github.com/biubug6/Face-Detector-1MB-with-landmark\nhttps://github.com/610265158/DSFD-tensorflow", "votes": 1 }, { "id": 761474, "postDate": "2020-03-02T14:59:06.247Z", "content": "A python package to analyze and compare voices with deep learning\nhttps://github.com/resemble-ai/Resemblyzer
Real-Time Voice Cloning\nhttps://github.com/CorentinJ/Real-Time-Voice-Cloning
LibriSpeech ASR corpus\nhttp://www.openslr.org/12/
LibriTTS corpus\nhttp://www.openslr.org/60/
The VoxCeleb Dataset\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1.html\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox2.html
The M-AILABS Speech Dataset\nhttps://www.caito.de/2019/01/the-m-ailabs-speech-dataset/
MrDeepfakes\nhttps://mrdeepfakes.com/terms
English Multi-speaker Corpus for CSTR Voice Cloning Toolkit\nhttps://homepages.inf.ed.ac.uk/jyamagis/page3/page58/page58.html\nhttps://datashare.is.ed.ac.uk/handle/10283/2651
FaceForensics original repository\nhttps://github.com/ondyari/FaceForensics/tree/original
IEEE's Signal Processing Society - Camera Model Identification\nhttps://www.kaggle.com/c/sp-society-camera-model-identification
BlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch
", "rawMarkdown": "BlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\n\n\n", "votes": 1, "replies": [ { "id": 763741, "postDate": "2020-03-04T20:01:29.143Z", "content": "https://github.com/AlexanderParkin/ChaLearn_liveness_challenge
", "rawMarkdown": "https://github.com/AlexanderParkin/ChaLearn_liveness_challenge" } ] }, { "id": 760744, "postDate": "2020-03-01T17:28:03.687Z", "content": "dlib: https://github.com/davisking/dlib\ndlib-models: https://github.com/davisking/dlib-models/blob/master/shape_predictor_68_face_landmarks.dat.bz2\nnumpy: https://github.com/numpy/numpy\nopencv: https://github.com/opencv/opencv\nscipy: https://github.com/scipy/scipy\nimutils: https://pypi.org/project/imutils/\nripser: https://github.com/scikit-tda/ripser.py\nglob: https://docs.python.org/3/library/glob.html\nos: https://docs.python.org/3/library/os.html\nmultiprocessing: https://docs.python.org/3/library/multiprocessing.html\nscikit-learn: https://scikit-learn.org/stable/index.html\nScikit-TDA: https://github.com/scikit-tda\ncsv: https://docs.python.org/3/library/csv.html
", "rawMarkdown": "dlib: https://github.com/davisking/dlib\ndlib-models: https://github.com/davisking/dlib-models/blob/master/shape_predictor_68_face_landmarks.dat.bz2\nnumpy: https://github.com/numpy/numpy\nopencv: https://github.com/opencv/opencv\nscipy: https://github.com/scipy/scipy\nimutils: https://pypi.org/project/imutils/\nripser: https://github.com/scikit-tda/ripser.py\nglob: https://docs.python.org/3/library/glob.html\nos: https://docs.python.org/3/library/os.html\nmultiprocessing: https://docs.python.org/3/library/multiprocessing.html\nscikit-learn: https://scikit-learn.org/stable/index.html\nScikit-TDA: https://github.com/scikit-tda\ncsv: https://docs.python.org/3/library/csv.html", "votes": 1 }, { "id": 758759, "postDate": "2020-02-28T05:41:37.520Z", "content": "opencv extra modules\nhttps://github.com/opencv/opencv_contrib
\n\npytables\nhttps://pypi.org/project/tables/
\n\nh5py\nhttps://pypi.org/project/h5py/
\n\npandas\nhttps://pypi.org/project/pandas/
\n\nalbumetations\nhttps://pypi.org/project/albumentations/
\n\nTacotron2\nhttps://github.com/NVIDIA/tacotron2
\n\nWaveglow\nhttps://github.com/NVIDIA/waveglow
\n\nParallelWaveGAN\nhttps://github.com/kan-bayashi/ParallelWaveGAN
\n\nMozilla TTS\nhttps://github.com/mozilla/TTS
\n\nCatalyst framework\nhttps://github.com/catalyst-team/catalyst
\n\nPytorch-toolbelt & Examples\nhttps://github.com/BloodAxe/pytorch-toolbelt\nhttps://github.com/BloodAxe/Kaggle-2019-Blindness-Detection
\n\nKornia package\nhttps://github.com/kornia/kornia
", "rawMarkdown": "opencv extra modules\nhttps://github.com/opencv/opencv_contrib\n\npytables\nhttps://pypi.org/project/tables/\n\nh5py\nhttps://pypi.org/project/h5py/\n\npandas\nhttps://pypi.org/project/pandas/\n\nalbumetations\nhttps://pypi.org/project/albumentations/\n\nTacotron2\nhttps://github.com/NVIDIA/tacotron2\n\nWaveglow\nhttps://github.com/NVIDIA/waveglow\n\nParallelWaveGAN\nhttps://github.com/kan-bayashi/ParallelWaveGAN\n\nMozilla TTS\nhttps://github.com/mozilla/TTS\n\nCatalyst framework\nhttps://github.com/catalyst-team/catalyst\n\nPytorch-toolbelt & Examples\nhttps://github.com/BloodAxe/pytorch-toolbelt\nhttps://github.com/BloodAxe/Kaggle-2019-Blindness-Detection\n\nKornia package\nhttps://github.com/kornia/kornia", "votes": 1 }, { "id": 758018, "postDate": "2020-02-27T11:12:15.163Z", "content": "https://github.com/yxlijun/S3FD.pytorch
", "rawMarkdown": "https://github.com/yxlijun/S3FD.pytorch", "votes": 1 }, { "id": 757847, "postDate": "2020-02-27T07:02:35.427Z", "content": "Pretrained models disclosure from two repositories:\nhttps://pytorch.org/docs/stable/torchvision/models.html\nand\nhttps://modelzoo.co/framework/pytorch
", "rawMarkdown": "Pretrained models disclosure from two repositories:\nhttps://pytorch.org/docs/stable/torchvision/models.html\nand\nhttps://modelzoo.co/framework/pytorch", "votes": 1 }, { "id": 757253, "postDate": "2020-02-26T15:16:09.447Z", "content": "https://keras.io/applications\nhttps://www.tensorflow.org/lite/models\nhttps://github.com/NVIDIA/tensorflow-determinism\nhttps://github.com/tensorflow/models/blob/master/official/README.md\nhttps://github.com/tensorflow/models/blob/master/official/README-TPU.md
", "rawMarkdown": "https://keras.io/applications\nhttps://www.tensorflow.org/lite/models\nhttps://github.com/NVIDIA/tensorflow-determinism\nhttps://github.com/tensorflow/models/blob/master/official/README.md\nhttps://github.com/tensorflow/models/blob/master/official/README-TPU.md", "votes": 1, "replies": [ { "id": 757807, "postDate": "2020-02-27T05:54:40.543Z", "content": "+1
", "rawMarkdown": "+1" } ] }, { "id": 756800, "postDate": "2020-02-26T05:02:26.607Z", "content": "External Data Disclosure
\n\nNVlabs ffhq dataset \nhttps://drive.google.com/open?id=1u2xu7bSrWxrbUxk-dT-UvEJq8IjdmNTP
\n\nUADFV dataset \nhttps://drive.google.com/drive/folders/1GEk1DSxmlV_61JtpEGzC9Fo_BffvyxpH?usp=sharing
\n\nDataset from ASVSpoof 2019 competition \nhttps://datashare.is.ed.ac.uk/handle/10283/3336
\n\nKaggle CIPLAB\nhttps://www.kaggle.com/ciplab/real-and-fake-face-detection/download
\n\nOpenCV face detector model
\n\nDLIB face detector model
\n\nVoxCeleb\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/
\n\nresnet and xception for learning transfer (I'm testing with both tensorflow and pytorch) using imagenet as baseline.
\n\nhttps://github.com/thiago1080/SphereFace\nhttps://github.com/vlad3996/FaceDetection-DSFD\n https://github.com/EndlessSora/DeeperForensics-1.0
", "rawMarkdown": "External Data Disclosure\n\nNVlabs ffhq dataset \nhttps://drive.google.com/open?id=1u2xu7bSrWxrbUxk-dT-UvEJq8IjdmNTP\n\nUADFV dataset \nhttps://drive.google.com/drive/folders/1GEk1DSxmlV_61JtpEGzC9Fo_BffvyxpH?usp=sharing\n\nDataset from ASVSpoof 2019 competition \nhttps://datashare.is.ed.ac.uk/handle/10283/3336\n\nKaggle CIPLAB\nhttps://www.kaggle.com/ciplab/real-and-fake-face-detection/download\n\nOpenCV face detector model\n\nDLIB face detector model\n\nVoxCeleb\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/\n\nresnet and xception for learning transfer (I'm testing with both tensorflow and pytorch) using imagenet as baseline.\n\nhttps://github.com/thiago1080/SphereFace\nhttps://github.com/vlad3996/FaceDetection-DSFD\n https://github.com/EndlessSora/DeeperForensics-1.0\n", "votes": 1 }, { "id": 756063, "postDate": "2020-02-25T11:30:07.583Z", "content": "DLIB: https://github.com/davisking/dlib\nfacenet: https://github.com/timesler/facenet-pytorch\nPre-trained models of VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nVGGFace2: https://drive.google.com/uc?export=download&id=1TDZVEBudGaEd5POR5X4ZsMvdsh1h68T1\nface-recognition: https://github.com/ageitgey/face_recognition\nface-recognition-models: https://github.com/ageitgey/face_recognition_models\nautokeras: https://github.com/keras-team/autokeras\nkeras-tuner: https://github.com/keras-team/keras-tuner\nterminaltables: https://github.com/Robpol86/terminaltables
", "rawMarkdown": "DLIB: https://github.com/davisking/dlib\nfacenet: https://github.com/timesler/facenet-pytorch\nPre-trained models of VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nVGGFace2: https://drive.google.com/uc?export=download&id=1TDZVEBudGaEd5POR5X4ZsMvdsh1h68T1\nface-recognition: https://github.com/ageitgey/face_recognition\nface-recognition-models: https://github.com/ageitgey/face_recognition_models\nautokeras: https://github.com/keras-team/autokeras\nkeras-tuner: https://github.com/keras-team/keras-tuner\nterminaltables: https://github.com/Robpol86/terminaltables", "votes": 1 }, { "id": 755549, "postDate": "2020-02-24T21:57:53.050Z", "content": "keras xception imagenet weights https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xception_weights_tf_dim_ordering_tf_kernels_notop.h5
", "rawMarkdown": "keras xception imagenet weights https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xception_weights_tf_dim_ordering_tf_kernels_notop.h5", "votes": 1 }, { "id": 753252, "postDate": "2020-02-21T23:38:53.847Z", "content": "youtube-bb dataset\nhttps://research.google.com/youtube-bb/
", "rawMarkdown": "youtube-bb dataset\nhttps://research.google.com/youtube-bb/", "votes": 1 }, { "id": 750901, "postDate": "2020-02-19T19:56:42.073Z", "content": "vggface2 dataset : http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nand also\nhttps://www.kaggle.com/c/imagenet-object-localization-challenge/data
\n\nJust a question, are those licenses ok with dataset limitation : \n- Creative Commons Attribution-ShareAlike ?\n- and Creative Commons Attribution-NonCommercial-ShareAlike ?
", "rawMarkdown": "vggface2 dataset : http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nand also\nhttps://www.kaggle.com/c/imagenet-object-localization-challenge/data\n\nJust a question, are those licenses ok with dataset limitation : \n- Creative Commons Attribution-ShareAlike ?\n- and Creative Commons Attribution-NonCommercial-ShareAlike ?\n", "votes": 1 }, { "id": 750278, "postDate": "2020-02-19T08:48:53.387Z", "content": "http://www.robots.ox.ac.uk/~vgg/data/voxceleb/
", "rawMarkdown": "http://www.robots.ox.ac.uk/~vgg/data/voxceleb/", "votes": 1, "replies": [ { "id": 750898, "postDate": "2020-02-19T19:51:11.923Z", "content": "this requires permission/ask for login so I don't think it's allowed @juliaelliott
", "rawMarkdown": "this requires permission/ask for login so I don't think it's allowed @juliaelliott " }, { "id": 752690, "postDate": "2020-02-21T09:59:47.333Z", "content": "They provide direct links w/o login requirement:
\n\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/data/vox2_dev_txt.zip\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/data/vox2_test_txt.zip
", "rawMarkdown": "They provide direct links w/o login requirement:\n\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/data/vox2_dev_txt.zip\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/data/vox2_test_txt.zip", "votes": 1 } ] }, { "id": 749115, "postDate": "2020-02-18T10:07:37.107Z", "content": "Pretrained models from:\nhttps://github.com/rwightman/pytorch-image-models\nhttps://github.com/rwightman/gen-efficientnet-pytorch
", "rawMarkdown": "Pretrained models from:\nhttps://github.com/rwightman/pytorch-image-models\nhttps://github.com/rwightman/gen-efficientnet-pytorch", "votes": 1 }, { "id": 748620, "postDate": "2020-02-17T18:51:56.490Z", "content": "https://github.com/qubvel/segmentation_models.pytorch
", "rawMarkdown": "https://github.com/qubvel/segmentation_models.pytorch", "votes": 1 }, { "id": 740961, "postDate": "2020-02-10T02:30:58.627Z", "content": "Keras applications models\nhttps://keras.io/applications/
", "rawMarkdown": "Keras applications models\nhttps://keras.io/applications/\n", "votes": 1 }, { "id": 737834, "postDate": "2020-02-05T20:40:10.610Z", "content": "https://github.com/NVlabs/ffhq-dataset\nhttps://www.kaggle.com/jessicali9530/celeba-dataset\nhttps://susanqq.github.io/UTKFace/
", "rawMarkdown": "https://github.com/NVlabs/ffhq-dataset\nhttps://www.kaggle.com/jessicali9530/celeba-dataset\nhttps://susanqq.github.io/UTKFace/ ", "votes": 1 }, { "id": 736429, "postDate": "2020-02-04T07:04:34.133Z", "content": "Using this set of video tools uploaded by sheldon robinson on kaggle. Didn't find a comment here, so I hope it helps someone else too. Dataset here. Tutorial for installation here. Tools include scikit-video, ffmpeg, moviepy etc. I will primarily be using scikit-video and moviepy.
", "rawMarkdown": "Using this set of video tools uploaded by sheldon robinson on kaggle. Didn't find a comment here, so I hope it helps someone else too. Dataset [here](https://www.kaggle.com/sheldonrobinson/video-tools). Tutorial for installation [here](https://www.kaggle.com/sheldonrobinson/starter-video-tools/notebook). Tools include scikit-video, ffmpeg, moviepy etc. I will primarily be using scikit-video and moviepy.", "votes": 1 }, { "id": 730492, "postDate": "2020-01-27T14:53:11.950Z", "content": "Keras ResNet50 (MIT license) and Resnet50V2 (BSD license) models and pre-trained weights from https://keras.io/applications/#resnet
", "rawMarkdown": "Keras ResNet50 (MIT license) and Resnet50V2 (BSD license) models and pre-trained weights from https://keras.io/applications/#resnet", "votes": 1 }, { "id": 707331, "postDate": "2019-12-31T16:52:23.010Z", "content": "opencvdnnfp16 - Dataset containing opencv face detection Caffe model - courtesy Milandu Keith\ntimesler/facenet-pytorch-vggface2 Facenet with pretrained weights - courtesy timesler\nResNet with pre-trained weights
", "rawMarkdown": "[opencvdnnfp16](https://www.kaggle.com/mbmk92/opencvdnnfp16) - Dataset containing opencv face detection Caffe model - courtesy Milandu Keith\n[timesler/facenet-pytorch-vggface2](https://www.kaggle.com/timesler/facenet-pytorch-vggface2) Facenet with pretrained weights - courtesy timesler\nResNet with pre-trained weights", "votes": 1 }, { "id": 705166, "postDate": "2019-12-28T15:20:46.887Z", "content": "Pretrained models on VggFace2: https://github.com/cydonia999/VGGFace2-pytorch#pretrained-models
", "rawMarkdown": "Pretrained models on VggFace2: https://github.com/cydonia999/VGGFace2-pytorch#pretrained-models", "votes": 1 }, { "id": 704584, "postDate": "2019-12-27T16:50:18.093Z", "content": "BlazeFace for face and landmark detection: https://sites.google.com/view/perception-cv4arvr/blazeface
\n\nEdit: I added this as a data source and a demo kernel.
", "rawMarkdown": "BlazeFace for face and landmark detection: https://sites.google.com/view/perception-cv4arvr/blazeface\n\nEdit: I added this as a [data source](https://www.kaggle.com/humananalog/blazeface-pytorch) and a [demo kernel](https://www.kaggle.com/humananalog/starter-blazeface-pytorch).", "votes": 1 }, { "id": 701639, "postDate": "2019-12-23T17:49:52.270Z", "content": "RetinaFace Pytorch detector: https://github.com/biubug6/Pytorch_Retinaface
", "rawMarkdown": "RetinaFace Pytorch detector: https://github.com/biubug6/Pytorch_Retinaface", "votes": 1 }, { "id": 695988, "postDate": "2019-12-16T01:02:16.160Z", "content": "opencv dnn face detector: \nhttps://raw.githubusercontent.com/opencv/opencv_3rdparty/dnn_samples_face_detector_20180220_uint8/opencv_face_detector_uint8.pb\nhttps://github.com/opencv/opencv/blob/master/samples/dnn/face_detector/opencv_face_detector.pbtxt
", "rawMarkdown": "opencv dnn face detector: \nhttps://raw.githubusercontent.com/opencv/opencv_3rdparty/dnn_samples_face_detector_20180220_uint8/opencv_face_detector_uint8.pb\nhttps://github.com/opencv/opencv/blob/master/samples/dnn/face_detector/opencv_face_detector.pbtxt", "votes": 1, "replies": [ { "id": 743205, "postDate": "2020-02-11T22:02:57.940Z", "content": "opencv pretrainde haar cascade: https://github.com/opencv/opencv/tree/master/data/haarcascades
", "rawMarkdown": "opencv pretrainde haar cascade: https://github.com/opencv/opencv/tree/master/data/haarcascades" } ] }, { "id": 695274, "postDate": "2019-12-14T22:13:24.190Z", "content": "MTCNN-package\nhttps://pypi.org/project/mtcnn/
", "rawMarkdown": "MTCNN-package\n[https://pypi.org/project/mtcnn/](https://pypi.org/project/mtcnn/)", "votes": 1 }, { "id": 762816, "postDate": "2020-03-03T20:30:55.987Z", "content": "models and weights :\nhttps://github.com/hukkelas/DSFD-Pytorch-Inference\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/sfzhang15/FaceBoxes\nhttps://github.com/TropComplique/FaceBoxes-tensorflow\nhttps://github.com/XiaXuehai/faceboxes\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/peteryuX/retinaface-tf2\nhttps://github.com/OFRIN/Tensorflow_RetinaFace\nhttps://github.com/yeephycho/tensorflow-face-detection\nhttps://github.com/deepmind/kinetics-i3d\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/yaohungt/Gated-Spatio-Temporal-Energy-Graph\nhttps://github.com/dlpbc/keras-kinetics-i3d\nhttps://github.com/huangyangyu/SeqFace\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/ronghuaiyang/arcface-pytorch\nhttps://github.com/happynear/AMSoftmax\nhttps://github.com/foamliu/InsightFace-v2\nhttps://github.com/1996scarlet/ArcFace-Multiplex-Recognition\nhttps://github.com/xiaoboCASIA/SV-X-Softmax\nhttps://github.com/foamliu/InsightFace-PyTorch\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/iloveuu2011/Face-Detection-Library\nhttps://github.com/iloveuu2011/retinaface-tf2\nhttps://github.com/pvskand/DisguiseNet\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/EricZgw/PyramidBox\nhttps://github.com/swghosh/DeepFace\nhttps://github.com/HRNet/HRNet-Facial-Landmark-Detection\nhttps://github.com/zma-c-137/VarGFaceNet\nhttps://github.com/cvtower/seesawfacenet_pytorch\nhttps://github.com/kk7nc/RMDL\nhttps://github.com/ZhaoJ9014/High-Performance-Face-Recognition\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/rlaengud123/CMC_LRCN\nhttps://github.com/doronharitan/human_activity_recognition_LRCN\nhttps://github.com/piergiaj/representation-flow-cvpr19\nhttps://github.com/piergiaj/evanet-iccv19\nhttps://github.com/piergiaj\nhttps://github.com/piergiaj/mlb-youtube\nhttps://github.com/craston/MARS\nRetinaFace : https://github.com/deepinsight/insightface/tree/master/RetinaFace\nKeras Xception Imagenet Weights https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xceptionweightstfdimorderingtfkernels_notop.h5\nEfficient Net Weights https://github.com/qubvel/efficientnet\nImgaug: https://imgaug.readthedocs.io/en/latest/\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/albumentations-team/albumentations\nMesoNet - https://github.com/DariusAf/MesoNet\nopencv extra modules\nhttps://github.com/opencv/opencv_contrib\npytables\nhttps://pypi.org/project/tables/\nh5py\nhttps://pypi.org/project/h5py/\npandas\nhttps://pypi.org/project/pandas/\nalbumetations\nhttps://pypi.org/project/albumentations/\nTacotron2\nhttps://github.com/NVIDIA/tacotron2\nWaveglow\nhttps://github.com/NVIDIA/waveglow\nParallelWaveGAN\nhttps://github.com/kan-bayashi/ParallelWaveGAN\nMozilla TTS\nhttps://github.com/mozilla/TTS\nCatalyst framework\nhttps://github.com/catalyst-team/catalyst\nPytorch-toolbelt & Examples\nhttps://github.com/BloodAxe/pytorch-toolbelt\nhttps://github.com/BloodAxe/Kaggle-2019-Blindness-Detection\nKornia package\nhttps://github.com/kornia/kornia\nNetworks and weights in github\neffecientnet ( https://github.com/lukemelas/EfficientNet-PyTorch ) - Google weights\n'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b0-355c32eb.pth',\n'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b1-f1951068.pth',\n'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b2-8bb594d6.pth',\n'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b3-5fb5a3c3.pth',\n'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b4-6ed6700e.pth',\n'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b5-b6417697.pth',\n'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b6-c76e70fd.pth',\n'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b7-dcc49843.pth'
\n\n'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b0-b64d5a18.pth',\n'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b1-0f3ce85a.pth',\n'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b2-6e9d97e5.pth',\n'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b3-cdd7c0f4.pth',\n'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b4-44fb3a87.pth',\n'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b5-86493f6b.pth',\n'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b6-ac80338e.pth',\n'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b7-4652b6dd.pth',\n'efficientnet-b8': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b8-22a8fe65.pth'
\n\nModels and weights:\nEfficientNet Keras (and TensorFlow Keras)\nhttps://github.com/qubvel/efficientnet\nClassification models Zoo - Keras (and TensorFlow Keras)\nhttps://github.com/qubvel/classification_models\nPython library with Neural Networks for Image Segmentation based on Keras and TensorFlow. \nhttps://github.com/qubvel/segmentation_models
\n\nmtccn\npytorch model zoo\nopencv2\nscikit learn\nfastai\nEfficientNet-pyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nresnext101: https://github.com/facebookresearch/WSL-Images\nface.evoLVe.PyTorch: https://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nResidualAttentionNetwork-pytorch: https://github.com/tengshaofeng/ResidualAttentionNetwork-pytorch\nhrnet: https://github.com/HRNet/HRNet-Image-Classification\ntorchvision models: https://github.com/pytorch/vision/tree/master/torchvision/models\nFaceForensics++ dataset: https://github.com/ondyari/FaceForensics\ntrainmsra.tar.gz traincelebrity.tar.gz: http://trillionpairs.deepglint.com/data
\n\nwider face datasets: http://shuoyang1213.me/WIDERFACE/\nresnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n'resnext5032x4d': 'https://download.pytorch.org/models/resnext5032x4d-7cdf4587.pth',\n'resnext10132x8d': 'https://download.pytorch.org/models/resnext10132x8d-8ba56ff5.pth',\n'wideresnet502': 'https://download.pytorch.org/models/wideresnet502-95faca4d.pth',\n'wideresnet1012': 'https://download.pytorch.org/models/wideresnet1012-32ee1156.pth',\nhttps://github.com/qubvel/segmentation_models.pytorch\nPretrained weights from https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nKeras applications models\nhttps://keras.io/applications/\nInceptionV3 imagenet weights from Keras:\nhttps://keras.io/applications/#inceptionv3\nfacenet-pytorch pre-trained models: https://github.com/timesler/facenet-pytorch\n(MIT license)\nPre-trained MTCNN model from here: https://github.com/timesler/facenet-pytorch\nimutils package: https://github.com/jrosebr1/imutils\nRetinaFace Pytorch detector: https://github.com/biubug6/Pytorch_Retinaface\nopencv dnn face detector:\nhttps://raw.githubusercontent.com/opencv/opencv_3rdparty/dnn_samples_face_detector_20180220_uint8/opencv_face_detector_uint8.pb\nhttps://github.com/opencv/opencv/blob/master/samples/dnn/face_detector/opencv_face_detector.pbtxt\nhttps://github.com/nyoki-mtl/keras-facenet\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/huawei-noah/ghostnet\nhttps://github.com/iamhankai/ghostnet.pytorch\nhttps://github.com/kuan-wang/pytorch-mobilenet-v3\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nhttps://keras.io/applications/\nDatasets:\nImageNet\nhttps://github.com/Daniil-Osokin/lightweight-human-pose-estimation.pytorch\nhttps://github.com/HRNet\nhttps://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py
", "rawMarkdown": "models and weights :\nhttps://github.com/hukkelas/DSFD-Pytorch-Inference\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/sfzhang15/FaceBoxes\nhttps://github.com/TropComplique/FaceBoxes-tensorflow\nhttps://github.com/XiaXuehai/faceboxes\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/peteryuX/retinaface-tf2\nhttps://github.com/OFRIN/Tensorflow_RetinaFace\nhttps://github.com/yeephycho/tensorflow-face-detection\nhttps://github.com/deepmind/kinetics-i3d\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/yaohungt/Gated-Spatio-Temporal-Energy-Graph\nhttps://github.com/dlpbc/keras-kinetics-i3d\nhttps://github.com/huangyangyu/SeqFace\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/ronghuaiyang/arcface-pytorch\nhttps://github.com/happynear/AMSoftmax\nhttps://github.com/foamliu/InsightFace-v2\nhttps://github.com/1996scarlet/ArcFace-Multiplex-Recognition\nhttps://github.com/xiaoboCASIA/SV-X-Softmax\nhttps://github.com/foamliu/InsightFace-PyTorch\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/iloveuu2011/Face-Detection-Library\nhttps://github.com/iloveuu2011/retinaface-tf2\nhttps://github.com/pvskand/DisguiseNet\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/EricZgw/PyramidBox\nhttps://github.com/swghosh/DeepFace\nhttps://github.com/HRNet/HRNet-Facial-Landmark-Detection\nhttps://github.com/zma-c-137/VarGFaceNet\nhttps://github.com/cvtower/seesawfacenet_pytorch\nhttps://github.com/kk7nc/RMDL\nhttps://github.com/ZhaoJ9014/High-Performance-Face-Recognition\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/rlaengud123/CMC_LRCN\nhttps://github.com/doronharitan/human_activity_recognition_LRCN\nhttps://github.com/piergiaj/representation-flow-cvpr19\nhttps://github.com/piergiaj/evanet-iccv19\nhttps://github.com/piergiaj\nhttps://github.com/piergiaj/mlb-youtube\nhttps://github.com/craston/MARS\nRetinaFace : https://github.com/deepinsight/insightface/tree/master/RetinaFace\nKeras Xception Imagenet Weights https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xceptionweightstfdimorderingtfkernels_notop.h5\nEfficient Net Weights https://github.com/qubvel/efficientnet\nImgaug: https://imgaug.readthedocs.io/en/latest/\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/albumentations-team/albumentations\nMesoNet - https://github.com/DariusAf/MesoNet\nopencv extra modules\nhttps://github.com/opencv/opencv_contrib\npytables\nhttps://pypi.org/project/tables/\nh5py\nhttps://pypi.org/project/h5py/\npandas\nhttps://pypi.org/project/pandas/\nalbumetations\nhttps://pypi.org/project/albumentations/\nTacotron2\nhttps://github.com/NVIDIA/tacotron2\nWaveglow\nhttps://github.com/NVIDIA/waveglow\nParallelWaveGAN\nhttps://github.com/kan-bayashi/ParallelWaveGAN\nMozilla TTS\nhttps://github.com/mozilla/TTS\nCatalyst framework\nhttps://github.com/catalyst-team/catalyst\nPytorch-toolbelt & Examples\nhttps://github.com/BloodAxe/pytorch-toolbelt\nhttps://github.com/BloodAxe/Kaggle-2019-Blindness-Detection\nKornia package\nhttps://github.com/kornia/kornia\nNetworks and weights in github\neffecientnet ( https://github.com/lukemelas/EfficientNet-PyTorch ) - Google weights\n'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b0-355c32eb.pth',\n'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b1-f1951068.pth',\n'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b2-8bb594d6.pth',\n'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b3-5fb5a3c3.pth',\n'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b4-6ed6700e.pth',\n'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b5-b6417697.pth',\n'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b6-c76e70fd.pth',\n'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b7-dcc49843.pth'\n\n'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b0-b64d5a18.pth',\n'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b1-0f3ce85a.pth',\n'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b2-6e9d97e5.pth',\n'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b3-cdd7c0f4.pth',\n'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b4-44fb3a87.pth',\n'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b5-86493f6b.pth',\n'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b6-ac80338e.pth',\n'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b7-4652b6dd.pth',\n'efficientnet-b8': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b8-22a8fe65.pth'\n\nModels and weights:\nEfficientNet Keras (and TensorFlow Keras)\nhttps://github.com/qubvel/efficientnet\nClassification models Zoo - Keras (and TensorFlow Keras)\nhttps://github.com/qubvel/classification_models\nPython library with Neural Networks for Image Segmentation based on Keras and TensorFlow. \nhttps://github.com/qubvel/segmentation_models\n\nmtccn\npytorch model zoo\nopencv2\nscikit learn\nfastai\nEfficientNet-pyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nresnext101: https://github.com/facebookresearch/WSL-Images\nface.evoLVe.PyTorch: https://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nResidualAttentionNetwork-pytorch: https://github.com/tengshaofeng/ResidualAttentionNetwork-pytorch\nhrnet: https://github.com/HRNet/HRNet-Image-Classification\ntorchvision models: https://github.com/pytorch/vision/tree/master/torchvision/models\nFaceForensics++ dataset: https://github.com/ondyari/FaceForensics\ntrainmsra.tar.gz traincelebrity.tar.gz: http://trillionpairs.deepglint.com/data\n\nwider face datasets: http://shuoyang1213.me/WIDERFACE/\nresnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n'resnext5032x4d': 'https://download.pytorch.org/models/resnext5032x4d-7cdf4587.pth',\n'resnext10132x8d': 'https://download.pytorch.org/models/resnext10132x8d-8ba56ff5.pth',\n'wideresnet502': 'https://download.pytorch.org/models/wideresnet502-95faca4d.pth',\n'wideresnet1012': 'https://download.pytorch.org/models/wideresnet1012-32ee1156.pth',\nhttps://github.com/qubvel/segmentation_models.pytorch\nPretrained weights from https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nKeras applications models\nhttps://keras.io/applications/\nInceptionV3 imagenet weights from Keras:\nhttps://keras.io/applications/#inceptionv3\nfacenet-pytorch pre-trained models: https://github.com/timesler/facenet-pytorch\n(MIT license)\nPre-trained MTCNN model from here: https://github.com/timesler/facenet-pytorch\nimutils package: https://github.com/jrosebr1/imutils\nRetinaFace Pytorch detector: https://github.com/biubug6/Pytorch_Retinaface\nopencv dnn face detector:\nhttps://raw.githubusercontent.com/opencv/opencv_3rdparty/dnn_samples_face_detector_20180220_uint8/opencv_face_detector_uint8.pb\nhttps://github.com/opencv/opencv/blob/master/samples/dnn/face_detector/opencv_face_detector.pbtxt\nhttps://github.com/nyoki-mtl/keras-facenet\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/huawei-noah/ghostnet\nhttps://github.com/iamhankai/ghostnet.pytorch\nhttps://github.com/kuan-wang/pytorch-mobilenet-v3\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nhttps://keras.io/applications/\nDatasets:\nImageNet\nhttps://github.com/Daniil-Osokin/lightweight-human-pose-estimation.pytorch\nhttps://github.com/HRNet\nhttps://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py\n", "votes": 2 }, { "id": 761637, "postDate": "2020-03-02T19:44:19.143Z", "content": "http://www.openslr.org/17\nhttp://www.openslr.org/26
\n\nhttp://kaldi-asr.org/models/m4
\n\nhttps://github.com/hitachi-speech/EEND\nhttps://github.com/WeidiXie/VGG-Speaker-Recognition
\n\nhttps://github.com/mozilla/DeepSpeech
\n\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb
\n\nhttps://github.com/Fengdalu/LipNet-PyTorch\nhttps://github.com/astorfi/lip-reading-deeplearning\nhttp://www.robots.ox.ac.uk/~vgg/research/deep_lip_reading
", "rawMarkdown": "http://www.openslr.org/17\nhttp://www.openslr.org/26\n\nhttp://kaldi-asr.org/models/m4\n\nhttps://github.com/hitachi-speech/EEND\nhttps://github.com/WeidiXie/VGG-Speaker-Recognition\n\nhttps://github.com/mozilla/DeepSpeech\n\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb\n\nhttps://github.com/Fengdalu/LipNet-PyTorch\nhttps://github.com/astorfi/lip-reading-deeplearning\nhttp://www.robots.ox.ac.uk/~vgg/research/deep_lip_reading", "votes": 2 }, { "id": 761391, "postDate": "2020-03-02T13:15:24.470Z", "content": "https://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttps://github.com/rwightman/pytorch-image-models\nhttps://pypi.org/project/gluoncv2/\nhttps://github.com/ondyari/FaceForensics/tree/master/classification\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/mnikitin/EfficientNet\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://github.com/narumiruna/efficientnet-pytorch\nhttps://github.com/zsef123/EfficientNets-PyTorch\nhttps://mxnet.apache.org/api/python/docs/api/gluon/model_zoo/index.html\nhttp://www.cs.albany.edu/~lsw/celeb-deepfakeforensics.html\nhttps://github.com/ox-vgg/vgg_face2\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/grib0ed0v/face_recognition.pytorch\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nhttps://github.com/deepinsight/insightface\nhttp://shuoyang1213.me/WIDERFACE/\nhttps://github.com/lijiannuist/lightDSFD\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/yxlijun/S3FD.pytorch\nhttps://github.com/supernotman/RetinaFace_Pytorch\nhttps://github.com/bogireddytejareddy/retinaface-pytorch\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/clovaai/EXTD_Pytorch\nhttps://github.com/ElvishElvis/68-Retinaface-Pytorch-version\nhttps://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/sfzhang15/SFD\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/hcl14/retinaface-pytorch-inference\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD
", "rawMarkdown": "https://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttps://github.com/rwightman/pytorch-image-models\nhttps://pypi.org/project/gluoncv2/\nhttps://github.com/ondyari/FaceForensics/tree/master/classification\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/mnikitin/EfficientNet\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://github.com/narumiruna/efficientnet-pytorch\nhttps://github.com/zsef123/EfficientNets-PyTorch\nhttps://mxnet.apache.org/api/python/docs/api/gluon/model_zoo/index.html\nhttp://www.cs.albany.edu/~lsw/celeb-deepfakeforensics.html\nhttps://github.com/ox-vgg/vgg_face2\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/grib0ed0v/face_recognition.pytorch\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nhttps://github.com/deepinsight/insightface\nhttp://shuoyang1213.me/WIDERFACE/\nhttps://github.com/lijiannuist/lightDSFD\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/yxlijun/S3FD.pytorch\nhttps://github.com/supernotman/RetinaFace_Pytorch\nhttps://github.com/bogireddytejareddy/retinaface-pytorch\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/clovaai/EXTD_Pytorch\nhttps://github.com/ElvishElvis/68-Retinaface-Pytorch-version\nhttps://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/sfzhang15/SFD\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/hcl14/retinaface-pytorch-inference\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\n", "votes": 2, "replies": [ { "id": 762013, "postDate": "2020-03-03T05:47:11.137Z", "content": "Note: you may or may not be allowed to use them:
\n\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/133316
", "rawMarkdown": "Note: you may or may not be allowed to use them: \n\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/133316" } ] }, { "id": 760263, "postDate": "2020-03-01T03:35:52.287Z", "content": "RetinaFace : https://github.com/deepinsight/insightface/tree/master/RetinaFace
\n\nKeras Xception Imagenet Weights https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xception_weights_tf_dim_ordering_tf_kernels_notop.h5
\n\nEfficient Net Weights https://github.com/qubvel/efficientnet
\n\nImgaug: https://imgaug.readthedocs.io/en/latest/
", "rawMarkdown": "RetinaFace : [https://github.com/deepinsight/insightface/tree/master/RetinaFace](https://github.com/deepinsight/insightface/tree/master/RetinaFace)\n\nKeras Xception Imagenet Weights [https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xception_weights_tf_dim_ordering_tf_kernels_notop.h5](https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xception_weights_tf_dim_ordering_tf_kernels_notop.h5)\n\nEfficient Net Weights [https://github.com/qubvel/efficientnet](https://github.com/qubvel/efficientnet)\n\nImgaug: [https://imgaug.readthedocs.io/en/latest/](https://imgaug.readthedocs.io/en/latest/)\n\n\n", "votes": 2 }, { "id": 759185, "postDate": "2020-02-28T17:22:28.390Z", "content": "MesoNet - https://github.com/DariusAf/MesoNet
", "rawMarkdown": "MesoNet - https://github.com/DariusAf/MesoNet", "votes": 2 }, { "id": 758640, "postDate": "2020-02-28T01:59:33.917Z", "content": "MTCNN - package
", "rawMarkdown": "MTCNN - package", "votes": 2 }, { "id": 742139, "postDate": "2020-02-11T05:05:10.307Z", "content": "https://www.kaggle.com/caffeinism/helpers
\n\nsome changes in https://www.kaggle.com/humananalog/deepfakes-inference-demo's helpers code
", "rawMarkdown": "https://www.kaggle.com/caffeinism/helpers\n\nsome changes in https://www.kaggle.com/humananalog/deepfakes-inference-demo's helpers code", "votes": 2 }, { "id": 762386, "postDate": "2020-03-03T13:38:51.120Z", "content": "https://www.kaggle.com/c/deepfake-detection-challenge/discussion/123665\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/121173\nhttps://github.com/ondyari/FaceForensics \nhttp://kaldir.vc.in.tum.de/FaceForensics/models/faceforensics++_models.zip\nhttps://youtube.com CC videos
\n\nhttps://github.com/DariusAf/MesoNet\nhttps://github.com/titu1994/keras-efficientnets \nhttps://keras.io/applications/\nhttps://www.tensorflow.org/api_docs/python/tf/keras/applications/\nhttps://github.com/qubvel/efficientnet
\n\nhttps://github.com/KaiyangZhou/deep-person-reid\nhttps://github.com/rwightman/pytorch-image-models\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/pytorch/vision/tree/master/torchvision/models
", "rawMarkdown": "https://www.kaggle.com/c/deepfake-detection-challenge/discussion/123665\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/121173\nhttps://github.com/ondyari/FaceForensics \nhttp://kaldir.vc.in.tum.de/FaceForensics/models/faceforensics++_models.zip\nhttps://youtube.com CC videos\n\nhttps://github.com/DariusAf/MesoNet\nhttps://github.com/titu1994/keras-efficientnets \nhttps://keras.io/applications/\nhttps://www.tensorflow.org/api_docs/python/tf/keras/applications/\nhttps://github.com/qubvel/efficientnet\n\nhttps://github.com/KaiyangZhou/deep-person-reid\nhttps://github.com/rwightman/pytorch-image-models\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/pytorch/vision/tree/master/torchvision/models" }, { "id": 760274, "postDate": "2020-03-01T03:59:05.320Z", "content": "https://github.com/ipazc/mtcnn
\n\nhttps://github.com/albumentations-team/albumentations
", "rawMarkdown": "https://github.com/ipazc/mtcnn\n\nhttps://github.com/albumentations-team/albumentations" }, { "id": 732874, "postDate": "2020-01-30T12:08:56.300Z", "content": "EfficientNet implementation in PyTorch plus pretrained weights found here https://github.com/lukemelas/EfficientNet-PyTorch (Apache 2.0 license)
", "rawMarkdown": "EfficientNet implementation in PyTorch plus pretrained weights found here https://github.com/lukemelas/EfficientNet-PyTorch (Apache 2.0 license)", "votes": 2 }, { "id": 723979, "postDate": "2020-01-20T17:14:24.320Z", "content": "Dual Shot Face Detection - Pytorch Inference Code from https://github.com/hukkelas/DSFD-Pytorch-Inference (Apache 2.0 license)\nWiderFACE pretrained model from https://github.com/TencentYoutuResearch/FaceDetection-DSFD (Apache 2.0 license)
\n\n@inproceedings{li2018dsfd,\n title={DSFD: Dual Shot Face Detector},\n author={Li, Jian and Wang, Yabiao and Wang, Changan and Tai, Ying and Qian, Jianjun and Yang, Jian and Wang, Chengjie and Li, Jilin and Huang, Feiyue},\n booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},\n year={2019}\n}
", "rawMarkdown": "Dual Shot Face Detection - Pytorch Inference Code from https://github.com/hukkelas/DSFD-Pytorch-Inference (Apache 2.0 license)\nWiderFACE pretrained model from https://github.com/TencentYoutuResearch/FaceDetection-DSFD (Apache 2.0 license)\n\n@inproceedings{li2018dsfd,\n title={DSFD: Dual Shot Face Detector},\n author={Li, Jian and Wang, Yabiao and Wang, Changan and Tai, Ying and Qian, Jianjun and Yang, Jian and Wang, Chengjie and Li, Jilin and Huang, Feiyue},\n booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},\n year={2019}\n}", "votes": 2, "replies": [ { "id": 728400, "postDate": "2020-01-24T17:22:59.320Z", "content": "Also noteworthy is LightDSFD https://github.com/lijiannuist/lightDSFD
", "rawMarkdown": "Also noteworthy is LightDSFD https://github.com/lijiannuist/lightDSFD", "votes": 2 }, { "id": 732964, "postDate": "2020-01-30T13:47:42.893Z", "content": "I created a wheel package to install with pip:\nhttps://www.kaggle.com/soviet/dsfdpytorchinference
", "rawMarkdown": "I created a wheel package to install with pip:\nhttps://www.kaggle.com/soviet/dsfdpytorchinference", "votes": 2 } ] }, { "id": 723311, "postDate": "2020-01-19T20:35:53.563Z", "content": "not sure if this helps but there is the YouTube 8m challenge videos that are free to use and already on Kaggle.\nhttps://research.google.com/youtube8m/index.html
\n\ncurl data.yt8m.org/download.py | partition=3/frame/validate mirror=us python\ncurl data.yt8m.org/download.py | partition=3/frame/test mirror=us python
", "rawMarkdown": "not sure if this helps but there is the YouTube 8m challenge videos that are free to use and already on Kaggle.\nhttps://research.google.com/youtube8m/index.html\n\ncurl data.yt8m.org/download.py | partition=3/frame/validate mirror=us python\ncurl data.yt8m.org/download.py | partition=3/frame/test mirror=us python", "votes": 2, "replies": [ { "id": 728468, "postDate": "2020-01-24T18:52:12.643Z", "content": "And use this to pull actual video: https://github.com/gsssrao/youtube-8m-videos-frames
", "rawMarkdown": "And use this to pull actual video: https://github.com/gsssrao/youtube-8m-videos-frames", "votes": 1 } ] }, { "id": 722928, "postDate": "2020-01-19T10:15:58.280Z", "content": "May we use code and models from here: https://github.com/deepinsight/insightface/ ?\nTheir license states:
\n\n\n\n\nThe code of InsightFace is released under the MIT License. There is no limitation for both acadmic and commercial usage.
\n \nThe training data containing the annotation (and the models trained with these data) are available for non-commercial research purposes only.
\n
I'm not entirely sure if this challenge counts as non-commerical research.
", "rawMarkdown": "May we use code and models from here: https://github.com/deepinsight/insightface/ ?\nTheir license states:\n> The code of InsightFace is released under the MIT License. There is no limitation for both acadmic and commercial usage.\n>\n> The training data containing the annotation (and the models trained with these data) are available for non-commercial research purposes only.\n\nI'm not entirely sure if this challenge counts as non-commerical research.", "votes": 2, "replies": [ { "id": 727343, "postDate": "2020-01-23T16:37:15.350Z", "content": "I would also want to know the answer to this question.
", "rawMarkdown": "I would also want to know the answer to this question.", "isDeleted": true } ] }, { "id": 719539, "postDate": "2020-01-15T15:24:11Z", "content": "Pre-trained MTCNN model from here: https://github.com/timesler/facenet-pytorch\nimutils package: https://github.com/jrosebr1/imutils
", "rawMarkdown": "Pre-trained MTCNN model from here: https://github.com/timesler/facenet-pytorch\nimutils package: https://github.com/jrosebr1/imutils", "votes": 2 }, { "id": 705785, "postDate": "2019-12-29T13:20:11.673Z", "content": "MTCNN-package\nhttps://pypi.org/project/mtcnn/
", "rawMarkdown": "MTCNN-package\nhttps://pypi.org/project/mtcnn/", "votes": 2 }, { "id": 693452, "postDate": "2019-12-12T11:39:02.057Z", "content": "“External data is allowed up to 1 GB in size. External data must be freely & publicly available, including pre-trained models”
\n\nSince we must upload our locally trained/pre-trained models as a data source, and all data sources must be made publicly available, won't this make it so that everyone can simply always take the best current model and have everyone score mostly the same?
\n\nMaybe I am missing something.
", "rawMarkdown": "“External data is allowed up to 1 GB in size. External data must be freely & publicly available, including pre-trained models”\n\nSince we must upload our locally trained/pre-trained models as a data source, and all data sources must be made publicly available, won't this make it so that everyone can simply always take the best current model and have everyone score mostly the same? \n\nMaybe I am missing something.", "votes": 2, "replies": [ { "id": 693550, "postDate": "2019-12-12T13:55:36.580Z", "content": "\"Pre-trained models\" does not apply to your privately trained model. It applies only to other models that you, for example, used as starting point. You might take some face detection model from Model Zoo (of your choice) that is already pre-trained. In that case you have to disclose such model usage in this thread: https://www.kaggle.com/c/deepfake-detection-challenge/discussion/121203
", "rawMarkdown": "**\"Pre-trained models\" does not apply to your privately trained model.** It applies only to other models that you, for example, used as starting point. You might take some face detection model from Model Zoo (of your choice) that is already pre-trained. In that case you have to disclose such model usage in this thread: https://www.kaggle.com/c/deepfake-detection-challenge/discussion/121203 ", "votes": 2 }, { "id": 693874, "postDate": "2019-12-12T21:58:24.897Z", "content": "@simi2525 -- What @leanglider stated is correct. External data or pre-trained models you might use does not mean your entire model. For example, if you were to use imagenet - that would be considered external data that must be reported. But if you train your model using whatever techniques locally, let's say, and then load that trained model into Kaggle as an input into your submission notebook, you do not need to share that trained model. Just any external data or pre-trained models you may have used which contributed to the trained model.
", "rawMarkdown": "@simi2525 -- What @leanglider stated is correct. External data or pre-trained models you might use does not mean your entire model. For example, if you were to use imagenet - that would be considered external data that must be reported. But if you train your model using whatever techniques locally, let's say, and then load that trained model into Kaggle as an input into your submission notebook, you do not need to share that trained model. Just any external data or pre-trained models you may have used which contributed to the trained model. ", "votes": 2 }, { "id": 694702, "postDate": "2019-12-14T01:37:15.203Z", "content": "@leanglider @juliaelliott Thank you, I must have missed that part.
", "rawMarkdown": "@leanglider @juliaelliott Thank you, I must have missed that part." }, { "id": 712381, "postDate": "2020-01-07T07:01:54.423Z", "rawMarkdown": "" }, { "id": 717173, "postDate": "2020-01-12T20:28:03.380Z", "content": "Another detail:
\n\nSuppose, in a kernel run, I simply detect faces and output boxes representing these faces. Then, in subsequent kernel run (generating a submission), I input these boxes (as part of my 1 GByte allowed). I use these boxes to alleviate a great deal of CPU in my kernel run(purely for face detection). Is this allowed?
", "rawMarkdown": "Another detail:\n\nSuppose, in a kernel run, I simply detect faces and output boxes representing these faces. Then, in subsequent kernel run (generating a submission), I input these boxes (as part of my 1 GByte allowed). I use these boxes to alleviate a great deal of CPU in my kernel run(purely for face detection). Is this allowed?" }, { "id": 717190, "postDate": "2020-01-12T21:05:49.643Z", "content": "That wont help you much, even if possible, as the second stage is blind.
", "rawMarkdown": "That wont help you much, even if possible, as the second stage is blind." } ] }, { "id": 769124, "postDate": "2020-03-11T14:50:46.837Z", "content": "A python package to analyze and compare voices with deep learning\nhttps://github.com/resemble-ai/Resemblyzer\nhttps://www.kaggle.com/masterhyj/packagefile\nhttps://www.kaggle.com/sheldonrobinson/video-tools
", "rawMarkdown": "A python package to analyze and compare voices with deep learning\nhttps://github.com/resemble-ai/Resemblyzer\nhttps://www.kaggle.com/masterhyj/packagefile\nhttps://www.kaggle.com/sheldonrobinson/video-tools", "votes": 1 }, { "id": 759242, "postDate": "2020-02-28T19:17:16.817Z", "content": "https://pypi.org/, in case I didn't post it. :)
", "rawMarkdown": "https://pypi.org/, in case I didn't post it. :)", "votes": -1 }, { "id": 703168, "postDate": "2019-12-25T18:07:41.013Z", "content": "youtube.com has a nice dataset of videos
", "rawMarkdown": "youtube.com has a nice dataset of videos", "votes": 1 }, { "id": 761667, "postDate": "2020-03-02T20:39:51.413Z", "content": "All videos from https://youtube.com avialable at this moment.
", "rawMarkdown": "All videos from https://youtube.com avialable at this moment.", "votes": -1 }, { "id": 756837, "postDate": "2020-02-26T06:09:04.353Z", "content": "https://github.com/moabitcoin/ig65m-pytorch\nNetworks and weights in github
\n\neffecientnet ( https://github.com/lukemelas/EfficientNet-PyTorch ) - Google weights\n'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b0-355c32eb.pth',\n 'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b1-f1951068.pth',\n 'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b2-8bb594d6.pth',\n 'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b3-5fb5a3c3.pth',\n 'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b4-6ed6700e.pth',\n 'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b5-b6417697.pth',\n 'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b6-c76e70fd.pth',\n 'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b7-dcc49843.pth'
\n\n'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b0-b64d5a18.pth',\n 'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b1-0f3ce85a.pth',\n 'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b2-6e9d97e5.pth',\n 'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b3-cdd7c0f4.pth',\n 'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b4-44fb3a87.pth',\n 'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b5-86493f6b.pth',\n 'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b6-ac80338e.pth',\n 'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b7-4652b6dd.pth',\n 'efficientnet-b8': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b8-22a8fe65.pth'
\n\nmtccn\n*pytroch model zoo*\nopencv2\n*scikit learn*\nfastai
\n\nEfficientNet-pyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nresnext101: https://github.com/facebookresearch/WSL-Images\nface.evoLVe.PyTorch: https://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nResidualAttentionNetwork-pytorch: https://github.com/tengshaofeng/ResidualAttentionNetwork-pytorch\nhrnet: https://github.com/HRNet/HRNet-Image-Classification\ntorchvision models: https://github.com/pytorch/vision/tree/master/torchvision/models\nFaceForensics++ dataset: https://github.com/ondyari/FaceForensics\ntrainmsra.tar.gz traincelebrity.tar.gz: http://trillionpairs.deepglint.com/data
\n\nwider face datasets: http://shuoyang1213.me/WIDERFACE/\nresnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n'resnext5032x4d': 'https://download.pytorch.org/models/resnext5032x4d-7cdf4587.pth',\n'resnext10132x8d': 'https://download.pytorch.org/models/resnext10132x8d-8ba56ff5.pth',\n'wideresnet502': 'https://download.pytorch.org/models/wideresnet502-95faca4d.pth',\n'wideresnet1012': 'https://download.pytorch.org/models/wideresnet1012-32ee1156.pth',
\n\nhttps://www.kaggle.com/tunguz/70000-real-faces-1\nhttps://www.kaggle.com/tunguz/70000-real-faces-2\nhttps://www.kaggle.com/tunguz/70000-real-faces-3\nhttps://www.kaggle.com/tunguz/70000-real-faces-4\nhttps://www.kaggle.com/tunguz/70000-real-faces-5\nhttps://www.kaggle.com/tunguz/70000-real-faces-6
\n\nmmdetection: https://github.com/open-mmlab/mmdetection
\n\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch
\n\nhttps://github.com/EndlessSora/DeeperForensics-1.0\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/
\n\nUADFV: The UADFV dataset - Xin Yang, Yuezun Li, and Siwei Lyu. Exposing deep fakes\nusing inconsistent head poses. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),\n2019.
\n\nDF-TIMIT: Pavel Korshunov and Sebastien Marcel. Deepfakes: a new ´\nthreat to face recognition? assessment and detection. arXiv\npreprint arXiv:1812.08685, 2018.
\n\nFF-DF: Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Chris- ¨\ntian Riess, Justus Thies, and Matthias Nießner. FaceForensics++: Learning to detect manipulated facial images. In\nICCV, 2019.
\n\nDFD: Nicholas Dufour, Andrew Gully, Per Karlsson, Alexey Victor Vorbyov, Thomas Leung, Jeremiah Childs, and Christoph Bregler. Deepfakes detection dataset by google & jigsaw.
\n\nhttps://github.com/dessa-research/DeepFake-Detection
", "rawMarkdown": "**https://github.com/moabitcoin/ig65m-pytorch**\nNetworks and weights in github\n\n\n**effecientnet ( https://github.com/lukemelas/EfficientNet-PyTorch ) - Google weights**\n'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b0-355c32eb.pth',\n 'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b1-f1951068.pth',\n 'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b2-8bb594d6.pth',\n 'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b3-5fb5a3c3.pth',\n 'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b4-6ed6700e.pth',\n 'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b5-b6417697.pth',\n 'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b6-c76e70fd.pth',\n 'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b7-dcc49843.pth'\n\n 'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b0-b64d5a18.pth',\n 'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b1-0f3ce85a.pth',\n 'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b2-6e9d97e5.pth',\n 'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b3-cdd7c0f4.pth',\n 'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b4-44fb3a87.pth',\n 'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b5-86493f6b.pth',\n 'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b6-ac80338e.pth',\n 'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b7-4652b6dd.pth',\n 'efficientnet-b8': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b8-22a8fe65.pth'\n\n**mtccn**\n**pytroch model zoo**\n**opencv2**\n**scikit learn**\n**fastai**\n\nEfficientNet-pyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nresnext101: https://github.com/facebookresearch/WSL-Images\nface.evoLVe.PyTorch: https://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nResidualAttentionNetwork-pytorch: https://github.com/tengshaofeng/ResidualAttentionNetwork-pytorch\nhrnet: https://github.com/HRNet/HRNet-Image-Classification\ntorchvision models: https://github.com/pytorch/vision/tree/master/torchvision/models\nFaceForensics++ dataset: https://github.com/ondyari/FaceForensics\ntrainmsra.tar.gz traincelebrity.tar.gz: http://trillionpairs.deepglint.com/data\n\n\nwider face datasets: http://shuoyang1213.me/WIDERFACE/\nresnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n'resnext5032x4d': 'https://download.pytorch.org/models/resnext5032x4d-7cdf4587.pth',\n'resnext10132x8d': 'https://download.pytorch.org/models/resnext10132x8d-8ba56ff5.pth',\n'wideresnet502': 'https://download.pytorch.org/models/wideresnet502-95faca4d.pth',\n'wideresnet1012': 'https://download.pytorch.org/models/wideresnet1012-32ee1156.pth',\n\n\n\n\nhttps://www.kaggle.com/tunguz/70000-real-faces-1\nhttps://www.kaggle.com/tunguz/70000-real-faces-2\nhttps://www.kaggle.com/tunguz/70000-real-faces-3\nhttps://www.kaggle.com/tunguz/70000-real-faces-4\nhttps://www.kaggle.com/tunguz/70000-real-faces-5\nhttps://www.kaggle.com/tunguz/70000-real-faces-6\n\n\nmmdetection: https://github.com/open-mmlab/mmdetection\n\n\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\n\n\nhttps://github.com/EndlessSora/DeeperForensics-1.0\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/\n\n\nUADFV: The UADFV dataset - Xin Yang, Yuezun Li, and Siwei Lyu. Exposing deep fakes\nusing inconsistent head poses. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),\n2019.\n\nDF-TIMIT: Pavel Korshunov and Sebastien Marcel. Deepfakes: a new ´\nthreat to face recognition? assessment and detection. arXiv\npreprint arXiv:1812.08685, 2018.\n\nFF-DF: Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Chris- ¨\ntian Riess, Justus Thies, and Matthias Nießner. FaceForensics++: Learning to detect manipulated facial images. In\nICCV, 2019.\n\nDFD: Nicholas Dufour, Andrew Gully, Per Karlsson, Alexey Victor Vorbyov, Thomas Leung, Jeremiah Childs, and Christoph Bregler. Deepfakes detection dataset by google & jigsaw.\n\nhttps://github.com/dessa-research/DeepFake-Detection", "votes": -1 }, { "id": 823973, "postDate": "2020-04-28T04:22:42.973Z", "content": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/facebookresearch/SlowFast\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/elliottzheng/face-detection
", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/facebookresearch/SlowFast\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/elliottzheng/face-detection", "replies": [ { "id": 826858, "postDate": "2020-04-29T22:51:43.620Z", "content": "isn't that after the deadline hence this would be a DQ?
", "rawMarkdown": "isn't that after the deadline hence this would be a DQ?" }, { "id": 826891, "postDate": "2020-04-29T23:29:18.617Z", "content": "\n\n\nOnce it has been posted, you do not need to post it again.
\n
The repo I used has already been posted by other people and the face-detection repo is created by myself and I made it open after the competition. I comment here just to share with other people.
", "rawMarkdown": "> Once it has been posted, you do not need to post it again.\n\nThe repo I used has already been posted by other people and the face-detection repo is created by myself and I made it open after the competition. I comment here just to share with other people." } ] }, { "id": 792237, "postDate": "2020-03-31T02:21:51.710Z", "content": "", "rawMarkdown": "https://www.kaggle.com/humananalog/deepfakes-inference-demo/download/aioBap8W5sTGGJr16faR%2Fversions%2FvTJB32h13oA5mti8DfTI%2Ffiles%2Fresnext.pth?datasetVersionNumber=1" }, { "id": 791438, "postDate": "2020-03-30T12:25:48.600Z", "content": "https://github.com/rwightman/pytorch-image-models\nhttps://github.com/danmohaha/celeb-deepfakeforensics\nhttps://github.com/ondyari/FaceForensics\nhttp://mmlab.ie.cuhk.edu.hk/projects/CelebA.html
", "rawMarkdown": "https://github.com/rwightman/pytorch-image-models\nhttps://github.com/danmohaha/celeb-deepfakeforensics\nhttps://github.com/ondyari/FaceForensics\nhttp://mmlab.ie.cuhk.edu.hk/projects/CelebA.html" }, { "id": 790959, "postDate": "2020-03-30T01:46:11.507Z", "content": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/pytorch/vision/tree/master/torchvision/models\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/hollance/BlazeFace-PyTorch\nhttps://github.com/albumentations-team/albumentations\nhttps://github.com/nerox8664/pytorch2keras\nhttps://github.com/danmohaha/celeb-deepfakeforensics
", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/pytorch/vision/tree/master/torchvision/models\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/hollance/BlazeFace-PyTorch\nhttps://github.com/albumentations-team/albumentations\nhttps://github.com/nerox8664/pytorch2keras\nhttps://github.com/danmohaha/celeb-deepfakeforensics" }, { "id": 788654, "postDate": "2020-03-27T21:00:20.390Z", "content": "http://parnec.nuaa.edu.cn/xtan/data/ClosedEyeDatabases.html\ndata set of closed eyes
", "rawMarkdown": "http://parnec.nuaa.edu.cn/xtan/data/ClosedEyeDatabases.html\ndata set of closed eyes" }, { "id": 784766, "postDate": "2020-03-24T13:46:09.753Z", "content": "https://github.com/cleardusk/MeGlass
", "rawMarkdown": "https://github.com/cleardusk/MeGlass" }, { "id": 780671, "postDate": "2020-03-20T14:02:14.180Z", "rawMarkdown": "" }, { "id": 780440, "postDate": "2020-03-20T09:14:46.300Z", "content": "mtcnn\nhttps://www.kaggle.com/unkownhihi/mtcnn-package\nhttps://github.com/ipazc/mtcnn
", "rawMarkdown": "mtcnn\nhttps://www.kaggle.com/unkownhihi/mtcnn-package\nhttps://github.com/ipazc/mtcnn" }, { "id": 776871, "postDate": "2020-03-17T17:33:57.170Z", "content": "http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttp://megaface.cs.washington.edu
", "rawMarkdown": "http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttp://megaface.cs.washington.edu" }, { "id": 772680, "postDate": "2020-03-15T19:01:12.953Z", "content": "face_recognition https://github.com/ageitgey/face_recognition\nvgg-19 http://www.vlfeat.org/matconvnet/models/beta16/imagenet-vgg-verydeep-19.mat\nhttps://github.com/shekkizh/FCN.tensorflow
", "rawMarkdown": "face_recognition https://github.com/ageitgey/face_recognition\nvgg-19 http://www.vlfeat.org/matconvnet/models/beta16/imagenet-vgg-verydeep-19.mat\nhttps://github.com/shekkizh/FCN.tensorflow\n" }, { "id": 770519, "postDate": "2020-03-13T03:30:26.840Z", "content": "I found a landmark pretrained model today, and I rewrite the postprocess code. Does this case also need to be declared?\nhttps://github.com/justusschock/shapenet
", "rawMarkdown": "I found a landmark pretrained model today, and I rewrite the postprocess code. Does this case also need to be declared?\nhttps://github.com/justusschock/shapenet" }, { "id": 768928, "postDate": "2020-03-11T11:08:28.543Z", "content": "https://github.com/tensorflow/models/\nhttps://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md\nhttps://github.com/yeephycho/tensorflow-face-detection\nhttps://github.com/EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10
", "rawMarkdown": "https://github.com/tensorflow/models/\nhttps://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md\nhttps://github.com/yeephycho/tensorflow-face-detection\nhttps://github.com/EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10" }, { "id": 768701, "postDate": "2020-03-11T05:19:20.367Z", "content": "Using Tensorflow 2 latest
", "rawMarkdown": "Using Tensorflow 2 latest" }, { "id": 768482, "postDate": "2020-03-10T21:24:54.230Z", "content": "https://github.com/kpzhang93/MTCNN_face_detection_alignment
\n\nhttps://github.com/davisking/dlib\nhttp://dlib.net/files/mmod_human_face_detector.dat.bz2\nhttps://www.kaggle.com/pedromoya/dlibpackage
\n\nhttps://github.com/weiliu89/caffe/tree/ssd\nhttps://github.com/opencv/opencv/tree/master/samples/dnn/face_detector
\n\n\n\nhttp://dlib.net/files/data/dlib_face_detector_training_data.tar.gz\nhttp://dlib.net/files/data/dlib_faces_5points.tar\nhttp://dlib.net/files/data/dlib_face_detector_training_data.tar.gz
\n\nhttps://github.com/blaueck/tf-mtcnn\nhttps://www.kaggle.com/pedromoya/mtcmtflib
\n\nfaceRecognition\nhttps://github.com/ydwen/caffe-face
\n\nhearth-rate\nhttps://www.kaggle.com/pedromoya/hearthrate
", "rawMarkdown": "https://github.com/kpzhang93/MTCNN_face_detection_alignment\n\nhttps://github.com/davisking/dlib\nhttp://dlib.net/files/mmod_human_face_detector.dat.bz2\nhttps://www.kaggle.com/pedromoya/dlibpackage\n\nhttps://github.com/weiliu89/caffe/tree/ssd\nhttps://github.com/opencv/opencv/tree/master/samples/dnn/face_detector\n\nhttps://github.com/nhatthai/opencv-face-recognition/blob/master/src/face_detection_model/res10_300x300_ssd_iter_140000.caffemodel\n\nhttp://dlib.net/files/data/dlib_face_detector_training_data.tar.gz\nhttp://dlib.net/files/data/dlib_faces_5points.tar\nhttp://dlib.net/files/data/dlib_face_detector_training_data.tar.gz\n\nhttps://github.com/blaueck/tf-mtcnn\nhttps://www.kaggle.com/pedromoya/mtcmtflib\n\nfaceRecognition\nhttps://github.com/ydwen/caffe-face\n\nhearth-rate\nhttps://www.kaggle.com/pedromoya/hearthrate" }, { "id": 768119, "postDate": "2020-03-10T13:37:56.950Z", "content": "imagenet pretrained models resnet,xception ,resnext .. not sure if i will use anything in future ..
", "rawMarkdown": "imagenet pretrained models resnet,xception ,resnext .. not sure if i will use anything in future .." }, { "id": 767031, "postDate": "2020-03-09T04:09:52.210Z", "content": "Test dataset from https://github.com/PeterWang512/CNNDetection
", "rawMarkdown": "Test dataset from https://github.com/PeterWang512/CNNDetection\n" }, { "id": 766490, "postDate": "2020-03-08T08:36:11.447Z", "content": "https://github.com/open-mmlab/mmaction
", "rawMarkdown": "https://github.com/open-mmlab/mmaction" }, { "id": 766360, "postDate": "2020-03-08T03:49:38.140Z", "content": "Where it says that we have to disclose pre-trained models, does it mean all pre-trained models, including our original model that we trained outside of Kaggle and then uploaded to Kaggle (as essentially all models are, since you can't really train your models on Kaggle) or is it only referring to open source models that others have built (and possibly pre-trained, so we can use transfer-learning), and we are burrowing to incorporate in our own model?
\n\nThanks in advance.
", "rawMarkdown": "Where it says that we have to disclose pre-trained models, does it mean all pre-trained models, including our original model that we trained outside of Kaggle and then uploaded to Kaggle (as essentially all models are, since you can't really train your models on Kaggle) or is it only referring to open source models that others have built (and possibly pre-trained, so we can use transfer-learning), and we are burrowing to incorporate in our own model?\n\nThanks in advance.", "replies": [ { "id": 767505, "postDate": "2020-03-09T18:11:06.933Z", "content": "@rafwaf0101 It does not include your trained model. But it should include open-source pre-trained models, if you used any that are not your original work.
", "rawMarkdown": "@rafwaf0101 It does not include your trained model. But it should include open-source pre-trained models, if you used any that are not your original work. ", "votes": 1 }, { "id": 772029, "postDate": "2020-03-15T00:00:37.370Z", "content": "Thanks
", "rawMarkdown": "Thanks" } ] }, { "id": 764307, "postDate": "2020-03-05T10:20:37.790Z", "content": "All of freely available debian deb packages and shared libs.
", "rawMarkdown": "All of freely available debian deb packages and shared libs." }, { "id": 763911, "postDate": "2020-03-05T01:01:06.137Z", "content": "https://github.com/yxlijun/S3FD.pytorch\nhttps://www.kaggle.com/pedromoya/s3fdpytorch\nhttps://www.kaggle.com/pedromoya/mtcnnbib\nhttps://www.kaggle.com/pedromoya/mtcnnmtcnn
", "rawMarkdown": "https://github.com/yxlijun/S3FD.pytorch\nhttps://www.kaggle.com/pedromoya/s3fdpytorch\nhttps://www.kaggle.com/pedromoya/mtcnnbib\nhttps://www.kaggle.com/pedromoya/mtcnnmtcnn" }, { "id": 763841, "postDate": "2020-03-04T22:30:38.743Z", "content": "https://github.com/pytorch/vision/blob/master/torchvision/models\npytorch Resnext\n 'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n 'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n 'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n 'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n 'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n 'resnext50_32x4d': 'https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pth',\n 'resnext101_32x8d': 'https://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pth',\n 'wide_resnet50_2': 'https://download.pytorch.org/models/wide_resnet50_2-95faca4d.pth',\n 'wide_resnet101_2': 'https://download.pytorch.org/models/wide_resnet101_2-32ee1156.pth'\npytorch inception\n'inception_v3_google': 'https://download.pytorch.org/models/inception_v3_google-1a9a5a14.pth'
", "rawMarkdown": "https://github.com/pytorch/vision/blob/master/torchvision/models\npytorch Resnext\n 'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n 'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n 'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n 'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n 'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n 'resnext50_32x4d': 'https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pth',\n 'resnext101_32x8d': 'https://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pth',\n 'wide_resnet50_2': 'https://download.pytorch.org/models/wide_resnet50_2-95faca4d.pth',\n 'wide_resnet101_2': 'https://download.pytorch.org/models/wide_resnet101_2-32ee1156.pth'\npytorch inception\n'inception_v3_google': 'https://download.pytorch.org/models/inception_v3_google-1a9a5a14.pth'" }, { "id": 763804, "postDate": "2020-03-04T21:59:42.457Z", "content": "https://github.com/shijianjian/EfficientNet-PyTorch-3D
", "rawMarkdown": "https://github.com/shijianjian/EfficientNet-PyTorch-3D" }, { "id": 763701, "postDate": "2020-03-04T18:58:05.020Z", "content": "https://www.kaggle.com/humananalog/blazeface-pytorch\nhttps://github.com/timesler/facenet-pytorch
", "rawMarkdown": "https://www.kaggle.com/humananalog/blazeface-pytorch\nhttps://github.com/timesler/facenet-pytorch\n" }, { "id": 763256, "postDate": "2020-03-04T09:52:39.797Z", "content": "Some video classification project\nhttps://github.com/HHTseng/video-classification\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/boyaolin/Video-Classification
", "rawMarkdown": "Some video classification project\nhttps://github.com/HHTseng/video-classification\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/boyaolin/Video-Classification" }, { "id": 762986, "postDate": "2020-03-04T01:57:30.093Z", "content": "https://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://pypi.org/project/mtcnn/\nhttps://pypi.org/project/python_speech_features/0.4/\nhttps://pypi.org/project/pytorchcv/\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/ageitgey/face_recognition_models\nwww.youtube.com with permitted license
", "rawMarkdown": "https://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://pypi.org/project/mtcnn/\nhttps://pypi.org/project/python_speech_features/0.4/\nhttps://pypi.org/project/pytorchcv/\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/ageitgey/face_recognition_models\nwww.youtube.com with permitted license\n" }, { "id": 762922, "postDate": "2020-03-03T23:58:25.580Z", "content": "Resemblyzer: https://github.com/resemble-ai/Resemblyzer\nmmdetection: https://github.com/open-mmlab/mmdetection\nDual Shot Face Detector\nhttps://keithito.com/LJ-Speech-Dataset/\nhttps://github.com/thiago1080/SphereFace\nVoice Datasets: https://github.com/jim-schwoebel/voice_datasets (Only ones that comply with the rules, dataset in this list with any restrictions will not be used).\nCommonVoice:https://www.kaggle.com/mozillaorg/common-voice/home\nhttps://github.com/nyoki-mtl/keras-facenet\nhttps://github.com/onnx/onnx\nTFFD: https://github.com/yeephycho/tensorflow-face-detection code not training data used (which is not allowed)\nHelen: http://www.ifp.illinois.edu/~vuongle2/helen/\nSFEW: http://cs.anu.edu.au/few\nFacenet: https://github.com/davidsandberg/facenet\nhttps://www.kaggle.com/selfishgene/youtube-faces-with-facial-keypoints\nFace Detection in Images: https://www.kaggle.com/dataturks/face-detection-in-images\nYoutube Faces Dataset: https://www.cs.tau.ac.il/~wolf/ytfaces/\nhttps://research.google.com/audioset/download.html\nFace Detection Data Set & Benchmark http://vis-www.cs.umass.edu/fddb/\nFaces in the wild dataset: http://tamaraberg.com/faceDataset/index.html\nFlickr Faces HQ Dataset: https://github.com/NVlabs/ffhq-dataset\nTufts Face Database: https://www.kaggle.com/kpvisionlab/tufts-face-database\nReal and Fake Face Detection: https://www.kaggle.com/ciplab/real-and-fake-face-detection\nGoogle Facial Expression: https://research.google/tools/datasets/google-facial-expression/\nhttps://www.kaggle.com/drgilermo/face-images-with-marked-landmark-points\nhttps://github.com/thiago1080/SphereFace\nhttps://ai.googleblog.com/2017/08/launching-speech-commands-dataset.html\nLibrispeech: https://www.openslr.org/12\nhttps://homepages.inf.ed.ac.uk/jyamagis/page3/page58/page58.html\nhttps://github.com/ageitgey/face_recognition\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nhttps://github.com/iitzco/faced\nhttps://github.com/qubvel/efficientnet\nhttps://github.com/Star-Clouds/CenterFace\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/NVlabs/ffhq-dataset\nhttps://github.com/matterport/Mask_RCNN\nhttps://github.com/deepfakes/faceswap\nhttps://www.youtube.com/ Creative Commons\nEverything a :https://keras.io/applications/ and\nhttps://github.com/keras-team/keras-applications and any of their pre-trained models and weights that comply with licensing rules of competition.\nhttps://github.com/ageitgey/face_recognition\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/fyr91/facedetection https://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices/tree/master/facedetection\nSome architectures and pretrained models and datasets:\nhttp://www.robots.ox.ac.uk/~vgg/data/\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/ondyari/FaceForensics\nhttp://www.robots.ox.ac.uk/~vgg/data/lipreading/ https://github.com/astorfi/lip-reading-deeplearning https://github.com/joseph-zhong/LipReading https://github.com/afourast/deeplip_reading\nhttps://github.com/hassanhub/LipReading\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/qubvel/efficientnet\nhttps://github.com/qubvel/classification_models\nhttps://github.com/1adrianb/face-alignment\nhttps://www.kaggle.com/humananalog/deepfakes-inference-demo
\n\nDisclaimer:\nDatasets I post may be used in the final submission. They will be thoroughly vetted before final submission and any ones that do not comply with rules will not be used in the final submission, subject to your further review during verification.\nThank you.
", "rawMarkdown": "Resemblyzer: https://github.com/resemble-ai/Resemblyzer\nmmdetection: https://github.com/open-mmlab/mmdetection\nDual Shot Face Detector\nhttps://keithito.com/LJ-Speech-Dataset/\nhttps://github.com/thiago1080/SphereFace\nVoice Datasets: https://github.com/jim-schwoebel/voice_datasets (Only ones that comply with the rules, dataset in this list with any restrictions will not be used).\nCommonVoice:https://www.kaggle.com/mozillaorg/common-voice/home\nhttps://github.com/nyoki-mtl/keras-facenet\nhttps://github.com/onnx/onnx\nTFFD: https://github.com/yeephycho/tensorflow-face-detection code not training data used (which is not allowed)\nHelen: http://www.ifp.illinois.edu/~vuongle2/helen/\nSFEW: http://cs.anu.edu.au/few\nFacenet: https://github.com/davidsandberg/facenet\nhttps://www.kaggle.com/selfishgene/youtube-faces-with-facial-keypoints\nFace Detection in Images: https://www.kaggle.com/dataturks/face-detection-in-images\nYoutube Faces Dataset: https://www.cs.tau.ac.il/~wolf/ytfaces/\nhttps://research.google.com/audioset/download.html\nFace Detection Data Set & Benchmark http://vis-www.cs.umass.edu/fddb/\nFaces in the wild dataset: http://tamaraberg.com/faceDataset/index.html\nFlickr Faces HQ Dataset: https://github.com/NVlabs/ffhq-dataset\nTufts Face Database: https://www.kaggle.com/kpvisionlab/tufts-face-database\nReal and Fake Face Detection: https://www.kaggle.com/ciplab/real-and-fake-face-detection\nGoogle Facial Expression: https://research.google/tools/datasets/google-facial-expression/\nhttps://www.kaggle.com/drgilermo/face-images-with-marked-landmark-points\nhttps://github.com/thiago1080/SphereFace\nhttps://ai.googleblog.com/2017/08/launching-speech-commands-dataset.html\nLibrispeech: https://www.openslr.org/12\nhttps://homepages.inf.ed.ac.uk/jyamagis/page3/page58/page58.html\nhttps://github.com/ageitgey/face_recognition\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nhttps://github.com/iitzco/faced\nhttps://github.com/qubvel/efficientnet\nhttps://github.com/Star-Clouds/CenterFace\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/NVlabs/ffhq-dataset\nhttps://github.com/matterport/Mask_RCNN\nhttps://github.com/deepfakes/faceswap\nhttps://www.youtube.com/ Creative Commons\nEverything a :https://keras.io/applications/ and\nhttps://github.com/keras-team/keras-applications and any of their pre-trained models and weights that comply with licensing rules of competition.\nhttps://github.com/ageitgey/face_recognition\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/fyr91/facedetection https://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices/tree/master/facedetection\nSome architectures and pretrained models and datasets:\nhttp://www.robots.ox.ac.uk/~vgg/data/\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/ondyari/FaceForensics\nhttp://www.robots.ox.ac.uk/~vgg/data/lipreading/ https://github.com/astorfi/lip-reading-deeplearning https://github.com/joseph-zhong/LipReading https://github.com/afourast/deeplip_reading\nhttps://github.com/hassanhub/LipReading\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/qubvel/efficientnet\nhttps://github.com/qubvel/classification_models\nhttps://github.com/1adrianb/face-alignment\nhttps://www.kaggle.com/humananalog/deepfakes-inference-demo\n \nDisclaimer:\nDatasets I post may be used in the final submission. They will be thoroughly vetted before final submission and any ones that do not comply with rules will not be used in the final submission, subject to your further review during verification.\nThank you." }, { "id": 762919, "postDate": "2020-03-03T23:56:54.580Z", "content": "pip install ffmpeg-python\nconda install -c conda-forge ffmpeg\npip install librosa\nTensorflow 1.*
", "rawMarkdown": "pip install ffmpeg-python\nconda install -c conda-forge ffmpeg\npip install librosa\nTensorflow 1.*" }, { "id": 762905, "postDate": "2020-03-03T23:39:24.733Z", "content": "1. https://github.com/EndlessSora/DeeperForensics-1.0\n2. https://github.com/TencentYoutuResearch/FaceDetection-DSFD\n3. https://github.com/iitzco/faced\n4. https://github.com/matterport/Mask_RCNN\n5. https://github.com/timesler/facenet-pytorch\n6. https://github.com/yeephycho/tensorflow-face-detection\n7. https://github.com/qubvel/efficientnet\n8. https://github.com/ipazc/mtcnn\n9. https://github.com/ondyari/FaceForensics\n10. https://github.com/Star-Clouds/CenterFace\n11. http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1.html\n12. http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox2.html\n13. https://github.com/NVlabs/ffhq-dataset\n14. https://github.com/resemble-ai/Resemblyzer\n15. https://github.com/thiago1080/SphereFace\n16. https://github.com/HRNet/HRNet-Image-Classification\n17. https://github.com/deepfakes/faceswap\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/resemble-ai/Resemblyzer\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/opencv/opencv_contrib\nhttps://pypi.org/project/tables/\nhttps://pypi.org/project/h5py/\nhttps://pypi.org/project/pandas/\nhttps://github.com/ondyari/FaceForensics/\nmtccn\npytorch model zoo\nopencv2\nscikit learn\nfastai\nskvideo\nffmpeg-python\nhttps://www.kaggle.com/rakibilly/ffmpeg-static-build\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/mnikitin/EfficientNet\nhttps://pypi.org/project/librosa/\nhttps://www.kaggle.com/ciplab/real-and-fake-face-detection
", "rawMarkdown": "https://github.com/timesler/facenet-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/resemble-ai/Resemblyzer\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/opencv/opencv_contrib\nhttps://pypi.org/project/tables/\nhttps://pypi.org/project/h5py/\nhttps://pypi.org/project/pandas/\nhttps://github.com/ondyari/FaceForensics/\nmtccn\npytorch model zoo\nopencv2\nscikit learn\nfastai\nskvideo\nffmpeg-python\nhttps://www.kaggle.com/rakibilly/ffmpeg-static-build\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/mnikitin/EfficientNet\nhttps://pypi.org/project/librosa/\nhttps://www.kaggle.com/ciplab/real-and-fake-face-detection" }, { "id": 762875, "postDate": "2020-03-03T22:13:30.223Z", "content": "https://github.com/ondyari/FaceForensics\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/hollance/BlazeFace-PyTorch\nhttps://github.com/thiago1080/SphereFace\nhttps://github.com/vlad3996/FaceDetection-DSFD\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://johnvansickle.com/ffmpeg/\nhttps://www.kaggle.com/rakibilly/ffmpeg-static-build\nhttps://github.com/astorfi/speechpy
\n\nPretrained models from:\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttps://github.com/fchollet/deep-learning-models
", "rawMarkdown": "https://github.com/ondyari/FaceForensics\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/hollance/BlazeFace-PyTorch\nhttps://github.com/thiago1080/SphereFace\nhttps://github.com/vlad3996/FaceDetection-DSFD\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://johnvansickle.com/ffmpeg/\nhttps://www.kaggle.com/rakibilly/ffmpeg-static-build\nhttps://github.com/astorfi/speechpy\n\nPretrained models from:\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttps://github.com/fchollet/deep-learning-models\n" }, { "id": 762862, "postDate": "2020-03-03T21:59:35.143Z", "content": "https://github.com/ondyari/FaceForensics\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/hollance/BlazeFace-PyTorch
", "rawMarkdown": "https://github.com/ondyari/FaceForensics\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/hollance/BlazeFace-PyTorch\n" }, { "id": 762853, "postDate": "2020-03-03T21:40:19.220Z", "content": "https://github.com/opencv/opencv/blob/master/data/haarcascades/haarcascade_frontalface_default.xml
", "rawMarkdown": "https://github.com/opencv/opencv/blob/master/data/haarcascades/haarcascade_frontalface_default.xml" }, { "id": 762850, "postDate": "2020-03-03T21:29:59.770Z", "content": "https://github.com/sampepose/flownet2-tf
", "rawMarkdown": "https://github.com/sampepose/flownet2-tf" }, { "id": 762847, "postDate": "2020-03-03T21:17:38.373Z", "content": "https://datashare.is.ed.ac.uk/handle/10283/3336
", "rawMarkdown": "https://datashare.is.ed.ac.uk/handle/10283/3336" }, { "id": 762841, "postDate": "2020-03-03T21:12:12.317Z", "content": "https://pytorch.org/hub/research-models\nhttps://github.com/pytorch/examples\nhttps://github.com/taki0112/SPADE-Tensorflow\nhttps://github.com/ajbrock/BigGAN-PyTorch\nhttps://github.com/rosinality/style-based-gan-pytorch\nhttps://github.com/junyanz/pytorch-CycleGAN-and-pix2pix\nhttps://github.com/yunjey/stargan\nhttps://github.com/WojciechMormul/crn\nhttps://github.com/johnathanlouie/crn\nhttps://github.com/zth667/Diverse-Image-Synthesis-from-Semantic-Layout\nhttps://github.com/KupynOrest/DeblurGAN\nhttps://github.com/TAMU-VITA/DeblurGANv2\nhttp://cchen156.web.engr.illinois.edu/SID.html\nhttps://github.com/daitao/SAN\nhttps://modelzoo.co/
", "rawMarkdown": "https://pytorch.org/hub/research-models\nhttps://github.com/pytorch/examples\nhttps://github.com/taki0112/SPADE-Tensorflow\nhttps://github.com/ajbrock/BigGAN-PyTorch\nhttps://github.com/rosinality/style-based-gan-pytorch\nhttps://github.com/junyanz/pytorch-CycleGAN-and-pix2pix\nhttps://github.com/yunjey/stargan\nhttps://github.com/WojciechMormul/crn\nhttps://github.com/johnathanlouie/crn\nhttps://github.com/zth667/Diverse-Image-Synthesis-from-Semantic-Layout\nhttps://github.com/KupynOrest/DeblurGAN\nhttps://github.com/TAMU-VITA/DeblurGANv2\nhttp://cchen156.web.engr.illinois.edu/SID.html\nhttps://github.com/daitao/SAN\nhttps://modelzoo.co/\n" }, { "id": 762833, "postDate": "2020-03-03T20:57:29.867Z", "content": "MesoNet\nhttps://github.com/DariusAf/MesoNet\nFaceForensics code, no dataset\nhttps://github.com/ondyari/FaceForensics\nMulti-task\nhttps://github.com/nii-yamagishilab/ClassNSeg\nCapsule\nhttps://github.com/nii-yamagishilab/Capsule-Forensics-v2\npretrainedmodels\nhttps://github.com/cadene/pretrained-models.pytorch\nhttps://data.lip6.fr/cadene/pretrainedmodels/\ntfhub models \nhttps://tfhub.dev/\ndlib\nhttps://github.com/davisking/dlib\nface_recognition\nhttps://github.com/ageitgey/face_recognition\nmtcnn\nhttps://github.com/ipazc/mtcnn\nfacenet-pytorch\nhttps://github.com/timesler/facenet-pytorch\nBlazeFace-PyTorch\nhttps://github.com/hollance/BlazeFace-PyTorch\nffmpeg\nhttps://johnvansickle.com/ffmpeg/\nffmpeg-python\nhttps://pypi.org/project/ffmpeg-python/#files
", "rawMarkdown": "MesoNet\nhttps://github.com/DariusAf/MesoNet\nFaceForensics code, no dataset\nhttps://github.com/ondyari/FaceForensics\nMulti-task\nhttps://github.com/nii-yamagishilab/ClassNSeg\nCapsule\nhttps://github.com/nii-yamagishilab/Capsule-Forensics-v2\npretrainedmodels\nhttps://github.com/cadene/pretrained-models.pytorch\nhttps://data.lip6.fr/cadene/pretrainedmodels/\ntfhub models \nhttps://tfhub.dev/\ndlib\nhttps://github.com/davisking/dlib\nface_recognition\nhttps://github.com/ageitgey/face_recognition\nmtcnn\nhttps://github.com/ipazc/mtcnn\nfacenet-pytorch\nhttps://github.com/timesler/facenet-pytorch\nBlazeFace-PyTorch\nhttps://github.com/hollance/BlazeFace-PyTorch\nffmpeg\nhttps://johnvansickle.com/ffmpeg/\nffmpeg-python\nhttps://pypi.org/project/ffmpeg-python/#files" }, { "id": 762824, "postDate": "2020-03-03T20:46:18.437Z", "content": "Fakes generated by, and Datasets and pretrained models (where license permits) used by, the following:\n https://github.com/deepfakes/faceswap\n https://github.com/joshua-wu/deepfakes_faceswap\n https://github.com/wuhuikai/FaceSwap\n https://github.com/jinfagang/faceswap_pytorch\n https://github.com/iperov/DeepFaceLab\n https://github.com/shaoanlu/faceswap-GAN\n https://github.com/goberoi/faceit\n https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix\n https://github.com/LynnHo/AttGAN-Tensorflow\n https://github.com/LynnHo/DCGAN-LSGAN-WGAN-GP-DRAGAN-Tensorflow-2\n https://github.com/gsurma/face_generator\n https://github.com/snknitin/DeepfakeCapsuleGAN\n https://github.com/tkarras/progressive_growing_of_gans\n https://github.com/pfnet-research/sngan_projection\n https://github.com/mbinkowski/MMD-GAN\n https://github.com/dfaker/df\n https://github.com/NVlabs/stylegan2\n https://github.com/NVlabs/stylegan\n https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix\n https://github.com/NVIDIA/unsupervised-video-interpolation\n https://github.com/NVIDIA/pix2pixHD\n https://github.com/NVIDIA/flownet2-pytorch\n https://github.com/NVIDIA/OpenSeq2Seq\n https://github.com/andabi/deep-voice-conversion\n https://github.com/keithito/tacotron\n https://github.com/NVIDIA/tacotron2\n https://github.com/MycroftAI/mimic2\n https://github.com/Sharad24/Neural-Voice-Cloning-with-Few-Samples\n https://github.com/r9y9/deepvoice3_pytorch\n https://github.com/CorentinJ/Real-Time-Voice-Cloning\nhttps://keithito.com/LJ-Speech-Dataset/\nUNet3D etc at https://github.com/qubvel/tpu/tree/master/models\nhttps://github.com/Res2Net/Res2Net-PretrainedModels\nhttps://github.com/facebookresearch/semi-supervised-ImageNet1K-models\nModels at https://pytorch.org/hub/\nhttps://github.com/matterport/Mask_RCNN\nSegmentation model Unet with different backbone models (resnet, vgg16 etc)\nFastai model zoo and pretrained weights https://docs.fast.ai/vision.models.html\nhttps://github.com/divamgupta/image-segmentation-keras\nModels at https://github.com/NVIDIA/semantic-segmentation\n such as DeepLabV3+ architecture with different backbones, including WideResNet38, SEResNeXt(50, 101) and ResNet(50,101)\nhttps://github.com/gasvn/Res2Net\nhttps://github.com/ansleliu/LightNet\nhttps://github.com/adobe/antialiased-cnns\nhttps://github.com/Media-Smart/vedaseg\nhttps://github.com/tyiannak/pyAudioAnalysis\nhttps://github.com/amsehili/auditok\nhttps://github.com/ina-foss/inaSpeechSegmenter\nhttps://github.com/albietz/online_hmm\nhttps://github.com/qubvel/ttach
", "rawMarkdown": "Fakes generated by, and Datasets and pretrained models (where license permits) used by, the following:\n https://github.com/deepfakes/faceswap\n https://github.com/joshua-wu/deepfakes_faceswap\n https://github.com/wuhuikai/FaceSwap\n https://github.com/jinfagang/faceswap_pytorch\n https://github.com/iperov/DeepFaceLab\n https://github.com/shaoanlu/faceswap-GAN\n https://github.com/goberoi/faceit\n https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix\n https://github.com/LynnHo/AttGAN-Tensorflow\n https://github.com/LynnHo/DCGAN-LSGAN-WGAN-GP-DRAGAN-Tensorflow-2\n https://github.com/gsurma/face_generator\n https://github.com/snknitin/DeepfakeCapsuleGAN\n https://github.com/tkarras/progressive_growing_of_gans\n https://github.com/pfnet-research/sngan_projection\n https://github.com/mbinkowski/MMD-GAN\n https://github.com/dfaker/df\n https://github.com/NVlabs/stylegan2\n https://github.com/NVlabs/stylegan\n https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix\n https://github.com/NVIDIA/unsupervised-video-interpolation\n https://github.com/NVIDIA/pix2pixHD\n https://github.com/NVIDIA/flownet2-pytorch\n https://github.com/NVIDIA/OpenSeq2Seq\n https://github.com/andabi/deep-voice-conversion\n https://github.com/keithito/tacotron\n https://github.com/NVIDIA/tacotron2\n https://github.com/MycroftAI/mimic2\n https://github.com/Sharad24/Neural-Voice-Cloning-with-Few-Samples\n https://github.com/r9y9/deepvoice3_pytorch\n https://github.com/CorentinJ/Real-Time-Voice-Cloning\nhttps://keithito.com/LJ-Speech-Dataset/\nUNet3D etc at https://github.com/qubvel/tpu/tree/master/models\nhttps://github.com/Res2Net/Res2Net-PretrainedModels\nhttps://github.com/facebookresearch/semi-supervised-ImageNet1K-models\nModels at https://pytorch.org/hub/\nhttps://github.com/matterport/Mask_RCNN\nSegmentation model Unet with different backbone models (resnet, vgg16 etc)\nFastai model zoo and pretrained weights https://docs.fast.ai/vision.models.html\nhttps://github.com/divamgupta/image-segmentation-keras\nModels at https://github.com/NVIDIA/semantic-segmentation\n such as DeepLabV3+ architecture with different backbones, including WideResNet38, SEResNeXt(50, 101) and ResNet(50,101)\nhttps://github.com/gasvn/Res2Net\nhttps://github.com/ansleliu/LightNet\nhttps://github.com/adobe/antialiased-cnns\nhttps://github.com/Media-Smart/vedaseg\nhttps://github.com/tyiannak/pyAudioAnalysis\nhttps://github.com/amsehili/auditok\nhttps://github.com/ina-foss/inaSpeechSegmenter\nhttps://github.com/albietz/online_hmm\nhttps://github.com/qubvel/ttach" }, { "id": 762811, "postDate": "2020-03-03T20:28:17.103Z", "content": "https://github.com/1adrianb/face-alignment
", "rawMarkdown": "https://github.com/1adrianb/face-alignment" }, { "id": 762805, "postDate": "2020-03-03T20:23:04.003Z", "content": "https://github.com/YuvalNirkin/face_segmentation.git\nhttps://github.com/rwightman/posenet-python\nhttps://github.com/XifengGuo/CapsNet-Keras\nhttps://github.com/tensorflow/tfjs-models/tree/master/body-pix\nhttps://github.com/shamangary/FSA-Net\nhttps://github.com/1adrianb/face-alignment\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/deepinsight/insightface\nhttps://www.npmjs.com/package/@tensorflow-models/blazeface\nhttps://github.com/scikit-image/scikit-image\nhttps://github.com/librosa/librosa\nhttps://github.com/makcedward/nlpaug\nhttps://github.com/tyiannak/pyAudioAnalysis\nhttps://datashare.is.ed.ac.uk/handle/10283/3336\nhttps://github.com/jefflai108/ASSERT\nhttps://github.com/resemble-ai/Resemblyzer
", "rawMarkdown": "https://github.com/YuvalNirkin/face_segmentation.git\nhttps://github.com/rwightman/posenet-python\nhttps://github.com/XifengGuo/CapsNet-Keras\nhttps://github.com/tensorflow/tfjs-models/tree/master/body-pix\nhttps://github.com/shamangary/FSA-Net\nhttps://github.com/1adrianb/face-alignment\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/deepinsight/insightface\nhttps://www.npmjs.com/package/@tensorflow-models/blazeface\nhttps://github.com/scikit-image/scikit-image\nhttps://github.com/librosa/librosa\nhttps://github.com/makcedward/nlpaug\nhttps://github.com/tyiannak/pyAudioAnalysis\nhttps://datashare.is.ed.ac.uk/handle/10283/3336\nhttps://github.com/jefflai108/ASSERT\nhttps://github.com/resemble-ai/Resemblyzer" }, { "id": 762783, "postDate": "2020-03-03T19:55:51.537Z", "content": "Models
\n\nDarknet (https://github.com/AlexeyAB/darknet#yolo-v3-in-other-frameworks)\nyolov3-tiny-prn :\nhttps://drive.google.com/open?id=1_NnfVgj0EDtb_WLNoXV8Mo7WKgwdYZCc\nyolov3-tiny-prn:\nhttps://raw.githubusercontent.com/AlexeyAB/darknet/master/cfg/csresnext50-panet-spp.cfg
\n\nPytorch model zoo\nresnet18: https://download.pytorch.org/models/resnet18-5c106cde.pth\nresnet34: https://download.pytorch.org/models/resnet34-333f7ec4.pth\nresnet50: https://download.pytorch.org/models/resnet50-19c8e357.pth\nresnet101: https://download.pytorch.org/models/resnet101-5d3b4d8f.pth\nresnet152: https://download.pytorch.org/models/resnet152-b121ed2d.pth\nresnext5032x4d:\nhttps://download.pytorch.org/models/resnext5032x4d-7cdf4587.pth\nresnext10132x8d:\nhttps://download.pytorch.org/models/resnext10132x8d-8ba56ff5.pth\nwideresnet502:\n'https://download.pytorch.org/models/wideresnet502-95faca4d.pth\nwideresnet1012:\n'https://download.pytorch.org/models/wideresnet1012-32ee1156.pth
\n\nPackages\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nFastai2: https://github.com/fastai/fastai2\nOpenCV 4.2.0 (https://github.com/opencv/opencv) including all contrib\nmodules (https://github.com/opencv/opencv_contrib)
", "rawMarkdown": "**Models**\n\nDarknet (https://github.com/AlexeyAB/darknet#yolo-v3-in-other-frameworks)\nyolov3-tiny-prn :\nhttps://drive.google.com/open?id=1_NnfVgj0EDtb_WLNoXV8Mo7WKgwdYZCc\nyolov3-tiny-prn:\nhttps://raw.githubusercontent.com/AlexeyAB/darknet/master/cfg/csresnext50-panet-spp.cfg\n\nPytorch model zoo\nresnet18: https://download.pytorch.org/models/resnet18-5c106cde.pth\nresnet34: https://download.pytorch.org/models/resnet34-333f7ec4.pth\nresnet50: https://download.pytorch.org/models/resnet50-19c8e357.pth\nresnet101: https://download.pytorch.org/models/resnet101-5d3b4d8f.pth\nresnet152: https://download.pytorch.org/models/resnet152-b121ed2d.pth\nresnext5032x4d:\nhttps://download.pytorch.org/models/resnext5032x4d-7cdf4587.pth\nresnext10132x8d:\nhttps://download.pytorch.org/models/resnext10132x8d-8ba56ff5.pth\nwideresnet502:\n'https://download.pytorch.org/models/wideresnet502-95faca4d.pth\nwideresnet1012:\n'https://download.pytorch.org/models/wideresnet1012-32ee1156.pth\n\n\n**Packages**\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nFastai2: https://github.com/fastai/fastai2\nOpenCV 4.2.0 (https://github.com/opencv/opencv) including all contrib\nmodules (https://github.com/opencv/opencv_contrib)\n" }, { "id": 762780, "postDate": "2020-03-03T19:53:58.747Z", "content": "Pretrained weights and code:\nhttps://github.com/sthanhng/yoloface\nhttps://github.com/pjreddie/darknet\nhttps://pjreddie.com/media/files/yolov3.weights\nhttps://pjreddie.com/media/files/yolov3-tiny.weights\nhttps://sites.google.com/view/perception-cv4arvr/blazeface
\n\nDatasets:\nhttps://pjreddie.com/projects/pascal-voc-dataset-mirror/\nhttp://cocodataset.org
", "rawMarkdown": "Pretrained weights and code:\nhttps://github.com/sthanhng/yoloface\nhttps://github.com/pjreddie/darknet\nhttps://pjreddie.com/media/files/yolov3.weights\nhttps://pjreddie.com/media/files/yolov3-tiny.weights\nhttps://sites.google.com/view/perception-cv4arvr/blazeface\n\nDatasets:\nhttps://pjreddie.com/projects/pascal-voc-dataset-mirror/\nhttp://cocodataset.org" }, { "id": 762713, "postDate": "2020-03-03T18:46:03.850Z", "content": "https://github.com/nyoki-mtl/keras-facenet\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/danmohaha/celeb-deepfakeforensics
", "rawMarkdown": "https://github.com/nyoki-mtl/keras-facenet\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/danmohaha/celeb-deepfakeforensics" }, { "id": 762701, "postDate": "2020-03-03T18:25:08.683Z", "content": "pyopencl\ntimm\nSSIM-PIL
", "rawMarkdown": "pyopencl\ntimm\nSSIM-PIL" }, { "id": 762697, "postDate": "2020-03-03T18:19:00.840Z", "content": "Some people mentioned s3fd related resources. However, it's not clear to me if the underlying datasets and libraries are the same, so I posting what exactly we might use:\n'2DFAN-4': 'https://www.adrianbulat.com/downloads/python-fan/2DFAN4-11f355bf06.pth.tar',\n'3DFAN-4': 'https://www.adrianbulat.com/downloads/python-fan/3DFAN4-7835d9f11d.pth.tar',\n 'depth': 'https://www.adrianbulat.com/downloads/python-fan/depth-2a464da4ea.pth.tar',\n's3fd': 'https://www.adrianbulat.com/downloads/python-fan/s3fd-619a316812.pth
", "rawMarkdown": "Some people mentioned s3fd related resources. However, it's not clear to me if the underlying datasets and libraries are the same, so I posting what exactly we might use:\n'2DFAN-4': 'https://www.adrianbulat.com/downloads/python-fan/2DFAN4-11f355bf06.pth.tar',\n'3DFAN-4': 'https://www.adrianbulat.com/downloads/python-fan/3DFAN4-7835d9f11d.pth.tar',\n 'depth': 'https://www.adrianbulat.com/downloads/python-fan/depth-2a464da4ea.pth.tar',\n's3fd': 'https://www.adrianbulat.com/downloads/python-fan/s3fd-619a316812.pth" }, { "id": 762609, "postDate": "2020-03-03T16:36:34.983Z", "content": "pytorch, skimage, sklearn, libsvm, opencv\nhttps://github.com/cvondrick/soundnet\nhttps://github.com/keunhong/pytorch-soundnet\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/bukalapak/pybrisque\nhttps://pypi.org/project/pydub/\nhttps://github.com/timesler/facenet-pytorch
", "rawMarkdown": "pytorch, skimage, sklearn, libsvm, opencv\nhttps://github.com/cvondrick/soundnet\nhttps://github.com/keunhong/pytorch-soundnet\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/bukalapak/pybrisque\nhttps://pypi.org/project/pydub/\nhttps://github.com/timesler/facenet-pytorch" }, { "id": 762602, "postDate": "2020-03-03T16:26:37.120Z", "content": "These are some of the resources that has helped me/is helping me to understand deepFakes and deepFake detection properly!
\n\nhttps://arxiv.org/pdf/1909.11573.pdf\nhttps://arxiv.org/pdf/1812.08685.pdf\nhttps://publications.idiap.ch/downloads/papers/2019/Korshunov_ICB_2019.pdf\nhttps://arxiv.org/pdf/2001.00179.pdf
", "rawMarkdown": "These are some of the resources that has helped me/is helping me to understand deepFakes and deepFake detection properly!\n\nhttps://arxiv.org/pdf/1909.11573.pdf\nhttps://arxiv.org/pdf/1812.08685.pdf\nhttps://publications.idiap.ch/downloads/papers/2019/Korshunov_ICB_2019.pdf\nhttps://arxiv.org/pdf/2001.00179.pdf\n\n" }, { "id": 762589, "postDate": "2020-03-03T16:17:17.930Z", "content": "libfacedetection
", "rawMarkdown": "libfacedetection" }, { "id": 762562, "postDate": "2020-03-03T16:01:56.320Z", "content": "https://github.com/sthanhng/yoloface\nhttps://github.com/eriklindernoren/PyTorch-YOLOv3
", "rawMarkdown": "https://github.com/sthanhng/yoloface\nhttps://github.com/eriklindernoren/PyTorch-YOLOv3" }, { "id": 762538, "postDate": "2020-03-03T15:38:36.023Z", "content": "I don't know if this is allowed. If not, we will not use it.
\n\nhttps://www.kaggle.com/greatgamedota/ffhq-face-data-set
", "rawMarkdown": "I don't know if this is allowed. If not, we will not use it.\n\nhttps://www.kaggle.com/greatgamedota/ffhq-face-data-set" }, { "id": 762510, "postDate": "2020-03-03T15:12:41.367Z", "content": "ImageNet pre-trained models in pytorchcv: https://pypi.org/project/pytorchcv/\nImageNet pre-trained models in PyTorch Hub: https://pytorch.org/hub/\npython_speech_features: https://pypi.org/project/python_speech_features/0.4/
", "rawMarkdown": "ImageNet pre-trained models in pytorchcv: https://pypi.org/project/pytorchcv/\nImageNet pre-trained models in PyTorch Hub: https://pytorch.org/hub/\npython_speech_features: https://pypi.org/project/python_speech_features/0.4/\n" }, { "id": 762498, "postDate": "2020-03-03T15:04:40.873Z", "content": "https://github.com/iitzco/faced\nhttps://github.com/tensorflow/tfjs-models/tree/master/posenet\nhttps://github.com/matterport/Mask_RCNN
", "rawMarkdown": "https://github.com/iitzco/faced\nhttps://github.com/tensorflow/tfjs-models/tree/master/posenet\nhttps://github.com/matterport/Mask_RCNN" }, { "id": 762467, "postDate": "2020-03-03T14:34:47.330Z", "content": "BlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch
", "rawMarkdown": "BlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch" }, { "id": 762458, "postDate": "2020-03-03T14:29:18.623Z", "content": "facenet-pytorch: https://github.com/timesler/facenet-pytorch\nhttps://keras.io/applications/\nhttps://github.com/qubvel/efficientnet
\n\nDatasets:\nImageNet
", "rawMarkdown": "facenet-pytorch: https://github.com/timesler/facenet-pytorch\nhttps://keras.io/applications/\nhttps://github.com/qubvel/efficientnet\n\nDatasets:\nImageNet" }, { "id": 762417, "postDate": "2020-03-03T13:52:10.317Z", "content": "Datasets:\n- ImageNet \n- CIFAR https://www.cs.toronto.edu/~kriz/cifar.html \n- http://live.ece.utexas.edu/research/incaptureDatabase/index.html \n- http://shuoyang1213.me/WIDERFACE/
\n- http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/ \n- https://www.umdfaces.io/ \n- http://sintel.is.tue.mpg.de/
Apparently, mostly all resources are restricted due to poor licenses (as I understood). So, I’ll write down a list nice to use…\n- https://github.com/lidq92/VSFA\n- http://tamaraberg.com/faceDataset/index.html \n- http://www.helsinki.fi/psychology/groups/visualcognition/ \n- http://vision.eng.shizuoka.ac.jp/course/view.php?id=6\n- https://www.cs.tau.ac.il/~wolf/ytfaces/ \n- http://vis-www.cs.umass.edu/lfw/ \n- https://github.com/1adrianb/face-alignment \n- https://github.com/ondyari/FaceForensics/tree/original \n- https://github.com/ondyari/FaceForensics/tree/master/classification \n- https://github.com/deepinsight/insightface
", "rawMarkdown": "- FastAi and pre-trained models (depends on torchvision): https://github.com/fastai/fastai and https://github.com/fastai/fastai2\n- mtcnn: https://pypi.org/project/mtcnn/\n- https://github.com/daniilidis-group/neural_renderer \n- https://github.com/fbcotter/pytorch_wavelets \n- https://imgaug.readthedocs.io/en/latest/\n- Dlib (https://github.com/davisking/dlib) with pretrained models from http://dlib.net/files/ \n- https://github.com/hollance/BlazeFace-PyTorch\n- OpenCV and OpenCV_contrib with pretrained models downloaded via https://github.com/opencv/opencv/tree/master/samples/dnn and data https://github.com/opencv/opencv/tree/master/data \n- facenet-pytorch: https://github.com/timesler/facenet-pytorch\n- skimage (https://scikit-image.org/), sklearn (https://scikit-learn.org/stable/), filterpy (https://filterpy.readthedocs.io/en/latest/), scipy (https://scipy.org/), Pandas\n- https://github.com/scikit-video/scikit-video \n- https://github.com/google/mediapipe \n- https://github.com/rwightman/pytorch-image-models \n- https://github.com/grib0ed0v/face_recognition.pytorch \n- https://github.com/lukemelas/EfficientNet-PyTorch \n- https://github.com/rwightman/pytorch-image-models \n- https://github.com/lessw2020/Ranger-Deep-Learning-Optimizer\n- https://github.com/TencentYoutuResearch/FaceDetection-DSFD \n- https://github.com/ZhaoJ9014/face.evoLVe.PyTorch \n- https://github.com/marvis/pytorch-caffe\n- https://github.com/NVIDIA/flownet2-pytorch\n- https://github.com/RanhaoKang/PWC-Net_pytorch \n- https://www.kaggle.com/hmendonca/proper-clustering-with-facenet-embeddings-eda/data\n\nDatasets:\n- ImageNet \n- CIFAR https://www.cs.toronto.edu/~kriz/cifar.html \n- http://live.ece.utexas.edu/research/incaptureDatabase/index.html \n- http://shuoyang1213.me/WIDERFACE/ \n- http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/ \n- https://www.umdfaces.io/ \n- http://sintel.is.tue.mpg.de/\n\nApparently, mostly all resources are restricted due to poor licenses (as I understood). So, I’ll write down a list nice to use…\n- https://github.com/lidq92/VSFA\n- http://tamaraberg.com/faceDataset/index.html \n- http://www.helsinki.fi/psychology/groups/visualcognition/ \n- http://vision.eng.shizuoka.ac.jp/course/view.php?id=6\n- https://www.cs.tau.ac.il/~wolf/ytfaces/ \n- http://vis-www.cs.umass.edu/lfw/ \n- https://github.com/1adrianb/face-alignment \n- https://github.com/ondyari/FaceForensics/tree/original \n- https://github.com/ondyari/FaceForensics/tree/master/classification \n- https://github.com/deepinsight/insightface \n" }, { "id": 762354, "postDate": "2020-03-03T13:09:02Z", "content": "DeeperForensics-1.0 also : https://github.com/EndlessSora/DeeperForensics-1.0
\n\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nRetinaFace : https://github.com/deepinsight/insightface/tree/master/RetinaFace
\n\nKeras Xception Imagenet Weights https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xceptionweightstfdimorderingtfkernels_notop.h5
\n\nEfficient Net Weights https://github.com/qubvel/efficientnet
\n\nImgaug: https://imgaug.readthedocs.io/en/latest/\nhttps://github.com/ipazc/mtcnn
\n\nhttps://github.com/albumentations-team/albumentations
\n\nhttps://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttps://github.com/rwightman/pytorch-image-models\nhttps://pypi.org/project/gluoncv2/\nhttps://github.com/ondyari/FaceForensics/tree/master/classification\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/mnikitin/EfficientNet\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://github.com/narumiruna/efficientnet-pytorch\nhttps://github.com/zsef123/EfficientNets-PyTorch\nhttps://mxnet.apache.org/api/python/docs/api/gluon/model_zoo/index.html\nhttp://www.cs.albany.edu/~lsw/celeb-deepfakeforensics.html\nhttps://github.com/ox-vgg/vgg_face2\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/grib0ed0v/face_recognition.pytorch\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nhttps://github.com/deepinsight/insightface\nhttp://shuoyang1213.me/WIDERFACE/\nhttps://github.com/lijiannuist/lightDSFD\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/yxlijun/S3FD.pytorch\nhttps://github.com/supernotman/RetinaFace_Pytorch\nhttps://github.com/bogireddytejareddy/retinaface-pytorch\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/clovaai/EXTD_Pytorch\nhttps://github.com/ElvishElvis/68-Retinaface-Pytorch-version\nhttps://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/sfzhang15/SFD\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/hcl14/retinaface-pytorch-inference\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD
\n\nPretrained weights from https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nFaces from https://github.com/NVlabs/ffhq-dataset with open licenses (https://creativecommons.org/publicdomain/mark/1.0/, https://creativecommons.org/publicdomain/zero/1.0/, https://creativecommons.org/licenses/by/2.0/)\nVideo Tools to install packages https://www.kaggle.com/harangdev/video-tools\nfaceswap tool https://github.com/deepfakes/faceswap and models https://github.com/deepfakes-models/faceswap-models/releases
", "rawMarkdown": "DeeperForensics-1.0 also : https://github.com/EndlessSora/DeeperForensics-1.0\n\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nRetinaFace : https://github.com/deepinsight/insightface/tree/master/RetinaFace\n\nKeras Xception Imagenet Weights https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xceptionweightstfdimorderingtfkernels_notop.h5\n\nEfficient Net Weights https://github.com/qubvel/efficientnet\n\nImgaug: https://imgaug.readthedocs.io/en/latest/\nhttps://github.com/ipazc/mtcnn\n\nhttps://github.com/albumentations-team/albumentations\n\nhttps://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttps://github.com/rwightman/pytorch-image-models\nhttps://pypi.org/project/gluoncv2/\nhttps://github.com/ondyari/FaceForensics/tree/master/classification\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/mnikitin/EfficientNet\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://github.com/narumiruna/efficientnet-pytorch\nhttps://github.com/zsef123/EfficientNets-PyTorch\nhttps://mxnet.apache.org/api/python/docs/api/gluon/model_zoo/index.html\nhttp://www.cs.albany.edu/~lsw/celeb-deepfakeforensics.html\nhttps://github.com/ox-vgg/vgg_face2\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/grib0ed0v/face_recognition.pytorch\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nhttps://github.com/deepinsight/insightface\nhttp://shuoyang1213.me/WIDERFACE/\nhttps://github.com/lijiannuist/lightDSFD\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/yxlijun/S3FD.pytorch\nhttps://github.com/supernotman/RetinaFace_Pytorch\nhttps://github.com/bogireddytejareddy/retinaface-pytorch\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/clovaai/EXTD_Pytorch\nhttps://github.com/ElvishElvis/68-Retinaface-Pytorch-version\nhttps://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/sfzhang15/SFD\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/hcl14/retinaface-pytorch-inference\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\n\nPretrained weights from https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nFaces from https://github.com/NVlabs/ffhq-dataset with open licenses (https://creativecommons.org/publicdomain/mark/1.0/, https://creativecommons.org/publicdomain/zero/1.0/, https://creativecommons.org/licenses/by/2.0/)\nVideo Tools to install packages https://www.kaggle.com/harangdev/video-tools\nfaceswap tool https://github.com/deepfakes/faceswap and models https://github.com/deepfakes-models/faceswap-models/releases\n\n" }, { "id": 762353, "postDate": "2020-03-03T13:05:47.763Z", "content": "classy vision: https://github.com/facebookresearch/ClassyVision
", "rawMarkdown": "classy vision: https://github.com/facebookresearch/ClassyVision" }, { "id": 762335, "postDate": "2020-03-03T12:47:28.973Z", "content": "Pytorch and Dlib models: http://dlib.net/files/, https://pytorch.org/docs/stable/torchvision/models.html
", "rawMarkdown": "Pytorch and Dlib models: [http://dlib.net/files/](http://dlib.net/files/), [https://pytorch.org/docs/stable/torchvision/models.html](https://pytorch.org/docs/stable/torchvision/models.html)" }, { "id": 762333, "postDate": "2020-03-03T12:46:10.140Z", "content": "https://github.com/nesl/asvspoof2019 (MIT)
", "rawMarkdown": "https://github.com/nesl/asvspoof2019 (MIT)" }, { "id": 762313, "postDate": "2020-03-03T12:23:17.177Z", "content": "MTCNN package :- https://github.com/ipazc/mtcnn\nDSFD from tencent\nPytorch pre-trained models
", "rawMarkdown": "MTCNN package :- https://github.com/ipazc/mtcnn\nDSFD from tencent\nPytorch pre-trained models" }, { "id": 762302, "postDate": "2020-03-03T12:15:31.640Z", "content": "", "rawMarkdown": "https://www.deepfaker.xyz/" }, { "id": 762290, "postDate": "2020-03-03T12:02:29.633Z", "content": "https://github.com/qubvel/efficientnet\nhttps://github.com/Callidior/keras-applications/releases/tag/efficientnet
", "rawMarkdown": "https://github.com/qubvel/efficientnet\nhttps://github.com/Callidior/keras-applications/releases/tag/efficientnet" }, { "id": 762282, "postDate": "2020-03-03T11:55:28.947Z", "content": "https://pypi.org/project/ffmpeg-python/\nhttps://johnvansickle.com/ffmpeg/
", "rawMarkdown": "https://pypi.org/project/ffmpeg-python/\nhttps://johnvansickle.com/ffmpeg/" }, { "id": 762276, "postDate": "2020-03-03T11:51:03.487Z", "content": "https://pypi.org/project/mtcnn/
", "rawMarkdown": "https://pypi.org/project/mtcnn/" }, { "id": 762274, "postDate": "2020-03-03T11:49:12.610Z", "content": "Xception Imagenet weights (no top): https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xceptionweightstfdimorderingtfkernels_notop.h5
", "rawMarkdown": "Xception Imagenet weights (no top): https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xceptionweightstfdimorderingtfkernels_notop.h5" }, { "id": 762251, "postDate": "2020-03-03T11:16:37.300Z", "content": "Bob’s library of image-quality feature-extractors\nhttps://pypi.org/project/bob.ip.qualitymeasure/
Scikit image\nhttps://pypi.org/project/scikit-image/
Lightgbm \nhttps://pypi.org/project/lightgbm/
Scikit learn\nhttps://pypi.org/project/scikit-learn/
https://www.kaggle.com/xhlulu/flickrfaceshq-dataset-nvidia-resized-256px\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/123665\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/121173\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://ffmpeg.org/ffmpeg-all.html\nhttps://github.com/yeephycho/tensorflow-face-detection\nDCFC preview dataset\nyoutube.com\nCeleb-DF-v2(https://arxiv.org/abs/1909.12962)
", "rawMarkdown": "https://www.kaggle.com/xhlulu/flickrfaceshq-dataset-nvidia-resized-256px\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/123665\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/121173\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://ffmpeg.org/ffmpeg-all.html\nhttps://github.com/yeephycho/tensorflow-face-detection\nDCFC preview dataset\nyoutube.com\nCeleb-DF-v2(https://arxiv.org/abs/1909.12962)\n\n" }, { "id": 762180, "postDate": "2020-03-03T09:36:13.007Z", "content": "https://www.kaggle.com/josecarmona/ffmpeg-python-example-to-extract-audio-from-mp4
", "rawMarkdown": "https://www.kaggle.com/josecarmona/ffmpeg-python-example-to-extract-audio-from-mp4" }, { "id": 762178, "postDate": "2020-03-03T09:33:58.483Z", "content": "facenet: https://github.com/timesler/facenet-pytorch\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nVgg19: https://download.pytorch.org/models/vgg19-dcbb9e9d.pth\nPre-trained models of VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/
\n\nface-recognition: https://github.com/ageitgey/face_recognition\nface-recognition-models: https://github.com/ageitgey/face_recognition_models
\n\nblazeFace: https://github.com/hollance/BlazeFace-PyTorch\nFfmpeg: Static Build - https://www.kaggle.com/rakibilly/ffmpeg-static-build
\n\nFake Audio dataset: https://datashare.is.ed.ac.uk/handle/10283/3336
\n\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch \nhttps://github.com/hollance/BlazeFace-PyTorch\nRetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nPytorch Retinaface: https://github.com/biubug6/Pytorch_Retinaface\nWIDER FACE Dataset: http://shuoyang1213.me/WIDERFACE/
\n\nResemblyzer: https://github.com/resemble-ai/Resemblyzer\nReal-Time Voice Cloning: https://github.com/CorentinJ/Real-Time-Voice-Cloning\nThe VoxCeleb Dataset\n http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1.html\n http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox2.html\nThe M-AILABS Speech Dataset\n https://www.caito.de/2019/01/the-m-ailabs-speech-dataset/\nhttps://librosa.github.io/librosa/
", "rawMarkdown": "\nfacenet: https://github.com/timesler/facenet-pytorch\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nVgg19: https://download.pytorch.org/models/vgg19-dcbb9e9d.pth\nPre-trained models of VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\n\nface-recognition: https://github.com/ageitgey/face_recognition\nface-recognition-models: https://github.com/ageitgey/face_recognition_models\n\nblazeFace: https://github.com/hollance/BlazeFace-PyTorch\nFfmpeg: Static Build - https://www.kaggle.com/rakibilly/ffmpeg-static-build \n\nFake Audio dataset: https://datashare.is.ed.ac.uk/handle/10283/3336\n\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch \nhttps://github.com/hollance/BlazeFace-PyTorch\nRetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nPytorch Retinaface: https://github.com/biubug6/Pytorch_Retinaface\nWIDER FACE Dataset: http://shuoyang1213.me/WIDERFACE/\n\nResemblyzer: https://github.com/resemble-ai/Resemblyzer\nReal-Time Voice Cloning: https://github.com/CorentinJ/Real-Time-Voice-Cloning\nThe VoxCeleb Dataset\n http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1.html\n http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox2.html\nThe M-AILABS Speech Dataset\n https://www.caito.de/2019/01/the-m-ailabs-speech-dataset/\nhttps://librosa.github.io/librosa/" }, { "id": 762114, "postDate": "2020-03-03T08:08:25.040Z", "content": "https://librosa.github.io/librosa/
", "rawMarkdown": "https://librosa.github.io/librosa/" }, { "id": 762094, "postDate": "2020-03-03T07:50:40.513Z", "content": "Blazeface TensorFlow Lite model from https://github.com/google/mediapipe
", "rawMarkdown": "Blazeface TensorFlow Lite model from https://github.com/google/mediapipe " }, { "id": 762055, "postDate": "2020-03-03T06:58:24.527Z", "content": "https://www.kaggle.com/chopinforest1986/efficientnetb0b7-keras-weights
", "rawMarkdown": "https://www.kaggle.com/chopinforest1986/efficientnetb0b7-keras-weights" }, { "id": 762048, "postDate": "2020-03-03T06:45:24.620Z", "content": "External Data
\n\n1.https://www.kaggle.com/prashantkikani/efficientnet-pytorch(pretrainedmodel)\n2.https://www.kaggle.com/rishabhiitbhu/pretrainedmodels(pretrained model)\n3.torchvision.models (pretrained model)\n4.https://www.kaggle.com/timesler/facenet-pytorch-vggface2(face extract)\n5.https://www.kaggle.com/humananalog/blazeface-pytorch and https://www.kaggle.com/humananalog/deepfakes-inference-demo (face extract)\n6.https://github.com/ondyari/FaceForensics （FaceForensics++ dataset）\n7.https://github.com/tensorflow/tensorflow/blob/3d86d8ce14989ca65a59ad4cf37f690694bf6267/tensorflow/contrib/factorization/python/ops/gmm.py (gmm model code)
", "rawMarkdown": "External Data\n\n1.https://www.kaggle.com/prashantkikani/efficientnet-pytorch(pretrainedmodel)\n2.https://www.kaggle.com/rishabhiitbhu/pretrainedmodels(pretrained model)\n3.torchvision.models (pretrained model)\n4.https://www.kaggle.com/timesler/facenet-pytorch-vggface2(face extract)\n5.https://www.kaggle.com/humananalog/blazeface-pytorch and https://www.kaggle.com/humananalog/deepfakes-inference-demo (face extract)\n6.https://github.com/ondyari/FaceForensics （FaceForensics++ dataset）\n7.https://github.com/tensorflow/tensorflow/blob/3d86d8ce14989ca65a59ad4cf37f690694bf6267/tensorflow/contrib/factorization/python/ops/gmm.py (gmm model code)" }, { "id": 762038, "postDate": "2020-03-03T06:34:02.810Z", "content": "https://github.com/danmohaha/DSP-FWA
", "rawMarkdown": "https://github.com/danmohaha/DSP-FWA" }, { "id": 762025, "postDate": "2020-03-03T06:00:06.837Z", "content": "https://github.com/calmisential/TensorFlow2.0_ResNet\nhttps://www.kaggle.com/xiaofengmao/retinaface\nhttps://www.kaggle.com/daddyjin/scikitvideo1111\nhttps://files.pythonhosted.org/packages/31/d4/bcdbad92101430ff9a5161eb7612fc0e66cae96daa81953edb17aa3d7c37/librosa-0.7.0rc1-py3-none-any.whl\nhttps://files.pythonhosted.org/packages/73/63/ebf4332964fba68f72cd78723a2722d77cf78f988b13461c6fbf30fc96bc/SoundFile-0.10.3.post1-py2.py3.cp26.cp27.cp32.cp33.cp34.cp35.cp36.pp27.pp32.pp33-none-win32.whl
", "rawMarkdown": "https://github.com/calmisential/TensorFlow2.0_ResNet\nhttps://www.kaggle.com/xiaofengmao/retinaface\nhttps://www.kaggle.com/daddyjin/scikitvideo1111\nhttps://files.pythonhosted.org/packages/31/d4/bcdbad92101430ff9a5161eb7612fc0e66cae96daa81953edb17aa3d7c37/librosa-0.7.0rc1-py3-none-any.whl\nhttps://files.pythonhosted.org/packages/73/63/ebf4332964fba68f72cd78723a2722d77cf78f988b13461c6fbf30fc96bc/SoundFile-0.10.3.post1-py2.py3.cp26.cp27.cp32.cp33.cp34.cp35.cp36.pp27.pp32.pp33-none-win32.whl" }, { "id": 761998, "postDate": "2020-03-03T05:29:16.550Z", "content": "https://github.com/ondyari/FaceForensics
", "rawMarkdown": "https://github.com/ondyari/FaceForensics" }, { "id": 761995, "postDate": "2020-03-03T05:27:11.200Z", "content": "https://pypi.org/project/pytorchcv/0.0.13/\nhttps://github.com/yeephycho/tensorflow-face-detection
", "rawMarkdown": "https://pypi.org/project/pytorchcv/0.0.13/\nhttps://github.com/yeephycho/tensorflow-face-detection" }, { "id": 761987, "postDate": "2020-03-03T05:23:58.090Z", "content": "BlazeFace PyTorch: https://github.com/hollance/BlazeFace-PyTorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nAgeGenderDeepLearning: https://github.com/GilLevi/AgeGenderDeepLearning/tree/master/models\nKeras Pretrained Xception: https://keras.io/applications/#xception\nKeras Pretrained Inception: https://keras.io/applications/#inceptionv3\nKeras Pretrained Resnet: https://keras.io/applications/#resnet\nPublic Kaggle Kernel: \nFaceRecognition: https://github.com/grib0ed0v/face_recognition.pytorch\nPytorch Retinaface: https://github.com/biubug6/Pytorch_Retinaface\nEfficient Net Weights https://github.com/qubvel/efficientnet\nhttps://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md\nhttps://github.com/opencv/opencv/tree/master/data/haarcascades
", "rawMarkdown": "BlazeFace PyTorch: https://github.com/hollance/BlazeFace-PyTorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nAgeGenderDeepLearning: https://github.com/GilLevi/AgeGenderDeepLearning/tree/master/models\nKeras Pretrained Xception: https://keras.io/applications/#xception\nKeras Pretrained Inception: https://keras.io/applications/#inceptionv3\nKeras Pretrained Resnet: https://keras.io/applications/#resnet\nPublic Kaggle Kernel: \nFaceRecognition: https://github.com/grib0ed0v/face_recognition.pytorch\nPytorch Retinaface: https://github.com/biubug6/Pytorch_Retinaface\nEfficient Net Weights https://github.com/qubvel/efficientnet\nhttps://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md\nhttps://github.com/opencv/opencv/tree/master/data/haarcascades" }, { "id": 761933, "postDate": "2020-03-03T03:41:23.907Z", "content": "FaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nDeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nRetinaFace Pytorch detector: https://github.com/biubug6/Pytorch_Retinaface\nhrnet: https://github.com/HRNet/HRNet-Image-Classification\nFFMPEG: https://johnvansickle.com/ffmpeg/
", "rawMarkdown": "FaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nDeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nRetinaFace Pytorch detector: https://github.com/biubug6/Pytorch_Retinaface\nhrnet: https://github.com/HRNet/HRNet-Image-Classification\nFFMPEG: https://johnvansickle.com/ffmpeg/" }, { "id": 761891, "postDate": "2020-03-03T02:37:51.870Z", "content": "keras inception_v3 imagenet weights https://github.com/fchollet/deep-learning-models/releases/download/v0.5/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5
\n\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics
\n\nyou tube faces dataset - http://www.cs.tau.ac.il/~wolf/ytfaces/\n(I don't see anywhere on this site/page that it cannot be used for commercial purposes. So I must assume the opposite to be true.)
\n\nHoha Dataset - https://www.di.ens.fr/~laptev/actions/ (currently cant download but may use if commercially allowed)
\n\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/
", "rawMarkdown": "keras inception_v3 imagenet weights https://github.com/fchollet/deep-learning-models/releases/download/v0.5/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5\n\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\n\nyou tube faces dataset - http://www.cs.tau.ac.il/~wolf/ytfaces/\n(I don't see anywhere on this site/page that it cannot be used for commercial purposes. So I must assume the opposite to be true.)\n\nHoha Dataset - https://www.di.ens.fr/~laptev/actions/ (currently cant download but may use if commercially allowed)\n\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/\n" }, { "id": 761868, "postDate": "2020-03-03T02:14:17.253Z", "content": "https://drive.google.com/drive/folders/10AVxwiSundwHoJxONKYF5U9ieZHPkXXR?usp=sharing\ncontains 4 publicly shared files: yolo.weights, yolo.cfg, labels.txt & model.h5
", "rawMarkdown": "https://drive.google.com/drive/folders/10AVxwiSundwHoJxONKYF5U9ieZHPkXXR?usp=sharing\ncontains 4 publicly shared files: yolo.weights, yolo.cfg, labels.txt & model.h5" }, { "id": 761863, "postDate": "2020-03-03T02:00:53.497Z", "content": "https://github.com/foamliu/InsightFace-v2\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/facebookresearch/SlowFast\nhttps://github.com/biubug6/Pytorch_Retinaface
", "rawMarkdown": "https://github.com/foamliu/InsightFace-v2\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/facebookresearch/SlowFast\nhttps://github.com/biubug6/Pytorch_Retinaface" }, { "id": 761806, "postDate": "2020-03-03T00:27:06.483Z", "content": "1) ffmpeg: (all declared)
\n\nhttps://pypi.org/project/ffmpeg/ (offline pip installation)\nhttps://johnvansickle.com/ffmpeg/\nhttps://github.com/kkroening/ffmpeg-python\nhttps://www.kaggle.com/sheldonrobinson/video-tools\n2) OpenCV pre-trained weight XMLs: (declared)
\n\nhttps://github.com/opencv/opencv/tree/master/data/haarcascades\n3) pre-trained models and weights listed these TF-hub official sites below
\n\n[Mobilenet] https://tfhub.dev/s?fine-tunable=yes&module-type=image-classification&tf-version=tf2\n[Efficientnet] https://tfhub.dev/google/collections/efficientnet/1\n\nnote: these models above are pre-trained using ImageNet (according to GGL)\n4) additional Python modules to be installed (offline) using pip:
\n\npydub: https://pypi.org/project/pydub/\nmoviepy: https://pypi.org/project/moviepy/\npandas: https://pypi.org/project/pandas/\nseaborn: https://pypi.org/project/seaborn/\n\nnote: dependent modules on these above are included if not installed in default Kaggle Notebook platform\nQ: asking this just in case. we do NOT need to declare Python pip packages that can be installed additionally with '!pip install xxxx' even offline? am I understanding declare rule correct?
", "rawMarkdown": "1) ffmpeg: (all declared)\n\n https://pypi.org/project/ffmpeg/ (offline pip installation)\n https://johnvansickle.com/ffmpeg/\n https://github.com/kkroening/ffmpeg-python\n https://www.kaggle.com/sheldonrobinson/video-tools\n\n2) OpenCV pre-trained weight XMLs: (declared)\n\n https://github.com/opencv/opencv/tree/master/data/haarcascades\n \n3) pre-trained models and weights listed these TF-hub official sites below\n\n [Mobilenet] https://tfhub.dev/s?fine-tunable=yes&module-type=image-classification&tf-version=tf2\n [Efficientnet] https://tfhub.dev/google/collections/efficientnet/1\n\n note: these models above are pre-trained using ImageNet (according to GGL)\n\n4) additional Python modules to be installed (offline) using pip:\n\n pydub: https://pypi.org/project/pydub/\n moviepy: https://pypi.org/project/moviepy/\n pandas: https://pypi.org/project/pandas/\n seaborn: https://pypi.org/project/seaborn/\n\n note: dependent modules on these above are included if not installed in default Kaggle Notebook platform\n\nQ: asking this just in case. we do NOT need to declare Python pip packages that can be installed additionally with '!pip install xxxx' even offline? am I understanding declare rule correct?" }, { "id": 761799, "postDate": "2020-03-03T00:21:59.920Z", "content": "A light and fast face detector (lffd):\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/fyr91/face_detection\nhttps://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices/tree/master/face_detection
\n\nSome architectures and pretrained models and datasets:\nhttp://www.robots.ox.ac.uk/~vgg/data/ \nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/ondyari/FaceForensics
\n\nLip Reading:\nhttp://www.robots.ox.ac.uk/~vgg/data/lip_reading/\nhttps://github.com/astorfi/lip-reading-deeplearning (Apache License)\nhttps://github.com/joseph-zhong/LipReading\nhttps://github.com/afourast/deep_lip_reading\nhttps://github.com/hassanhub/LipReading
\n\nEverything at:\nhttps://keras.io/applications/\nhttps://github.com/keras-team/keras-applications
\n\nOnnx: https://github.com/onnx/onnx\nONNXRuntime: https://www.wheelodex.org/projects/onnxruntime/ (and in general, any wheel at https://www.wheelodex.org/projects/)\nFaceRecognition: https://github.com/ageitgey/face_recognition\nBlazeFace: https://www.kaggle.com/humananalog/blazeface-pytorch
", "rawMarkdown": "A light and fast face detector (lffd):\n[https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB](https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB)\n[https://github.com/fyr91/face_detection](https://github.com/fyr91/face_detection)\n[https://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices/tree/master/face_detection](https://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices/tree/master/face_detection)\n\nSome architectures and pretrained models and datasets:\n[http://www.robots.ox.ac.uk/~vgg/data/ ](http://www.robots.ox.ac.uk/~vgg/data/ )\n[http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/](http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/)\n[http://www.robots.ox.ac.uk/~vgg/data/voxceleb/](http://www.robots.ox.ac.uk/~vgg/data/voxceleb/)\n[https://github.com/davidsandberg/facenet](https://github.com/davidsandberg/facenet)\n[https://github.com/ondyari/FaceForensics](https://github.com/ondyari/FaceForensics)\n\nLip Reading:\n[http://www.robots.ox.ac.uk/~vgg/data/lip_reading/](http://www.robots.ox.ac.uk/~vgg/data/lip_reading/)\n[https://github.com/astorfi/lip-reading-deeplearning (Apache License)](https://github.com/astorfi/lip-reading-deeplearning)\n[https://github.com/joseph-zhong/LipReading](https://github.com/joseph-zhong/LipReading)\n[https://github.com/afourast/deep_lip_reading](https://github.com/afourast/deep_lip_reading)\n[https://github.com/hassanhub/LipReading](https://github.com/hassanhub/LipReading)\n\nEverything at:\n[https://keras.io/applications/](https://keras.io/applications/)\n[https://github.com/keras-team/keras-applications](https://github.com/keras-team/keras-applications)\n\nOnnx: https://github.com/onnx/onnx\nONNXRuntime: https://www.wheelodex.org/projects/onnxruntime/ (and in general, any wheel at https://www.wheelodex.org/projects/)\nFaceRecognition: https://github.com/ageitgey/face_recognition\nBlazeFace: https://www.kaggle.com/humananalog/blazeface-pytorch" }, { "id": 761796, "postDate": "2020-03-03T00:19:44.877Z", "content": "Pretrained XCeption, ResNet
", "rawMarkdown": "Pretrained XCeption, ResNet" }, { "id": 761784, "postDate": "2020-03-02T23:39:11.873Z", "content": "Pytorchcv: https://pypi.org/project/pytorchcv/\nBlazeface: https://www.kaggle.com/humananalog/blazeface-pytorch\nResnext pretrained: https://www.kaggle.com/humananalog/deepfakes-inference-demo
", "rawMarkdown": "Pytorchcv: https://pypi.org/project/pytorchcv/\nBlazeface: https://www.kaggle.com/humananalog/blazeface-pytorch\nResnext pretrained: https://www.kaggle.com/humananalog/deepfakes-inference-demo" }, { "id": 761755, "postDate": "2020-03-02T22:36:55.567Z", "content": "https://github.com/Cadene/pretrained-models.pytorch https://github.com/timesler/facenet-pytorch https://github.com/biubug6/Pytorch_Retinaface https://github.com/lukemelas/EfficientNet-PyTorch https://github.com/HRNet http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/ https://storage.googleapis.com/openimages/web/index.html https://www.cs.tau.ac.il/~wolf/ytfaces https://github.com/NVlabs/ffhq-dataset
", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch https://github.com/timesler/facenet-pytorch https://github.com/biubug6/Pytorch_Retinaface https://github.com/lukemelas/EfficientNet-PyTorch https://github.com/HRNet http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/ https://storage.googleapis.com/openimages/web/index.html https://www.cs.tau.ac.il/~wolf/ytfaces https://github.com/NVlabs/ffhq-dataset\n" }, { "id": 761675, "postDate": "2020-03-02T20:49:52.827Z", "content": "Pytorchcv: https://pypi.org/project/pytorchcv/0.0.13/\nA mobilenet SSD based face detector: https://github.com/yeephycho/tensorflow-face-detection
", "rawMarkdown": "Pytorchcv: https://pypi.org/project/pytorchcv/0.0.13/\nA mobilenet SSD based face detector: https://github.com/yeephycho/tensorflow-face-detection\n" }, { "id": 761655, "postDate": "2020-03-02T20:29:08.713Z", "content": "http://www.robots.ox.ac.uk/~vgg/data/voxceleb/\nDataset and pre-trained models from VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/deepfakes/faceswap\nhttps://github.com/YuvalNirkin/face_swap\nhttps://github.com/ipazc/mtcnn
", "rawMarkdown": "http://www.robots.ox.ac.uk/~vgg/data/voxceleb/\nDataset and pre-trained models from VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/deepfakes/faceswap\nhttps://github.com/YuvalNirkin/face_swap\nhttps://github.com/ipazc/mtcnn" }, { "id": 761639, "postDate": "2020-03-02T19:50:02.973Z", "content": "Face Detection and other packages:\nhttps://github.com/ipazc/mtcnn (wheel: https://www.kaggle.com/unkownhihi/mtcnn-package)\nhttps://www.kaggle.com/robikscube/deepfakemodelspackages
\n\nModel implementations and pretrained model weights:\nhttps://github.com/ondyari/FaceForensics\nhttps://data.lip6.fr\nhttps://github.com/ondyari/FaceForensics\nhttps://data.lip6.fr/cadene/pretrainedmodels/xception-b5690688.pth\nhttp://kaldir.vc.in.tum.de:/FaceForensics/models/faceforensics++_models.zip\nhttps://download.pytorch.org/models/r3d_18-b3b3357e.pth\nhttps://download.pytorch.org/models/r2plus1d_18-91a641e6.pth
", "rawMarkdown": "\nFace Detection and other packages:\nhttps://github.com/ipazc/mtcnn (wheel: https://www.kaggle.com/unkownhihi/mtcnn-package)\nhttps://www.kaggle.com/robikscube/deepfakemodelspackages\n\nModel implementations and pretrained model weights:\nhttps://github.com/ondyari/FaceForensics\nhttps://data.lip6.fr\nhttps://github.com/ondyari/FaceForensics\nhttps://data.lip6.fr/cadene/pretrainedmodels/xception-b5690688.pth\nhttp://kaldir.vc.in.tum.de:/FaceForensics/models/faceforensics++_models.zip\nhttps://download.pytorch.org/models/r3d_18-b3b3357e.pth\nhttps://download.pytorch.org/models/r2plus1d_18-91a641e6.pth\n" }, { "id": 761633, "postDate": "2020-03-02T19:37:39.497Z", "content": "https://www.kaggle.com/rakibilly/ffmpeg-static-build\nhttps://sites.google.com/view/perception-cv4arvr/blazeface\nhttps://github.com/keras-team/keras-applications
", "rawMarkdown": "https://www.kaggle.com/rakibilly/ffmpeg-static-build\nhttps://sites.google.com/view/perception-cv4arvr/blazeface\nhttps://github.com/keras-team/keras-applications" }, { "id": 761627, "postDate": "2020-03-02T19:19:12.137Z", "content": "https://github.com/marl/crepe\nhttps://github.com/santi-pdp/segan
", "rawMarkdown": "https://github.com/marl/crepe\nhttps://github.com/santi-pdp/segan" }, { "id": 761621, "postDate": "2020-03-02T19:03:24.473Z", "content": "https://github.com/sthanhng/yoloface\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Linzaer/Face-Track-Detect-Extract\nhttps://github.com/cc-hpc-itwm/DeepFakeDetection\nhttps://github.com/Xtra-Computing/thundersvm\nhttps://github.com/iperov/DeepFaceLab\nhttps://github.com/deepfakes/faceswap\nhttps://github.com/dessa-research/DeepFake-Detection\nhttps://github.com/VainF/DeepLabV3Plus-Pytorch\nhttps://pytorch.org/hub/pytorch_vision_deeplabv3_resnet101/\nhttps://pytorch.org/hub/pytorch_vision_fcn_resnet101/\nhttps://github.com/SConsul/Global_Convolutional_Network\nhttps://github.com/cutz-j/FDFtNet\nhttps://research.google.com/youtube-bb/\nhttps://research.google.com/youtube8m/index.html\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/121173\nhttps://github.com/DariusAf/MesoNet\nhttps://github.com/nii-yamagishilab/Capsule-Forensics-v2\nhttps://github.com/danmohaha/DSP-FWA\nhttps://github.com/mvaleriani/Shallow\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nhttps://github.com/grib0ed0v/face_recognition.pytorch\nhttps://github.com/nii-yamagishilab/ClassNSeg\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pypi.org/project/face-recognition/\nhttps://github.com/microsoft/LightGBM\nhttps://github.com/dmlc/xgboost\nhttps://github.com/adobe/antialiased-cnns\nhttps://github.com/amilworks/GanDetection\nhttps://github.com/tensorflow/models
", "rawMarkdown": "https://github.com/sthanhng/yoloface\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Linzaer/Face-Track-Detect-Extract\nhttps://github.com/cc-hpc-itwm/DeepFakeDetection\nhttps://github.com/Xtra-Computing/thundersvm\nhttps://github.com/iperov/DeepFaceLab\nhttps://github.com/deepfakes/faceswap\nhttps://github.com/dessa-research/DeepFake-Detection\nhttps://github.com/VainF/DeepLabV3Plus-Pytorch\nhttps://pytorch.org/hub/pytorch_vision_deeplabv3_resnet101/\nhttps://pytorch.org/hub/pytorch_vision_fcn_resnet101/\nhttps://github.com/SConsul/Global_Convolutional_Network\nhttps://github.com/cutz-j/FDFtNet\nhttps://research.google.com/youtube-bb/\nhttps://research.google.com/youtube8m/index.html\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/121173\nhttps://github.com/DariusAf/MesoNet\nhttps://github.com/nii-yamagishilab/Capsule-Forensics-v2\nhttps://github.com/danmohaha/DSP-FWA\nhttps://github.com/mvaleriani/Shallow\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nhttps://github.com/grib0ed0v/face_recognition.pytorch\nhttps://github.com/nii-yamagishilab/ClassNSeg\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pypi.org/project/face-recognition/\nhttps://github.com/microsoft/LightGBM\nhttps://github.com/dmlc/xgboost\nhttps://github.com/adobe/antialiased-cnns\nhttps://github.com/amilworks/GanDetection\nhttps://github.com/tensorflow/models" }, { "id": 761587, "postDate": "2020-03-02T17:46:24.400Z", "content": "Playing around with hybrid quantum-classical computations: \nhttps://github.com/XanaduAI/PennyLane\nhttps://github.com/XanaduAI/pennylane-qiskit (edited)
", "rawMarkdown": "Playing around with hybrid quantum-classical computations: \nhttps://github.com/XanaduAI/PennyLane\nhttps://github.com/XanaduAI/pennylane-qiskit (edited)" }, { "id": 761547, "postDate": "2020-03-02T16:41:16.063Z", "content": "https://github.com/stanfordnlp/mac-network
", "rawMarkdown": "https://github.com/stanfordnlp/mac-network" }, { "id": 761528, "postDate": "2020-03-02T16:18:26.130Z", "content": "https://github.com/PeterWang512/CNNDetection
", "rawMarkdown": "https://github.com/PeterWang512/CNNDetection" }, { "id": 761522, "postDate": "2020-03-02T16:12:49.260Z", "content": "Xception: https://www.dropbox.com/s/1hplpzet9d7dv29/xception-c0a72b38.pth.tar?dl=1\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nDeepfakes Inference Demo: https://www.kaggle.com/humananalog/deepfakes-inference-demo\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nEDIT. Not used: RetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nEDIT. Not used:FaceDetection-DSFD: https://github.com/TencentYoutuResearch/FaceDetection-DSFD\nEDIT. Not used:lightDSFD: https://github.com/lijiannuist/lightDSFD\nEDIT. Not used:FaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nEDIT. Not used:DeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html
\n\nImageNet pre-trained models in pytorchcv: https://pypi.org/project/pytorchcv/\nImageNet pre-trained models in PyTorch Hub: https://pytorch.org/hub/\npythonspeechfeatures: https://pypi.org/project/python_speech_features/0.4/\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/tengshaofeng/ResidualAttentionNetwork-pytorch\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch
", "rawMarkdown": "Xception: https://www.dropbox.com/s/1hplpzet9d7dv29/xception-c0a72b38.pth.tar?dl=1\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nDeepfakes Inference Demo: https://www.kaggle.com/humananalog/deepfakes-inference-demo\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nEDIT. Not used: RetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nEDIT. Not used:FaceDetection-DSFD: https://github.com/TencentYoutuResearch/FaceDetection-DSFD\nEDIT. Not used:lightDSFD: https://github.com/lijiannuist/lightDSFD\nEDIT. Not used:FaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nEDIT. Not used:DeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html\n\nImageNet pre-trained models in pytorchcv: https://pypi.org/project/pytorchcv/\nImageNet pre-trained models in PyTorch Hub: https://pytorch.org/hub/\npythonspeechfeatures: https://pypi.org/project/python_speech_features/0.4/\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/tengshaofeng/ResidualAttentionNetwork-pytorch\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch" }, { "id": 761501, "postDate": "2020-03-02T15:48:07.443Z", "content": "Youtube-df\nhttp://cs.uef.fi/deepfake_dataset/
", "rawMarkdown": "Youtube-df\nhttp://cs.uef.fi/deepfake_dataset/" }, { "id": 761494, "postDate": "2020-03-02T15:33:40.333Z", "content": "ImageNet\nJFT dataset\nWIDERFace\nhttps://github.com/tensorflow/models/ and corresponding training datasets\nhttps://pytorch.org/docs/stable/torchvision/models.html and corresponding training datasets\nhttps://github.com/rwightman/pytorch-image-models and corresponding training datasets\nhttps://github.com/rwightman/gen-efficientnet-pytorch and corresponding training datasets\nhttps://github.com/iitzco/faced and corresponding training datasets\nRetinaFace : https://github.com/deepinsight/insightface/tree/master/RetinaFace,\nhttps://github.com/biubug6/Pytorch_Retinaface and corresponding training datasets\nFacenet: https://github.com/davidsandberg/facenet, https://github.com/timesler/facenet-pytorch and corresponding training datasets\nhttps://github.com/open-mmlab/mmdetection
", "rawMarkdown": "ImageNet\nJFT dataset\nWIDERFace\nhttps://github.com/tensorflow/models/ and corresponding training datasets\nhttps://pytorch.org/docs/stable/torchvision/models.html and corresponding training datasets\nhttps://github.com/rwightman/pytorch-image-models and corresponding training datasets\nhttps://github.com/rwightman/gen-efficientnet-pytorch and corresponding training datasets\nhttps://github.com/iitzco/faced and corresponding training datasets\nRetinaFace : https://github.com/deepinsight/insightface/tree/master/RetinaFace,\nhttps://github.com/biubug6/Pytorch_Retinaface and corresponding training datasets\nFacenet: https://github.com/davidsandberg/facenet, https://github.com/timesler/facenet-pytorch and corresponding training datasets\nhttps://github.com/open-mmlab/mmdetection" }, { "id": 761488, "postDate": "2020-03-02T15:24:27.453Z", "content": "Stingray: https://github.com/StingraySoftware/stingray (MIT license)
", "rawMarkdown": "Stingray: https://github.com/StingraySoftware/stingray (MIT license)" }, { "id": 761479, "postDate": "2020-03-02T15:07:59.640Z", "content": "AudioSet: https://research.google.com/audioset/dataset/index.html\nUAFDV: https://drive.google.com/drive/folders/1GEk1DSxmlV_61JtpEGzC9Fo_BffvyxpH\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics/blob/master/Celeb-DF-v2/README.md\nFaceForensics++: https://github.com/ondyari/FaceForensics\n300W: https://ibug.doc.ic.ac.uk/resources/300-W/
", "rawMarkdown": "AudioSet: https://research.google.com/audioset/dataset/index.html\nUAFDV: https://drive.google.com/drive/folders/1GEk1DSxmlV_61JtpEGzC9Fo_BffvyxpH\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics/blob/master/Celeb-DF-v2/README.md\nFaceForensics++: https://github.com/ondyari/FaceForensics\n300W: https://ibug.doc.ic.ac.uk/resources/300-W/", "replies": [ { "id": 761539, "postDate": "2020-03-02T16:33:08.670Z", "content": "I may use , \npose_300W_LP dataset based https://github.com/natanielruiz/deep-head-pose\nhttps://github.com/OverEuro/deep-head-pose-lite
", "rawMarkdown": "I may use , \npose_300W_LP dataset based [https://github.com/natanielruiz/deep-head-pose](https://github.com/natanielruiz/deep-head-pose)\n[https://github.com/OverEuro/deep-head-pose-lite](https://github.com/OverEuro/deep-head-pose-lite)" } ] }, { "id": 761446, "postDate": "2020-03-02T14:25:13.323Z", "content": "https://github.com/deepfakeinthewild/deepfake_in_the_wild seems to be really interesting but it is available for download only with Baidu account... which clearly violates any competition rules
", "rawMarkdown": "https://github.com/deepfakeinthewild/deepfake_in_the_wild seems to be really interesting but it is available for download only with Baidu account... which clearly violates any competition rules", "votes": 2 } ] }, { "id": 761381, "postDate": "2020-03-02T12:58:12.697Z", "content": "mtcnn package:\nhttps://www.kaggle.com/unkownhihi/mtcnn-package\nffmpeg static build:\nhttps://www.kaggle.com/rakibilly/ffmpeg-static-build\ninception_resnet_v2:\nhttps://github.com/tensorflow/models/blob/master/research/slim/nets/inception_resnet_v2.py\nDCGAN:\nhttps://github.com/tensorflow/models/blob/master/research/slim/nets/dcgan.py\nCycleGAN:\nhttps://github.com/tensorflow/models/blob/master/research/slim/nets/cyclegan.py\nOpenCV Haarcascades:\nhttps://github.com/opencv/opencv/tree/master/data/haarcascades\nRealtime Glasses Detection:\nhttps://github.com/TianxingWu/realtime-glasses-detection\nKeras Applications:\nhttps://www.tensorflow.org/api_docs/python/tf/keras/applications/\nMobileNet face extractor:\nhttps://github.com/yeephycho/tensorflow-face-detection/tree/master/model
", "rawMarkdown": "mtcnn package:\n[https://www.kaggle.com/unkownhihi/mtcnn-package](https://www.kaggle.com/unkownhihi/mtcnn-package)\nffmpeg static build:\n[https://www.kaggle.com/rakibilly/ffmpeg-static-build](https://www.kaggle.com/rakibilly/ffmpeg-static-build)\ninception_resnet_v2:\n[https://github.com/tensorflow/models/blob/master/research/slim/nets/inception_resnet_v2.py](https://github.com/tensorflow/models/blob/master/research/slim/nets/inception_resnet_v2.py)\nDCGAN:\n[https://github.com/tensorflow/models/blob/master/research/slim/nets/dcgan.py](https://github.com/tensorflow/models/blob/master/research/slim/nets/dcgan.py)\nCycleGAN:\n[https://github.com/tensorflow/models/blob/master/research/slim/nets/cyclegan.py](https://github.com/tensorflow/models/blob/master/research/slim/nets/cyclegan.py)\nOpenCV Haarcascades:\n[https://github.com/opencv/opencv/tree/master/data/haarcascades](https://github.com/opencv/opencv/tree/master/data/haarcascades)\nRealtime Glasses Detection:\n[https://github.com/TianxingWu/realtime-glasses-detection](https://github.com/TianxingWu/realtime-glasses-detection)\nKeras Applications:\n[https://www.tensorflow.org/api_docs/python/tf/keras/applications/](https://www.tensorflow.org/api_docs/python/tf/keras/applications/)\nMobileNet face extractor:\nhttps://github.com/yeephycho/tensorflow-face-detection/tree/master/model" }, { "id": 761369, "postDate": "2020-03-02T12:48:58.250Z", "content": "librosa: https://github.com/librosa/librosa (ISC license)
", "rawMarkdown": "librosa: https://github.com/librosa/librosa (ISC license)" }, { "id": 761356, "postDate": "2020-03-02T12:28:49.053Z", "content": "FaceForensics+: https://github.com/ondyari/FaceForensics\nmodel\": https://data.lip6.fr/cadene/pretrainedmodels/xception-b5690688.pth\nmodel weights: : http://kaldir.vc.in.tum.de:/FaceForensics/models/faceforensics++_models.zip
", "rawMarkdown": "FaceForensics+: https://github.com/ondyari/FaceForensics\nmodel\": https://data.lip6.fr/cadene/pretrainedmodels/xception-b5690688.pth\nmodel weights: : http://kaldir.vc.in.tum.de:/FaceForensics/models/faceforensics++_models.zip" }, { "id": 761319, "postDate": "2020-03-02T11:39:45.333Z", "content": "The TIMIT dataset, reportedly used by the elite in this competition, is not available without a '.edu' or similar email address.\nAnd thus, should NOT be eligible to this competition.\nBTW : What's all this mess about availability of the data from the authors ???
", "rawMarkdown": "The TIMIT dataset, reportedly used by the elite in this competition, is not available without a '.edu' or similar email address.\nAnd thus, should NOT be eligible to this competition.\nBTW : What's all this mess about availability of the data from the authors ???" }, { "id": 761283, "postDate": "2020-03-02T11:01:50.500Z", "content": "https://github.com/biubug6/Pytorch_Retinaface
", "rawMarkdown": "https://github.com/biubug6/Pytorch_Retinaface\n" }, { "id": 761264, "postDate": "2020-03-02T10:27:28.543Z", "content": "https://github.com/open-mmlab/mmcv\nhttps://github.com/opencv/dldt - OpenVINO\nhttps://github.com/opencv/open_model_zoo
", "rawMarkdown": "https://github.com/open-mmlab/mmcv\nhttps://github.com/opencv/dldt - OpenVINO\nhttps://github.com/opencv/open_model_zoo" }, { "id": 761254, "postDate": "2020-03-02T10:14:46.817Z", "content": "https://github.com/rcmalli/keras-vggface/ \n(with all models)
\n\nhttps://github.com/the-house-of-black-and-white/hall-of-faces
", "rawMarkdown": "https://github.com/rcmalli/keras-vggface/ \n(with all models)\n\nhttps://github.com/the-house-of-black-and-white/hall-of-faces" }, { "id": 761249, "postDate": "2020-03-02T10:04:49.760Z", "content": "TensorFlow Speech Recognition Challenge data: https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/data\nRetinaFace MobileNet0.25: https://github.com/deepinsight/insightface/issues/669
", "rawMarkdown": "TensorFlow Speech Recognition Challenge data: https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/data\nRetinaFace MobileNet0.25: https://github.com/deepinsight/insightface/issues/669" }, { "id": 761239, "postDate": "2020-03-02T09:56:40.237Z", "content": "face_recognition: https://github.com/ageitgey/face_recognition\nopenface: https://github.com/cmusatyalab/openface\nDeep Residual Learning for Image Recognition: deep-residuahttps://github.com/KaimingHe/deep-residual-networks\nLFW dataset: http://vis-www.cs.umass.edu/lfw/
", "rawMarkdown": "face_recognition: https://github.com/ageitgey/face_recognition\nopenface: https://github.com/cmusatyalab/openface\nDeep Residual Learning for Image Recognition: deep-residuahttps://github.com/KaimingHe/deep-residual-networks\nLFW dataset: http://vis-www.cs.umass.edu/lfw/\n" }, { "id": 761228, "postDate": "2020-03-02T09:39:42.783Z", "content": "vggface : https://github.com/rcmalli/keras-vggface
", "rawMarkdown": "vggface : https://github.com/rcmalli/keras-vggface" }, { "id": 761182, "postDate": "2020-03-02T08:21:27.970Z", "content": "EfficientNet-pyTorch: https://github.com/lukemelas/EfficientNet-PyTorch \nresnext101: https://github.com/facebookresearch/WSL-Images \nface.evoLVe.PyTorch: https://github.com/ZhaoJ9014/face.evoLVe.PyTorch \nInsightface:https://github.com/deepinsight/insightface\nhrnet: https://github.com/HRNet/HRNet-Image-Classification \ntorchvision models: https://github.com/pytorch/vision/tree/master/torchvision/models \nCennternet:https://github.com/xingyizhou/CenterNet \nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nRetinafacePytorch:https://github.com/biubug6/Pytorch_Retinaface\nFaceForensics++ dataset: https://github.com/ondyari/FaceForensics \nFFMPEG: https://johnvansickle.com/ffmpeg/\nTensorRT: https://github.com/NVIDIA/TensorRT \nOnnx: https://github.com/onnx/onnx\nOnnx-tensorRT: https://github.com/onnx/onnx-tensorrt
", "rawMarkdown": "EfficientNet-pyTorch: https://github.com/lukemelas/EfficientNet-PyTorch \nresnext101: https://github.com/facebookresearch/WSL-Images \nface.evoLVe.PyTorch: https://github.com/ZhaoJ9014/face.evoLVe.PyTorch \nInsightface:https://github.com/deepinsight/insightface\nhrnet: https://github.com/HRNet/HRNet-Image-Classification \ntorchvision models: https://github.com/pytorch/vision/tree/master/torchvision/models \nCennternet:https://github.com/xingyizhou/CenterNet \nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nRetinafacePytorch:https://github.com/biubug6/Pytorch_Retinaface\nFaceForensics++ dataset: https://github.com/ondyari/FaceForensics \nFFMPEG: https://johnvansickle.com/ffmpeg/\nTensorRT: https://github.com/NVIDIA/TensorRT \nOnnx: https://github.com/onnx/onnx\nOnnx-tensorRT: https://github.com/onnx/onnx-tensorrt\n" }, { "id": 761162, "postDate": "2020-03-02T07:49:41.773Z", "content": "TensorRT: https://github.com/NVIDIA/TensorRT TensorRt-6.0.1.5.Ubuntu-14.04.x86_64-gnu.cuda-10.0.cudnn7.6.tar.gz\nOnnx: https://github.com/onnx/onnx\nOnnx-tensorRT: https://github.com/onnx/onnx-tensorrt
", "rawMarkdown": "TensorRT: [https://github.com/NVIDIA/TensorRT](https://github.com/NVIDIA/TensorRT) TensorRt-6.0.1.5.Ubuntu-14.04.x86_64-gnu.cuda-10.0.cudnn7.6.tar.gz\nOnnx: [https://github.com/onnx/onnx](https://github.com/onnx/onnx)\nOnnx-tensorRT: [https://github.com/onnx/onnx-tensorrt](https://github.com/onnx/onnx-tensorrt) " }, { "id": 761119, "postDate": "2020-03-02T06:54:58.390Z", "content": "MTCNN: https://pypi.org/project/mtcnn/\nFacenet: https://github.com/nyoki-mtl/keras-facenet
", "rawMarkdown": "MTCNN: https://pypi.org/project/mtcnn/\nFacenet: https://github.com/nyoki-mtl/keras-facenet\n" }, { "id": 761088, "postDate": "2020-03-02T05:51:49.510Z", "content": "FFMPEG: https://johnvansickle.com/ffmpeg/
\n\nFacenet: https://github.com/timesler/facenet-pytorch
\n\nResnext: https://pytorch.org/hub/facebookresearch_WSL-Images_resnext/
\n\nPytorch torch vision models https://pytorch.org/docs/stable/torchvision/models.html
\n\nNvidia DALI https://github.com/NVIDIA/DALI
", "rawMarkdown": "FFMPEG: https://johnvansickle.com/ffmpeg/\n\nFacenet: https://github.com/timesler/facenet-pytorch\n\nResnext: https://pytorch.org/hub/facebookresearch_WSL-Images_resnext/\n\nPytorch torch vision models https://pytorch.org/docs/stable/torchvision/models.html\n\nNvidia DALI https://github.com/NVIDIA/DALI" }, { "id": 761044, "postDate": "2020-03-02T04:21:16.217Z", "content": "https://gluon-cv.mxnet.io/ and their model zoo https://gluon-cv.mxnet.io/model_zoo/index.html including ImageNet (still not sure if this is allowed) pre-trained models\nfilterpy http://github.com/rlabbe/filterpy
", "rawMarkdown": "https://gluon-cv.mxnet.io/ and their model zoo https://gluon-cv.mxnet.io/model_zoo/index.html including ImageNet (still not sure if this is allowed) pre-trained models\nfilterpy http://github.com/rlabbe/filterpy\n" }, { "id": 761034, "postDate": "2020-03-02T04:03:11.603Z", "content": "youtube videos ds: https://deepfake-detection.s3.amazonaws.com/augment_deepfake.tar.gz
", "rawMarkdown": "youtube videos ds: https://deepfake-detection.s3.amazonaws.com/augment_deepfake.tar.gz" }, { "id": 761027, "postDate": "2020-03-02T03:42:36.157Z", "content": "pre-trained pytorch YOLO object detection https://github.com/eriklindernoren/PyTorch-YOLOv3
", "rawMarkdown": "pre-trained pytorch YOLO object detection https://github.com/eriklindernoren/PyTorch-YOLOv3" }, { "id": 760961, "postDate": "2020-03-02T01:06:36.393Z", "content": "MMSkeleton and some dependencies\nhttps://github.com/open-mmlab/mmskeleton\nhttps://pypi.org/project/addict/\nhttps://pypi.org/project/mmcv/\nhttps://pypi.org/project/lazy-import/\nhttps://github.com/cocodataset/cocoapi
", "rawMarkdown": "MMSkeleton and some dependencies\nhttps://github.com/open-mmlab/mmskeleton\nhttps://pypi.org/project/addict/\nhttps://pypi.org/project/mmcv/\nhttps://pypi.org/project/lazy-import/\nhttps://github.com/cocodataset/cocoapi", "replies": [ { "id": 762890, "postDate": "2020-03-03T22:59:12.873Z", "content": "https://github.com/mewwts/addict\nhttps://github.com/mnmelo/lazy_import
", "rawMarkdown": "https://github.com/mewwts/addict\nhttps://github.com/mnmelo/lazy_import" } ] }, { "id": 760957, "postDate": "2020-03-02T00:59:10.010Z", "content": "https://github.com/deepmind/sonnet
", "rawMarkdown": "https://github.com/deepmind/sonnet" }, { "id": 760852, "postDate": "2020-03-01T19:54:23.753Z", "content": "Models & weights from:\n* https://github.com/keras-team/keras-applications\n* https://github.com/qubvel/efficientnet\n* https://github.com/ipazc/mtcnn\n* https://github.com/Star-Clouds/CenterFace
", "rawMarkdown": "Models & weights from:\n* https://github.com/keras-team/keras-applications\n* https://github.com/qubvel/efficientnet\n* https://github.com/ipazc/mtcnn\n* https://github.com/Star-Clouds/CenterFace" }, { "id": 760848, "postDate": "2020-03-01T19:50:37.957Z", "content": "https://www.kaggle.com/sheldonrobinson/video-tools
", "rawMarkdown": "https://www.kaggle.com/sheldonrobinson/video-tools" }, { "id": 760846, "postDate": "2020-03-01T19:47:35.943Z", "content": "https://www.kaggle.com/pranavpulijala/moviepy
", "rawMarkdown": "https://www.kaggle.com/pranavpulijala/moviepy" }, { "id": 760715, "postDate": "2020-03-01T16:41:28.880Z", "content": "dLib Library: http://dlib.net https://www.kaggle.com/carlossouza/dlibpkg
\n\nKeras InceptionResNetV2: https://keras.io/applications
\n\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch
\n\nyolov3: https://github.com/ultralytics/yolov3
\n\nface_recognition: https://github.com/ageitgey/face_recognition
\n\nFaceForensics: https://github.com/ondyari/FaceForensics
\n\nreal-and-fake-face-detection: https://www.kaggle.com/ciplab/real-and-fake-face-detection
", "rawMarkdown": "dLib Library: http://dlib.net https://www.kaggle.com/carlossouza/dlibpkg\n\nKeras InceptionResNetV2: https://keras.io/applications\n\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\n\nyolov3: https://github.com/ultralytics/yolov3\n\nface_recognition: https://github.com/ageitgey/face_recognition\n\nFaceForensics: https://github.com/ondyari/FaceForensics\n\nreal-and-fake-face-detection: https://www.kaggle.com/ciplab/real-and-fake-face-detection" }, { "id": 760692, "postDate": "2020-03-01T16:07:59.940Z", "content": "EfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nDeepfakes Inference Demo: https://www.kaggle.com/humananalog/deepfakes-inference-demo\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nRetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nFaceDetection-DSFD: https://github.com/TencentYoutuResearch/FaceDetection-DSFD\nlightDSFD: https://github.com/lijiannuist/lightDSFD\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nDeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html\nWIDER FACE Dataset: http://shuoyang1213.me/WIDERFACE/
", "rawMarkdown": "EfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nDeepfakes Inference Demo: https://www.kaggle.com/humananalog/deepfakes-inference-demo\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nRetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nFaceDetection-DSFD: https://github.com/TencentYoutuResearch/FaceDetection-DSFD\nlightDSFD: https://github.com/lijiannuist/lightDSFD\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nDeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html\nWIDER FACE Dataset: http://shuoyang1213.me/WIDERFACE/" }, { "id": 760636, "postDate": "2020-03-01T14:57:46.060Z", "content": "Kaggle Facenet : https://www.kaggle.com/timesler/facenet-pytorch-vggface2
", "rawMarkdown": "Kaggle Facenet : https://www.kaggle.com/timesler/facenet-pytorch-vggface2" }, { "id": 760612, "postDate": "2020-03-01T14:23:22.847Z", "content": "https://github.com/foolwood/DaSiamRPN\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace
", "rawMarkdown": "https://github.com/foolwood/DaSiamRPN\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace" }, { "id": 760609, "postDate": "2020-03-01T14:21:41.273Z", "content": "https://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/foolwood/DaSiamRPN
", "rawMarkdown": "https://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/foolwood/DaSiamRPN" }, { "id": 760552, "postDate": "2020-03-01T13:07:49.110Z", "content": "https://github.com/thomasbrandon/mish-cuda\nhttps://developer.nvidia.com/tensorrt\nhttps://opencv.org/
", "rawMarkdown": "https://github.com/thomasbrandon/mish-cuda\nhttps://developer.nvidia.com/tensorrt\nhttps://opencv.org/" }, { "id": 760540, "postDate": "2020-03-01T12:42:09.517Z", "content": "@juliaelliott @addisonhoward \nI'm a bit confused about usage of external data. For example there is a public available dataset. It has license which implies only research usage(no commercial). I'm not using this dataset directly in my experiments, but I'm using a trained model on this dataset which is public available on GitHub. It will be violation of the rules or not?
", "rawMarkdown": "@juliaelliott @addisonhoward \nI'm a bit confused about usage of external data. For example there is a public available dataset. It has license which implies only research usage(no commercial). I'm not using this dataset directly in my experiments, but I'm using a trained model on this dataset which is public available on GitHub. It will be violation of the rules or not?", "replies": [ { "id": 761666, "postDate": "2020-03-02T20:39:08.270Z", "content": "If your model makes use of any external data that is prohibited by the rules of the competition, in training or otherwise, then you are subject to disqualification.
", "rawMarkdown": "If your model makes use of any external data that is prohibited by the rules of the competition, in training or otherwise, then you are subject to disqualification." }, { "id": 761677, "postDate": "2020-03-02T20:50:38.167Z", "content": "Ok, thanks!
", "rawMarkdown": "Ok, thanks!" } ] }, { "id": 760514, "postDate": "2020-03-01T11:58:56.367Z", "content": "MIT License\nhttps://pypi.org/project/mtcnn/\nFacenet: https://github.com/timesler/facenet-pytorch\nFFMPEG: https://johnvansickle.com/ffmpeg/\nResnext: https://pytorch.org/hub/facebookresearch_WSL-Images_resnext/\nBlazeface: https://github.com/hollance/BlazeFace-PyTorch\nhttps://opencv.org/\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nDeepfakes Inference Demo: https://www.kaggle.com/humananalog/deepfakes-inference-demo\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nRetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nFaceDetection-DSFD: https://github.com/TencentYoutuResearch/FaceDetection-DSFD\nlightDSFD: https://github.com/lijiannuist/lightDSFD\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nDeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html\nWIDER FACE Dataset: http://shuoyang1213.me/WIDERFACE/
\n\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch
\n\nhttps://github.com/yaojieliu/CVPR2019-DeepTreeLearningForZeroShotFaceAntispoofing\nhttps://github.com/yaojieliu/ECCV2018-FaceDeSpoofing
\n\nhttps://github.com/EndlessSora/DeeperForensics-1.0\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/
\n\nUADFV: The UADFV dataset - Xin Yang, Yuezun Li, and Siwei Lyu. Exposing deep fakes\nusing inconsistent head poses. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),\n2019.
\n\nDF-TIMIT: Pavel Korshunov and Sebastien Marcel. Deepfakes: a new ´\nthreat to face recognition? assessment and detection. arXiv\npreprint arXiv:1812.08685, 2018.
\n\nFF-DF: Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Chris- ¨\ntian Riess, Justus Thies, and Matthias Nießner. FaceForensics++: Learning to detect manipulated facial images. In\nICCV, 2019.
\n\nDFD: Nicholas Dufour, Andrew Gully, Per Karlsson, Alexey Victor Vorbyov, Thomas Leung, Jeremiah Childs, and Christoph Bregler. Deepfakes detection dataset by google & jigsaw.
\n\nhttps://github.com/dessa-research/DeepFake-Detection
\n\ndlib: https://github.com/davisking/dlib\ndlib-models: https://github.com/davisking/dlib-models/blob/master/shape_predictor_68_face_landmarks.dat.bz2\nnumpy: https://github.com/numpy/numpy\nopencv: https://github.com/opencv/opencv\nscipy: https://github.com/scipy/scipy\nimutils: https://pypi.org/project/imutils/\nripser: https://github.com/scikit-tda/ripser.py\nglob: https://docs.python.org/3/library/glob.html\nos: https://docs.python.org/3/library/os.html\nmultiprocessing: https://docs.python.org/3/library/multiprocessing.html\nscikit-learn: https://scikit-learn.org/stable/index.html\nScikit-TDA: https://github.com/scikit-tda\ncsv: https://docs.python.org/3/library/csv.html
", "rawMarkdown": "MIT License\nhttps://pypi.org/project/mtcnn/\nFacenet: https://github.com/timesler/facenet-pytorch\nFFMPEG: https://johnvansickle.com/ffmpeg/\nResnext: https://pytorch.org/hub/facebookresearch_WSL-Images_resnext/\nBlazeface: https://github.com/hollance/BlazeFace-PyTorch\nhttps://opencv.org/\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nDeepfakes Inference Demo: https://www.kaggle.com/humananalog/deepfakes-inference-demo\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nRetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nFaceDetection-DSFD: https://github.com/TencentYoutuResearch/FaceDetection-DSFD\nlightDSFD: https://github.com/lijiannuist/lightDSFD\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nDeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html\nWIDER FACE Dataset: http://shuoyang1213.me/WIDERFACE/\n\n\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\n\nhttps://github.com/yaojieliu/CVPR2019-DeepTreeLearningForZeroShotFaceAntispoofing\nhttps://github.com/yaojieliu/ECCV2018-FaceDeSpoofing\n\nhttps://github.com/EndlessSora/DeeperForensics-1.0\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/\n\nUADFV: The UADFV dataset - Xin Yang, Yuezun Li, and Siwei Lyu. Exposing deep fakes\nusing inconsistent head poses. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),\n2019.\n\nDF-TIMIT: Pavel Korshunov and Sebastien Marcel. Deepfakes: a new ´\nthreat to face recognition? assessment and detection. arXiv\npreprint arXiv:1812.08685, 2018.\n\nFF-DF: Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Chris- ¨\ntian Riess, Justus Thies, and Matthias Nießner. FaceForensics++: Learning to detect manipulated facial images. In\nICCV, 2019.\n\nDFD: Nicholas Dufour, Andrew Gully, Per Karlsson, Alexey Victor Vorbyov, Thomas Leung, Jeremiah Childs, and Christoph Bregler. Deepfakes detection dataset by google & jigsaw.\n\nhttps://github.com/dessa-research/DeepFake-Detection\n\ndlib: https://github.com/davisking/dlib\ndlib-models: https://github.com/davisking/dlib-models/blob/master/shape_predictor_68_face_landmarks.dat.bz2\nnumpy: https://github.com/numpy/numpy\nopencv: https://github.com/opencv/opencv\nscipy: https://github.com/scipy/scipy\nimutils: https://pypi.org/project/imutils/\nripser: https://github.com/scikit-tda/ripser.py\nglob: https://docs.python.org/3/library/glob.html\nos: https://docs.python.org/3/library/os.html\nmultiprocessing: https://docs.python.org/3/library/multiprocessing.html\nscikit-learn: https://scikit-learn.org/stable/index.html\nScikit-TDA: https://github.com/scikit-tda\ncsv: https://docs.python.org/3/library/csv.html" }, { "id": 760509, "postDate": "2020-03-01T11:42:33.390Z", "content": "Facenet: https://github.com/timesler/facenet-pytorch\nFFMPEG: https://johnvansickle.com/ffmpeg/\nResnext: https://pytorch.org/hub/facebookresearch_WSL-Images_resnext/\nBlazeface: https://github.com/hollance/BlazeFace-PyTorch
", "rawMarkdown": "Facenet: https://github.com/timesler/facenet-pytorch\nFFMPEG: https://johnvansickle.com/ffmpeg/\nResnext: https://pytorch.org/hub/facebookresearch_WSL-Images_resnext/\nBlazeface: https://github.com/hollance/BlazeFace-PyTorch\n" }, { "id": 760429, "postDate": "2020-03-01T09:23:19.487Z", "content": "Pytorch implementation of xception net with ImageNet pretrained weights by tstandley. [MIT License]\nModel: https://github.com/tstandley/Xception-PyTorch\nWeight: https://www.dropbox.com/s/1hplpzet9d7dv29/xception-c0a72b38.pth.tar?dl=1
\n\nPytorch VGG16 pretrained weights:\nhttps://download.pytorch.org/models/vgg19-dcbb9e9d.pth
", "rawMarkdown": "Pytorch implementation of xception net with ImageNet pretrained weights by tstandley. [MIT License]\nModel: https://github.com/tstandley/Xception-PyTorch\nWeight: https://www.dropbox.com/s/1hplpzet9d7dv29/xception-c0a72b38.pth.tar?dl=1\n\nPytorch VGG16 pretrained weights:\nhttps://download.pytorch.org/models/vgg19-dcbb9e9d.pth", "replies": [ { "id": 762190, "postDate": "2020-03-03T09:53:02.107Z", "content": "Yi, you provided the wrong link for VGG16 weight. Correct it before the deadline.
", "rawMarkdown": "Yi, you provided the wrong link for VGG16 weight. Correct it before the deadline." }, { "id": 776775, "postDate": "2020-03-17T16:17:57.487Z", "content": "[Edit March 18]\nSorry. The link is correct, but there is a typo in the text description. It should be VGG19 pretrained weights.
", "rawMarkdown": "[Edit March 18]\nSorry. The link is correct, but there is a typo in the text description. It should be VGG19 pretrained weights." } ] }, { "id": 760421, "postDate": "2020-03-01T09:12:05.900Z", "content": "https://pypi.org/project/librosa/
", "rawMarkdown": "https://pypi.org/project/librosa/" }, { "id": 760396, "postDate": "2020-03-01T08:17:48.400Z", "content": "https://www.cs.tau.ac.il/~wolf/ytfaces/\nhttps://www.kaggle.com/sophatvathana/casia-dataset#Sp_D_CND_A_sec0056_sec0015_0282.jpg\nhttps://github.com/Daniil-Osokin/lightweight-human-pose-estimation.pytorch\nhttps://github.com/HRNet\nhttps://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/MerHS/SASA-pytorch\nhttps://github.com/PeterWang512/CNNDetection\nhttps://github.com/tensorflow/magenta
", "rawMarkdown": "https://www.cs.tau.ac.il/~wolf/ytfaces/\nhttps://www.kaggle.com/sophatvathana/casia-dataset#Sp_D_CND_A_sec0056_sec0015_0282.jpg\nhttps://github.com/Daniil-Osokin/lightweight-human-pose-estimation.pytorch\nhttps://github.com/HRNet\nhttps://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/MerHS/SASA-pytorch\nhttps://github.com/PeterWang512/CNNDetection\nhttps://github.com/tensorflow/magenta" }, { "id": 760349, "postDate": "2020-03-01T06:45:08.953Z", "content": "Keras EfficientNet Noisy Student Weights as referred here::\nhttps://www.kaggle.com/c/bengaliai-cv19/discussion/132894
\n\nMay try : \nhttps://github.com/chen0040/keras-video-classifier\nhttps://github.com/sagarvegad/Video-Classification-CNN-and-LSTM-
", "rawMarkdown": "Keras EfficientNet Noisy Student Weights as referred here::\nhttps://www.kaggle.com/c/bengaliai-cv19/discussion/132894\n\nMay try : \nhttps://github.com/chen0040/keras-video-classifier\nhttps://github.com/sagarvegad/Video-Classification-CNN-and-LSTM-" }, { "id": 760324, "postDate": "2020-03-01T06:05:26.870Z", "content": "https://pypi.org/project/moviepy/
", "rawMarkdown": "https://pypi.org/project/moviepy/" }, { "id": 760167, "postDate": "2020-02-29T23:36:46.230Z", "content": "https://github.com/nii-yamagishilab/Capsule-Forensics-v2
", "rawMarkdown": "https://github.com/nii-yamagishilab/Capsule-Forensics-v2" }, { "id": 760166, "postDate": "2020-02-29T23:36:41.510Z", "content": "WIDER FACE: http://shuoyang1213.me/WIDERFACE/ (not allowed, no longer using)\nmmdetection: https://github.com/open-mmlab/mmdetection
\n\nEdited to add (March 1, 2020):\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch
", "rawMarkdown": "*WIDER FACE: http://shuoyang1213.me/WIDERFACE/ (not allowed, no longer using)*\nmmdetection: https://github.com/open-mmlab/mmdetection\n\nEdited to add (March 1, 2020):\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch", "replies": [ { "id": 760414, "postDate": "2020-03-01T08:57:45.667Z", "content": "I wanted to use Wider Face as well, but it is not allowed, as by terms it is for non-commercial purposes https://wider-challenge.org/terms_and_conditions_2018.html
", "rawMarkdown": "I wanted to use Wider Face as well, but it is not allowed, as by terms it is for non-commercial purposes https://wider-challenge.org/terms_and_conditions_2018.html", "votes": 2 }, { "id": 760766, "postDate": "2020-03-01T17:46:50.913Z", "content": "Good to know, will not use it. Thanks!
", "rawMarkdown": "Good to know, will not use it. Thanks!" } ] }, { "id": 760147, "postDate": "2020-02-29T22:40:09.413Z", "content": "facenet-pytorch: https://github.com/timesler/facenet-pytorch\nMesoNet: https://github.com/DariusAf/MesoNet\nPytorch xception net with ImageNet pretrained weights by tstandley. [MIT License]\nModel: https://github.com/tstandley/Xception-PyTorch\nWeight: https://www.dropbox.com/s/1hplpzet9d7dv29/xception-c0a72b38.pth.tar?dl=1\nPytorch VGG19 pretrained weights:\nhttps://download.pytorch.org/models/vgg19-dcbb9e9d.pth
", "rawMarkdown": "facenet-pytorch: https://github.com/timesler/facenet-pytorch\nMesoNet: https://github.com/DariusAf/MesoNet\nPytorch xception net with ImageNet pretrained weights by tstandley. [MIT License]\nModel: https://github.com/tstandley/Xception-PyTorch\nWeight: https://www.dropbox.com/s/1hplpzet9d7dv29/xception-c0a72b38.pth.tar?dl=1\nPytorch VGG19 pretrained weights:\nhttps://download.pytorch.org/models/vgg19-dcbb9e9d.pth\n" }, { "id": 760125, "postDate": "2020-02-29T22:11:51.337Z", "content": "https://serre-lab.clps.brown.edu/resource/hmdb-a-large-human-motion-database/
", "rawMarkdown": "https://serre-lab.clps.brown.edu/resource/hmdb-a-large-human-motion-database/" }, { "id": 759754, "postDate": "2020-02-29T12:24:16.960Z", "content": "Resnet Weights https://github.com/tensorflow/models/tree/master/official/r1/resnet\nEfficientnet Weights https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nLFFD Weights https://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices
", "rawMarkdown": "Resnet Weights https://github.com/tensorflow/models/tree/master/official/r1/resnet\nEfficientnet Weights https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nLFFD Weights https://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices\n" }, { "id": 759745, "postDate": "2020-02-29T12:13:17.927Z", "content": "https://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/ShiqiYu/libfacedetection\nhttps://github.com/timesler/facenet-pytorch
", "rawMarkdown": "[https://github.com/biubug6/Pytorch_Retinaface](https://github.com/biubug6/Pytorch_Retinaface)\n[https://github.com/ShiqiYu/libfacedetection](https://github.com/ShiqiYu/libfacedetection)\n[https://github.com/timesler/facenet-pytorch](https://github.com/timesler/facenet-pytorch)" }, { "id": 759719, "postDate": "2020-02-29T11:34:30.520Z", "content": "https://github.com/fyu/drn/\nhttps://github.com/biubug6/Pytorch_Retinaface.git\nhttps://github.com/ondyari/FaceForensics/tree/master/classification \nhttps://github.com/TreB1eN/InsightFace_Pytorch.git\n https://github.com/kenshohara/3D-ResNets-PyTorch\n https://github.com/facebookresearch/SlowFast\n https://github.com/open-mmlab/mmaction\n https://github.com/MIT-HAN-LAB/temporal-shift-module\n https://deepmind.com/research/open-source/kinetics
", "rawMarkdown": "https://github.com/fyu/drn/\nhttps://github.com/biubug6/Pytorch_Retinaface.git\nhttps://github.com/ondyari/FaceForensics/tree/master/classification \nhttps://github.com/TreB1eN/InsightFace_Pytorch.git\n https://github.com/kenshohara/3D-ResNets-PyTorch\n https://github.com/facebookresearch/SlowFast\n https://github.com/open-mmlab/mmaction\n https://github.com/MIT-HAN-LAB/temporal-shift-module\n https://deepmind.com/research/open-source/kinetics\n\n" }, { "id": 759715, "postDate": "2020-02-29T11:27:49.223Z", "content": "https://github.com/protossw512/AdaptiveWingLoss\nhttps://github.com/TadasBaltrusaitis/OpenFace\nhttps://github.com/NVIDIA/flownet2-pytorch\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/qijiezhao/py-denseflow\nhttps://github.com/wanglimin/dense_flow\nhttps://github.com/yjxiong/tsn-pytorch\nhttps://github.com/metalbubble/TRN-pytorch\nhttps://github.com/mit-han-lab/temporal-shift-module
", "rawMarkdown": "https://github.com/protossw512/AdaptiveWingLoss\nhttps://github.com/TadasBaltrusaitis/OpenFace\nhttps://github.com/NVIDIA/flownet2-pytorch\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/qijiezhao/py-denseflow\nhttps://github.com/wanglimin/dense_flow\nhttps://github.com/yjxiong/tsn-pytorch\nhttps://github.com/metalbubble/TRN-pytorch\nhttps://github.com/mit-han-lab/temporal-shift-module" }, { "id": 759491, "postDate": "2020-02-29T05:25:46.653Z", "content": "https://github.com/supernotman/RetinaFace_Pytorch
", "rawMarkdown": "https://github.com/supernotman/RetinaFace_Pytorch" }, { "id": 759455, "postDate": "2020-02-29T03:42:58.507Z", "content": "https://github.com/open-mmlab/mmdetection/
", "rawMarkdown": "https://github.com/open-mmlab/mmdetection/" }, { "id": 759380, "postDate": "2020-02-29T00:38:43.653Z", "content": "COCO pre trained object detection https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md#coco-trained-models-coco-models
\n\npretrained image net models as others mentioned
", "rawMarkdown": "COCO pre trained object detection https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md#coco-trained-models-coco-models\n\npretrained image net models as others mentioned" }, { "id": 759222, "postDate": "2020-02-28T18:41:35.553Z", "content": "https://pypi.org/project/Keras/
", "rawMarkdown": "https://pypi.org/project/Keras/" }, { "id": 759221, "postDate": "2020-02-28T18:39:51.563Z", "content": "https://pypi.org/project/ffmpeg-python/
", "rawMarkdown": "https://pypi.org/project/ffmpeg-python/" }, { "id": 759216, "postDate": "2020-02-28T18:32:53.193Z", "content": "https://www.kaggle.com/rakibilly/ffmpeg-static-build
", "rawMarkdown": "https://www.kaggle.com/rakibilly/ffmpeg-static-build" }, { "id": 759213, "postDate": "2020-02-28T18:23:59.687Z", "content": "https://www.kaggle.com/dagnelies/deepfake-faces
", "rawMarkdown": "https://www.kaggle.com/dagnelies/deepfake-faces" }, { "id": 759172, "postDate": "2020-02-28T16:55:09.753Z", "content": "https://github.com/dessa-research
", "rawMarkdown": "https://github.com/dessa-research" }, { "id": 759035, "postDate": "2020-02-28T13:39:17.433Z", "content": "https://pytorch.org/hub/huggingface_pytorch-transformers/
", "rawMarkdown": "https://pytorch.org/hub/huggingface_pytorch-transformers/" }, { "id": 758791, "postDate": "2020-02-28T06:52:00.083Z", "content": "https://www.kaggle.com/chooyoungjun/deepmtcnn\nhttps://www.kaggle.com/sheldonrobinson/video-tools\nhttps://www.kaggle.com/chooyoungjun/deepfake14\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/ShiqiYu/libfacedetection\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/fyu/drn/\nhttps://github.com/biubug6/Pytorch_Retinaface.git\nhttps://github.com/TreB1eN/InsightFace_Pytorch.git\nhttps://github.com/kenshohara/3D-ResNets-PyTorch\nhttps://github.com/facebookresearch/SlowFast\nhttps://github.com/open-mmlab/mmaction\nhttps://github.com/MIT-HAN-LAB/temporal-shift-module\nhttps://deepmind.com/research/open-source/kinetics
", "rawMarkdown": "https://www.kaggle.com/chooyoungjun/deepmtcnn\nhttps://www.kaggle.com/sheldonrobinson/video-tools\nhttps://www.kaggle.com/chooyoungjun/deepfake14\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/ShiqiYu/libfacedetection\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/fyu/drn/\nhttps://github.com/biubug6/Pytorch_Retinaface.git\nhttps://github.com/TreB1eN/InsightFace_Pytorch.git\nhttps://github.com/kenshohara/3D-ResNets-PyTorch\nhttps://github.com/facebookresearch/SlowFast\nhttps://github.com/open-mmlab/mmaction\nhttps://github.com/MIT-HAN-LAB/temporal-shift-module\nhttps://deepmind.com/research/open-source/kinetics" }, { "id": 758668, "postDate": "2020-02-28T02:42:23.680Z", "content": "https://github.com/sthanhng/yoloface\nhttps://github.com/ultralytics/yolov3\nhttps://github.com/sthanhng/yoloface/blob/master/model-weights/get_models.sh\nhttps://pytorch.org/hub/pytorch_vision_resnet/\nhttps://github.com/pytorch/examples/tree/master/imagenet\nhttps://docs.opencv.org/3.4/d4/dee/tutorial_optical_flow.html\nhttps://pytorch.org/tutorials/beginner/audio_classifier_tutorial.html?highlight=audio\nhttp://openaccess.thecvf.com/content_ICCVW_2019/papers/HBU/Amerini_Deepfake_Video_Detection_through_Optical_Flow_Based_CNN_ICCVW_2019_paper.pdf
", "rawMarkdown": "https://github.com/sthanhng/yoloface\nhttps://github.com/ultralytics/yolov3\nhttps://github.com/sthanhng/yoloface/blob/master/model-weights/get_models.sh\nhttps://pytorch.org/hub/pytorch_vision_resnet/\nhttps://github.com/pytorch/examples/tree/master/imagenet\nhttps://docs.opencv.org/3.4/d4/dee/tutorial_optical_flow.html\nhttps://pytorch.org/tutorials/beginner/audio_classifier_tutorial.html?highlight=audio\nhttp://openaccess.thecvf.com/content_ICCVW_2019/papers/HBU/Amerini_Deepfake_Video_Detection_through_Optical_Flow_Based_CNN_ICCVW_2019_paper.pdf" }, { "id": 758256, "postDate": "2020-02-27T15:34:43.540Z", "content": "https://github.com/aleju/imgaug\nhttps://github.com/deepmind/kinetics-i3d\nhttps://github.com/OanaIgnat/i3d_keras
", "rawMarkdown": "https://github.com/aleju/imgaug\nhttps://github.com/deepmind/kinetics-i3d\nhttps://github.com/OanaIgnat/i3d_keras", "replies": [ { "id": 759420, "postDate": "2020-02-29T02:29:31.650Z", "content": "Just in case : \nhttps://github.com/dlpbc/keras-kinetics-i3d\nhttps://gist.github.com/albertomontesg/d8b21a179c1e6cca0480ebdf292c34d2
", "rawMarkdown": "Just in case : \nhttps://github.com/dlpbc/keras-kinetics-i3d\nhttps://gist.github.com/albertomontesg/d8b21a179c1e6cca0480ebdf292c34d2" } ] }, { "id": 757983, "postDate": "2020-02-27T10:00:52.410Z", "content": "A Light and Fast Face Detector for Edge Devices\nhttps://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices\nUltra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB
", "rawMarkdown": "A Light and Fast Face Detector for Edge Devices\nhttps://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices\nUltra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB" }, { "id": 757863, "postDate": "2020-02-27T07:32:13.603Z", "content": "https://github.com/deepmind/kinetics-i3d/
", "rawMarkdown": "https://github.com/deepmind/kinetics-i3d/" }, { "id": 757194, "postDate": "2020-02-26T14:16:57.170Z", "content": "Pretrained models from here: https://github.com/pytorch/fairseq
", "rawMarkdown": "Pretrained models from here: https://github.com/pytorch/fairseq" }, { "id": 756818, "postDate": "2020-02-26T05:40:21.330Z", "content": "LFW - People (Face Recognition) License GNU Lesser General Public License 3.0\nhttps://www.kaggle.com/atulanandjha/lfwpeople\nor http://vis-www.cs.umass.edu/lfw/
", "rawMarkdown": "LFW - People (Face Recognition) License GNU Lesser General Public License 3.0\nhttps://www.kaggle.com/atulanandjha/lfwpeople\nor http://vis-www.cs.umass.edu/lfw/" }, { "id": 756553, "postDate": "2020-02-25T20:39:38.823Z", "content": "https://pypi.org/project/efficientnet/\nimagenet weights
", "rawMarkdown": "https://pypi.org/project/efficientnet/\nimagenet weights" }, { "id": 756518, "postDate": "2020-02-25T19:45:48.570Z", "content": "MTCNN package - https://www.kaggle.com/diamondsnake/mtcnnpackage\nKeras applications models - https://keras.io/applications/
", "rawMarkdown": "MTCNN package - https://www.kaggle.com/diamondsnake/mtcnnpackage\nKeras applications models - https://keras.io/applications/\n" }, { "id": 756079, "postDate": "2020-02-25T11:47:29.447Z", "content": "https://github.com/ufoym/imbalanced-dataset-sampler
", "rawMarkdown": "https://github.com/ufoym/imbalanced-dataset-sampler" }, { "id": 756071, "postDate": "2020-02-25T11:40:07.963Z", "content": "Pre-trained models of VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/
", "rawMarkdown": "Pre-trained models of VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/" }, { "id": 755919, "postDate": "2020-02-25T09:10:23.913Z", "content": "https://github.com/hollance/BlazeFace-PyTorch\nsome changes in https://www.kaggle.com/humananalog/deepfakes-inference-demo's helpers code
", "rawMarkdown": "https://github.com/hollance/BlazeFace-PyTorch\nsome changes in https://www.kaggle.com/humananalog/deepfakes-inference-demo's helpers code\n\n" }, { "id": 755630, "postDate": "2020-02-25T00:58:11.357Z", "content": "faceforensics++ model\nhttp://kaldir.vc.in.tum.de/FaceForensics/models/faceforensics++_models.zip\npretrainedmodels\nhttps://files.pythonhosted.org/packages/84/0e/be6a0e58447ac16c938799d49bfb5fb7a80ac35e137547fc6cee2c08c4cf/pretrainedmodels-0.7.4.tar.gz\nface_recognition\nhttps://files.pythonhosted.org/packages/3f/ed/ad9a28042f373d4633fc8b49109b623597d6f193d3bbbef7780a5ee8eef2/face_recognition-1.2.3-py2.py3-none-any.whl\nface-recognition-models\nhttps://files.pythonhosted.org/packages/cf/3b/4fd8c534f6c0d1b80ce0973d01331525538045084c73c153ee6df20224cf/face_recognition_models-0.3.0.tar.gz\ndlib\nhttps://files.pythonhosted.org/packages/63/92/05c3b98636661cb80d190a5a777dd94effcc14c0f6893222e5ca81e74fbc/dlib-19.19.0.tar.gz\nhaarcascades\nhttps://github.com/opencv/opencv/tree/master/data/haarcascades\nshape_predictor_68_face_landmarks\nhttps://ja.osdn.net/projects/sfnet_dclib/downloads/dlib/v18.10/shape_predictor_68_face_landmarks.dat.bz2/
", "rawMarkdown": "faceforensics++ model\nhttp://kaldir.vc.in.tum.de/FaceForensics/models/faceforensics++_models.zip\npretrainedmodels\nhttps://files.pythonhosted.org/packages/84/0e/be6a0e58447ac16c938799d49bfb5fb7a80ac35e137547fc6cee2c08c4cf/pretrainedmodels-0.7.4.tar.gz\nface_recognition\nhttps://files.pythonhosted.org/packages/3f/ed/ad9a28042f373d4633fc8b49109b623597d6f193d3bbbef7780a5ee8eef2/face_recognition-1.2.3-py2.py3-none-any.whl\nface-recognition-models\nhttps://files.pythonhosted.org/packages/cf/3b/4fd8c534f6c0d1b80ce0973d01331525538045084c73c153ee6df20224cf/face_recognition_models-0.3.0.tar.gz\ndlib\nhttps://files.pythonhosted.org/packages/63/92/05c3b98636661cb80d190a5a777dd94effcc14c0f6893222e5ca81e74fbc/dlib-19.19.0.tar.gz\nhaarcascades\nhttps://github.com/opencv/opencv/tree/master/data/haarcascades\nshape_predictor_68_face_landmarks\nhttps://ja.osdn.net/projects/sfnet_dclib/downloads/dlib/v18.10/shape_predictor_68_face_landmarks.dat.bz2/" }, { "id": 755565, "postDate": "2020-02-24T22:28:25.930Z", "content": "All pre-trained models referenced on https://tfhub.dev before the entry deadline.
", "rawMarkdown": "All pre-trained models referenced on https://tfhub.dev before the entry deadline." }, { "id": 755551, "postDate": "2020-02-24T21:58:35.217Z", "content": "I3D with pretrained weights for kinetics in pytorch - https://github.com/piergiaj/pytorch-i3d\n3D ResNets for pytorch https://github.com/kenshohara/3D-ResNets-PyTorch with pretrained weights from the google drive they link to
", "rawMarkdown": "I3D with pretrained weights for kinetics in pytorch - https://github.com/piergiaj/pytorch-i3d\n3D ResNets for pytorch https://github.com/kenshohara/3D-ResNets-PyTorch with pretrained weights from the google drive they link to" }, { "id": 755548, "postDate": "2020-02-24T21:54:56.283Z", "content": "DeepFace trained on VGGFace2\nGoogle facial expression comparison dataset (CC0) https://research.google/tools/datasets/google-facial-expression/
", "rawMarkdown": "DeepFace trained on VGGFace2\nGoogle facial expression comparison dataset (CC0) https://research.google/tools/datasets/google-facial-expression/" }, { "id": 755439, "postDate": "2020-02-24T19:24:20.573Z", "content": "https://github.com/protossw512/AdaptiveWingLoss\nhttps://wywu.github.io/projects/LAB/WFLW.html - Wider Facial Landmarks in-the-wild
", "rawMarkdown": "https://github.com/protossw512/AdaptiveWingLoss\nhttps://wywu.github.io/projects/LAB/WFLW.html - Wider Facial Landmarks in-the-wild " }, { "id": 755244, "postDate": "2020-02-24T15:36:59.573Z", "content": "\"Post links to your external data sources here before the deadline specified in the rules.\"
\n\nIs the date March 31st? There's an early March deadline of March 3rd for mergers, etc. I didn't see anywhere in the rules about \"external data disclosure\" deadlines.
\n\nPlease let me know and thank you!
\n\nRodney
", "rawMarkdown": "\"Post links to your external data sources here before the deadline specified in the rules.\"\n\nIs the date March 31st? There's an early March deadline of March 3rd for mergers, etc. I didn't see anywhere in the rules about \"external data disclosure\" deadlines.\n\nPlease let me know and thank you!\n\nRodney\n", "replies": [ { "id": 756459, "postDate": "2020-02-25T18:25:47.800Z", "content": "Per the competition rules,
\n\n\n\n\nC. External Data. You may use data other than the Competition Data (“External Data”) to develop and test your models and Submissions. However, you will (i) ensure the External Data is available to use by all participants of the competition for purposes of the competition at no cost to the other participants and (ii) post such access to the External Data for the participants to the official competition forum prior to the Entry Deadline.
\n
With the Entry Deadline being March 3rd, this is also the deadline for posting external data. as specified on the Timeline page.
", "rawMarkdown": "Per the competition rules, \n> C. External Data. You may use data other than the Competition Data (“External Data”) to develop and test your models and Submissions. However, you will (i) ensure the External Data is available to use by all participants of the competition for purposes of the competition at no cost to the other participants and (ii) **post such access to the External Data for the participants to the official competition forum prior to the Entry Deadline**.\n\nWith the Entry Deadline being March 3rd, this is also the deadline for posting external data. as specified on the Timeline page." } ] }, { "id": 754869, "postDate": "2020-02-24T06:40:38.323Z", "content": "RetinaFace : https://github.com/deepinsight/insightface/tree/master/RetinaFace
", "rawMarkdown": "RetinaFace : [https://github.com/deepinsight/insightface/tree/master/RetinaFace](https://github.com/deepinsight/insightface/tree/master/RetinaFace)" }, { "id": 754851, "postDate": "2020-02-24T05:56:20.267Z", "content": "https://github.com/scikit-video/scikit-video\nhttps://github.com/facebookresearch/detectron2\nhttps://github.com/pytorch/vision/tree/master/torchvision/models\nhttps://github.com/yhenon/pytorch-retinanet\nhttps://github.com/kenshohara/3D-ResNets-PyTorch\nhttps://github.com/toandaominh1997/EfficientDet.Pytorch
", "rawMarkdown": "https://github.com/scikit-video/scikit-video\nhttps://github.com/facebookresearch/detectron2\nhttps://github.com/pytorch/vision/tree/master/torchvision/models\nhttps://github.com/yhenon/pytorch-retinanet\nhttps://github.com/kenshohara/3D-ResNets-PyTorch\nhttps://github.com/toandaominh1997/EfficientDet.Pytorch" }, { "id": 754743, "postDate": "2020-02-24T02:04:37.703Z", "content": "Nice repo (MIT license) with pretrained models: https://github.com/dessa-public/DeepFake-Detection
\n\nAstronomy library with some useful image analysis tools: https://github.com/keflavich/agpy
", "rawMarkdown": "Nice repo (MIT license) with pretrained models: https://github.com/dessa-public/DeepFake-Detection\n\nAstronomy library with some useful image analysis tools: https://github.com/keflavich/agpy" }, { "id": 754488, "postDate": "2020-02-23T16:03:24.810Z", "content": "detectron2 model zoo (https://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md)
", "rawMarkdown": "\ndetectron2 model zoo (https://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md)" }, { "id": 754480, "postDate": "2020-02-23T15:50:15.697Z", "content": "http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://pypi.org/project/mtcnn/
", "rawMarkdown": "http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://pypi.org/project/mtcnn/" }, { "id": 754477, "postDate": "2020-02-23T15:41:36.513Z", "content": "https://pypi.org/project/face-recognition/
", "rawMarkdown": "https://pypi.org/project/face-recognition/" }, { "id": 754464, "postDate": "2020-02-23T15:22:19.917Z", "content": "BlazeFace Tensorflow light model (.tflite): https://github.com/google/mediapipe/blob/master/mediapipe/models/face_detection_front.tflite
", "rawMarkdown": "BlazeFace Tensorflow light model (.tflite): https://github.com/google/mediapipe/blob/master/mediapipe/models/face_detection_front.tflite" }, { "id": 754279, "postDate": "2020-02-23T10:41:52.643Z", "content": "ffmpeg-python\n- Github https://github.com/kkroening/ffmpeg-python\n- ffmpeg-python wheel for offline installation by @phoenix9032 https://www.kaggle.com/phoenix9032/ffmpegpython
", "rawMarkdown": "ffmpeg-python\n- Github [https://github.com/kkroening/ffmpeg-python](https://github.com/kkroening/ffmpeg-python)\n- ffmpeg-python wheel for offline installation by @phoenix9032 [https://www.kaggle.com/phoenix9032/ffmpegpython](https://www.kaggle.com/phoenix9032/ffmpegpython)\n", "replies": [ { "id": 754698, "postDate": "2020-02-24T00:12:25.873Z", "content": "When I include \"ffmpeg-python wheel for offline installation\" by @phoenix9032 https://www.kaggle.com/phoenix9032/ffmpegpython in my kernel the \"submit to competition\" button is inactive with \"Your notebook cannot use non-standard datasets in this competiton\" message. See discussion thread. What's the problem? @addisonhoward @juliaelliott?
", "rawMarkdown": "When I include \"ffmpeg-python wheel for offline installation\" by @phoenix9032 https://www.kaggle.com/phoenix9032/ffmpegpython in my kernel the \"submit to competition\" button is inactive with \"Your notebook cannot use non-standard datasets in this competiton\" message. See [discussion thread](https://www.kaggle.com/c/deepfake-detection-challenge/discussion/132005). What's the problem? @addisonhoward @juliaelliott?\n" } ] }, { "id": 754273, "postDate": "2020-02-23T10:25:59.347Z", "content": "Do we need to post implementations of specific layers as well (here and in general)? Like for example a special loss function we want to use that is not per default implemented in tensorflow / pytorch?
", "rawMarkdown": "Do we need to post implementations of specific layers as well (here and in general)? Like for example a special loss function we want to use that is not per default implemented in tensorflow / pytorch?" }, { "id": 753980, "postDate": "2020-02-22T22:18:47.200Z", "content": "Yolo V3 face detector: https://github.com/sthanhng/yoloface
", "rawMarkdown": "Yolo V3 face detector: https://github.com/sthanhng/yoloface" }, { "id": 753887, "postDate": "2020-02-22T19:27:06.580Z", "content": "https://github.com/ipazc/mtcnn\nAlso, a blanket declaration for all models under Module: tf.keras.applications (https://github.com/tensorflow/tensorflow)
", "rawMarkdown": "https://github.com/ipazc/mtcnn\nAlso, a blanket declaration for all models under Module: tf.keras.applications (https://github.com/tensorflow/tensorflow)" }, { "id": 753172, "postDate": "2020-02-21T19:51:55.260Z", "content": "https://github.com/Cadene/pretrained-models.pytorch/blob/master/pretrainedmodels/models/senet.py
", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch/blob/master/pretrainedmodels/models/senet.py" }, { "id": 752948, "postDate": "2020-02-21T15:10:05.030Z", "content": "https://github.com/qubvel/segmentation_models\nhttps://github.com/ipazc/mtcnn
", "rawMarkdown": "https://github.com/qubvel/segmentation_models\nhttps://github.com/ipazc/mtcnn" }, { "id": 752820, "postDate": "2020-02-21T12:51:21.533Z", "content": "Real and fake face dataset: https://www.kaggle.com/ciplab/real-and-fake-face-detection
", "rawMarkdown": "Real and fake face dataset: https://www.kaggle.com/ciplab/real-and-fake-face-detection" }, { "id": 752661, "postDate": "2020-02-21T09:21:41.140Z", "content": "https://github.com/d-li14/mobilenetv3.pytorch
", "rawMarkdown": "https://github.com/d-li14/mobilenetv3.pytorch" }, { "id": 752609, "postDate": "2020-02-21T08:00:18.870Z", "content": "Celebrity face dataset: \nhttps://github.com/prateekmehta59/Celebrity-Face-Recognition-Dataset
", "rawMarkdown": "Celebrity face dataset: \nhttps://github.com/prateekmehta59/Celebrity-Face-Recognition-Dataset" }, { "id": 752490, "postDate": "2020-02-21T04:27:48.747Z", "content": "Facenet Implementation by keras2 and model weight\nhttps://github.com/nyoki-mtl/keras-facenet
", "rawMarkdown": "Facenet Implementation by keras2 and model weight\nhttps://github.com/nyoki-mtl/keras-facenet" }, { "id": 752306, "postDate": "2020-02-20T21:26:45.937Z", "content": "face2 dataset : http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nimagenet of kaggle competition dataset : https://www.kaggle.com/c/imagenet-object-localization-challenge/data
\n\nI am very new on kaggle so could someone confirm me that Creative Commons Attribution-ShareAlike 4.0 International License is ok ?
", "rawMarkdown": "face2 dataset : http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nimagenet of kaggle competition dataset : https://www.kaggle.com/c/imagenet-object-localization-challenge/data\n\nI am very new on kaggle so could someone confirm me that Creative Commons Attribution-ShareAlike 4.0 International License is ok ?\n" }, { "id": 751682, "postDate": "2020-02-20T12:04:10.730Z", "content": "MTCNN package\nhttps://www.kaggle.com/diamondsnake/mtcnnpackage
", "rawMarkdown": "MTCNN package\nhttps://www.kaggle.com/diamondsnake/mtcnnpackage" }, { "id": 751588, "postDate": "2020-02-20T10:05:54.800Z", "content": "https://github.com/qubvel/efficientnet\nhttps://github.com/qubvel/classification_models\nhttps://github.com/1adrianb/face-alignment\nhttps://www.kaggle.com/humananalog/deepfakes-inference-demo
", "rawMarkdown": "https://github.com/qubvel/efficientnet\nhttps://github.com/qubvel/classification_models\nhttps://github.com/1adrianb/face-alignment\nhttps://www.kaggle.com/humananalog/deepfakes-inference-demo" }, { "id": 751556, "postDate": "2020-02-20T09:20:40.487Z", "content": "pytorch pretrained model collections:\nhttps://github.com/Cadene/pretrained-models.pytorch
", "rawMarkdown": "pytorch pretrained model collections:\nhttps://github.com/Cadene/pretrained-models.pytorch" }, { "id": 750470, "postDate": "2020-02-19T12:19:23.397Z", "content": "pytorch pretrained models:\n 'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n 'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n 'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n 'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n 'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n 'resnext50_32x4d': 'https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pth',\n 'resnext101_32x8d': 'https://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pth',\n 'wide_resnet50_2': 'https://download.pytorch.org/models/wide_resnet50_2-95faca4d.pth',\n 'wide_resnet101_2': 'https://download.pytorch.org/models/wide_resnet101_2-32ee1156.pth',
", "rawMarkdown": "pytorch pretrained models:\n 'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n 'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n 'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n 'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n 'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n 'resnext50_32x4d': 'https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pth',\n 'resnext101_32x8d': 'https://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pth',\n 'wide_resnet50_2': 'https://download.pytorch.org/models/wide_resnet50_2-95faca4d.pth',\n 'wide_resnet101_2': 'https://download.pytorch.org/models/wide_resnet101_2-32ee1156.pth'," }, { "id": 749074, "postDate": "2020-02-18T09:12:57.320Z", "content": "https://keras.io/applications/#inceptionresnetv2
", "rawMarkdown": "https://keras.io/applications/#inceptionresnetv2" }, { "id": 747037, "postDate": "2020-02-15T21:51:39.923Z", "content": "https://www.kaggle.com/selfishgene/youtube-faces-with-facial-keypoints\nPretrained models: https://github.com/osmr/imgclsmob
", "rawMarkdown": "https://www.kaggle.com/selfishgene/youtube-faces-with-facial-keypoints\nPretrained models: https://github.com/osmr/imgclsmob", "replies": [ { "id": 751858, "postDate": "2020-02-20T15:22:10.433Z", "content": "Does CelebA actually available to use in this contest? Some participants said, that datasets, witch not available to commercial use is not allowed.
", "rawMarkdown": "Does CelebA actually available to use in this contest? Some participants said, that datasets, witch not available to commercial use is not allowed.", "votes": 3 }, { "id": 752959, "postDate": "2020-02-21T15:21:39.780Z", "content": "Yea CelebA, UTKFace, and FFHQ aren't allowed I think
", "rawMarkdown": "Yea CelebA, UTKFace, and FFHQ aren't allowed I think", "votes": 1 } ] }, { "id": 744776, "postDate": "2020-02-13T06:10:35.637Z", "content": "https://github.com/qubvel/efficientnet\nefficientnet and pretrained weights
", "rawMarkdown": "https://github.com/qubvel/efficientnet\nefficientnet and pretrained weights " }, { "id": 744586, "postDate": "2020-02-13T00:24:54.367Z", "content": "Facenet weights for keras: https://drive.google.com/open?id=1pwQ3H4aJ8a6yyJHZkTwtjcL4wYWQb7bn
", "rawMarkdown": "Facenet weights for keras: https://drive.google.com/open?id=1pwQ3H4aJ8a6yyJHZkTwtjcL4wYWQb7bn" }, { "id": 744330, "postDate": "2020-02-12T18:40:36.117Z", "content": "ASVspoof 2019: The 3rd Automatic Speaker Verification Spoofing and Countermeasures Challenge database\nhttps://datashare.is.ed.ac.uk/handle/10283/3336\nlicensed under The Open Data Commons Attribution License
", "rawMarkdown": "ASVspoof 2019: The 3rd Automatic Speaker Verification Spoofing and Countermeasures Challenge database\n[https://datashare.is.ed.ac.uk/handle/10283/3336](https://datashare.is.ed.ac.uk/handle/10283/3336)\nlicensed under The Open Data Commons Attribution License\n" }, { "id": 743643, "postDate": "2020-02-12T06:48:27.283Z", "content": "I'm going to be using posenet and tensorflow lite runtime to run it.\nhttps://www.tensorflow.org/lite/guide/python\nhttps://dl.google.com/coral/python/tflite_runtime-2.1.0-cp36-cp36m-linux_x86_64.whl\nhttps://www.tensorflow.org/lite/models/pose_estimation/overview\nhttps://storage.googleapis.com/download.tensorflow.org/models/tflite/posenet_mobilenet_v1_100_257x257_multi_kpt_stripped.tflite
", "rawMarkdown": "I'm going to be using posenet and tensorflow lite runtime to run it.\nhttps://www.tensorflow.org/lite/guide/python\nhttps://dl.google.com/coral/python/tflite_runtime-2.1.0-cp36-cp36m-linux_x86_64.whl\nhttps://www.tensorflow.org/lite/models/pose_estimation/overview\nhttps://storage.googleapis.com/download.tensorflow.org/models/tflite/posenet_mobilenet_v1_100_257x257_multi_kpt_stripped.tflite", "replies": [ { "id": 745598, "postDate": "2020-02-14T01:21:57.217Z", "content": "I'm going to use this version of the posenet instead.\nhttps://storage.googleapis.com/download.tensorflow.org/models/tflite/gpu/multi_person_mobilenet_v1_075_float.tflite\nor any other available here:\nhttps://www.tensorflow.org/lite/guide/hosted_models#pose_estimation
", "rawMarkdown": "I'm going to use this version of the posenet instead.\nhttps://storage.googleapis.com/download.tensorflow.org/models/tflite/gpu/multi_person_mobilenet_v1_075_float.tflite\nor any other available here:\nhttps://www.tensorflow.org/lite/guide/hosted_models#pose_estimation" } ] }, { "id": 743361, "postDate": "2020-02-12T02:26:00.633Z", "content": "Can we use new external models/datasets after March 3rd ? because the deadline to disclose would have been passed then ?
", "rawMarkdown": "Can we use new external models/datasets after March 3rd ? because the deadline to disclose would have been passed then ?", "replies": [ { "id": 743403, "postDate": "2020-02-12T03:21:13.957Z", "content": "@basharallabadi Good question. The deadline to declare external data is on March 3rd. So you cannot add new external datasets after that deadline, but you can use any datasets that have been declared (which are not prohibited) on this thread.
", "rawMarkdown": "@basharallabadi Good question. The deadline to declare external data is on March 3rd. So you cannot add new external datasets after that deadline, but you can use any datasets that have been declared (which are not prohibited) on this thread.", "votes": 2 } ] }, { "id": 738129, "postDate": "2020-02-06T07:07:35.427Z", "content": "RetinaFace and ArcFace and pre-trained models in https://github.com/deepinsight/insightface
", "rawMarkdown": "RetinaFace and ArcFace and pre-trained models in https://github.com/deepinsight/insightface", "replies": [ { "id": 741220, "postDate": "2020-02-10T10:30:26.457Z", "content": "and\nyoutube8m https://research.google.com/youtube8m/index.html
", "rawMarkdown": "and\nyoutube8m https://research.google.com/youtube8m/index.html\n" }, { "id": 746506, "postDate": "2020-02-15T05:49:07.920Z", "content": "and YoutubeFace http://www.cslab.openu.ac.il/download/wolftau/
", "rawMarkdown": "and YoutubeFace http://www.cslab.openu.ac.il/download/wolftau/" }, { "id": 746521, "postDate": "2020-02-15T06:24:30.840Z", "content": "@hsienshen, YoutubeFace is not open source data
", "rawMarkdown": "@hsienshen, YoutubeFace is not open source data", "votes": 1 }, { "id": 746623, "postDate": "2020-02-15T09:28:59.987Z", "content": "@seshurajup Sorry that I'm not familiar with the rules. But I think it is fair use. Can anyone else correct me if I am wrong. Thanks!
", "rawMarkdown": "@seshurajup Sorry that I'm not familiar with the rules. But I think it is fair use. Can anyone else correct me if I am wrong. Thanks!" }, { "id": 763407, "postDate": "2020-03-04T13:00:15.693Z", "content": "hnsw https://github.com/nmslib/hnswlib
", "rawMarkdown": "hnsw https://github.com/nmslib/hnswlib" } ] }, { "id": 737844, "postDate": "2020-02-05T20:55:41.117Z", "content": "Pytorch pretrained models from: https://pytorch.org/docs/stable/torchvision/models.html
", "rawMarkdown": "Pytorch pretrained models from: https://pytorch.org/docs/stable/torchvision/models.html" }, { "id": 732513, "postDate": "2020-01-29T22:48:30.150Z", "content": "@tunguz' datasets under CC BY-NC 4.0 licence are allowed, or not allowed in this competition? A saw the question already below, sorry for repeat, but from the answer it is not clear for me.
", "rawMarkdown": "@tunguz' datasets under CC BY-NC 4.0 licence are allowed, or not allowed in this competition? A saw the question already below, sorry for repeat, but from the answer it is not clear for me.", "replies": [ { "id": 732865, "postDate": "2020-01-30T11:53:32.763Z", "content": "@addisonhoward @juliaelliott \nWould be nice to see info on all types of licences and restrictions:\nMain Github licenses like MIT, Apache, GNU(all allow commercial use), I assume, are allowed, right?\nWhat about explicit statement about availability for non-commercial research/education only? (BTW Imagenet is under such a license and models pretrained on it are allowed).\nWhat about Non-Commercial Creative Commons licenses that allow reuse/modification like BY-NC mentioned above?
", "rawMarkdown": "@addisonhoward @juliaelliott \nWould be nice to see info on all types of licences and restrictions:\nMain Github licenses like MIT, Apache, GNU(all allow commercial use), I assume, are allowed, right?\nWhat about explicit statement about availability for non-commercial research/education only? (BTW Imagenet is under such a license and models pretrained on it are allowed).\nWhat about Non-Commercial Creative Commons licenses that allow reuse/modification like BY-NC mentioned above?", "votes": 1 }, { "id": 732954, "postDate": "2020-01-30T13:41:31.710Z", "content": "@tetelias It wouldn’t be possible to capture a comprehensive list. So it is expected that competitors understand the external data they’re using and ensure it matches the requirements in the rules.
\n\nI’ve answered the question about BY-NC not being available for use by all (non-commercial use) and therefore violating the requirement that external data be available for use by all participants.
", "rawMarkdown": "@tetelias It wouldn’t be possible to capture a comprehensive list. So it is expected that competitors understand the external data they’re using and ensure it matches the requirements in the rules. \n\nI’ve [answered the question about BY-NC](https://www.kaggle.com/c/deepfake-detection-challenge/discussion/121203#721812) not being available for use by all (non-commercial use) and therefore violating the requirement that external data be available for use by all participants.", "votes": 1 }, { "id": 733044, "postDate": "2020-01-30T15:22:35.820Z", "content": "@juliaelliott Thank you for speedy answer!\nSo trying to summarize: would it be correct to assume that any code/dataset under license permitting commercial use is allowed, while anything that's allowed only for non-commercial purposes cannot be used?
\n\nEdit: Also looking through all major fake videos datasets they all seem to be restricted for non-commercial research only: FaceForensics, FaceForensics++, DeepFakes Detection, both versions of Celeb-DF and still unpublished DeeperForensics-1.0. Is it ok to use them?
", "rawMarkdown": "@juliaelliott Thank you for speedy answer!\nSo trying to summarize: would it be correct to assume that any code/dataset under license permitting commercial use is allowed, while anything that's allowed only for non-commercial purposes cannot be used?\n\nEdit: Also looking through all major fake videos datasets they all seem to be restricted for non-commercial research only: **FaceForensics, FaceForensics++, DeepFakes Detection**, both versions of **Celeb-DF** and still unpublished **DeeperForensics-1.0**. Is it ok to use them?" }, { "id": 738853, "postDate": "2020-02-07T05:12:02.583Z", "content": "@tetelias Yes, if there are restrictions on a dataset’s use, it would violate the competition rules.
", "rawMarkdown": "@tetelias Yes, if there are restrictions on a dataset’s use, it would violate the competition rules.", "votes": 1 }, { "id": 744264, "postDate": "2020-02-12T17:30:17.567Z", "content": "@juliaelliott , I am still confused by conflicting signals. You referenced the following:
\n\n> you will (i) ensure the External Data is available to use by all participants of the competition for purposes of the competition at no cost to the other participants and (ii) post such access to the External Data for the participants to the official competition forum prior to the Entry Deadline.
\n\nNote the phrase for purposes of the competition. The purpose of the competition is not commercial. So can you please resolve this inconsistency, is it allowed to use datasets which are for non-commercial research only, as @tetelias mentioned.
@zaharch You raise a fair question. The rule indicates to use by all participants -- Inherent in data licensing terms that constrain the data's use is the restriction that it is not available to be used by all participants. In many cases with research/academic-use licensing, only those in a research profession may gain access to the dataset. Sure, in some cases, everyone may be able to physically download it, but if a license prohibits its use for non-research uses, then it can be interpreted as also prohibiting use by users who are non-researching. So the issue here is one of legally-compliant accessibility to use of the dataset by all.
Likewise, the winners' licensing terms require that the solution be open sourced that that in no event limits commercial use of such code or model containing or depending on such code.
Hi, @juliaelliott . as you mention: \nI’ve answered the question about BY-NC not being available for use by all (non-commercial use)\n\nWe can not use any dataset which is non-commercial, because no all competitor can use it. But ImageNet is for non-commercial (http://image-net.org/download-faq). So it means we can not use ImageNet, and pretrained model by it?
interesting, if imagenet is out of picture it makes most of the pretrained model zoos unavailable(for example torchvision.models. @juliaelliott can you please clarify this?
", "rawMarkdown": "interesting, if imagenet is out of picture it makes most of the pretrained model zoos unavailable(for example torchvision.models. @juliaelliott can you please clarify this?", "votes": 5 } ] }, { "id": 731744, "postDate": "2020-01-29T01:53:38.003Z", "content": "The EfficientNet repository as uploaded by @xhlulu:
\n\nhttps://www.kaggle.com/xhlulu/efficientnet-keras-source-code
", "rawMarkdown": "The EfficientNet repository as uploaded by @xhlulu:\n\nhttps://www.kaggle.com/xhlulu/efficientnet-keras-source-code" }, { "id": 728072, "postDate": "2020-01-24T11:24:11.520Z", "content": "AVA dataset: https://research.google.com/ava/download.html
", "rawMarkdown": "AVA dataset: https://research.google.com/ava/download.html" }, { "id": 726941, "postDate": "2020-01-23T10:25:05.517Z", "content": "https://www.kaggle.com/carlossouza/dlibpkg\nhttps://www.kaggle.com/minhtam/imageio-ffmpeg\nhttps://www.kaggle.com/minhtam/face-recognition\nhttps://www.kaggle.com/minhtam/fake-detect-basic
", "rawMarkdown": "https://www.kaggle.com/carlossouza/dlibpkg\nhttps://www.kaggle.com/minhtam/imageio-ffmpeg\nhttps://www.kaggle.com/minhtam/face-recognition\nhttps://www.kaggle.com/minhtam/fake-detect-basic" }, { "id": 718340, "postDate": "2020-01-14T10:01:02.863Z", "content": "Open Images\nhttps://www.kaggle.com/bigquery/open-images
", "rawMarkdown": "Open Images\n[https://www.kaggle.com/bigquery/open-images](https://www.kaggle.com/bigquery/open-images)" }, { "id": 715444, "postDate": "2020-01-10T14:45:02.097Z", "content": "ResNeXt pre-trained weights available via torch.hub or TorchVision:\nhttps://pytorch.org/hub/pytorch_vision_resnext/
", "rawMarkdown": "ResNeXt pre-trained weights available via torch.hub or TorchVision:\nhttps://pytorch.org/hub/pytorch_vision_resnext/" }, { "id": 715442, "postDate": "2020-01-10T14:43:08.773Z", "content": "face.evoLVe: High-Performance Face Recognition Library based on PyTorch\n(including some pre-trained models):\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch
", "rawMarkdown": "face.evoLVe: High-Performance Face Recognition Library based on PyTorch\n(including some pre-trained models):\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch" }, { "id": 712780, "postDate": "2020-01-07T15:16:15.257Z", "content": "External data that must be 1 GB, is the only pre-trained model weights's data or it includes any external model (Including what i have trained with myself)? Question is about size of external data
", "rawMarkdown": "External data that must be 1 GB, is the only pre-trained model weights's data or it includes any external model (Including what i have trained with myself)? Question is about size of external data" }, { "id": 712699, "postDate": "2020-01-07T14:13:55.257Z", "content": "DenseNet pre-trained weights
", "rawMarkdown": "DenseNet pre-trained weights" }, { "id": 711150, "postDate": "2020-01-05T18:38:28.630Z", "content": "Mesonet pre-trained weights\nhttp://vis-www.cs.umass.edu/lfw/index.html#download
", "rawMarkdown": "Mesonet pre-trained weights\n[http://vis-www.cs.umass.edu/lfw/index.html#download](http://vis-www.cs.umass.edu/lfw/index.html#download)" }, { "id": 707231, "postDate": "2019-12-31T12:57:46.800Z", "content": "Keras FaceNet Model: https://www.kaggle.com/nikhil1011/facenet
", "rawMarkdown": "Keras FaceNet Model: https://www.kaggle.com/nikhil1011/facenet" }, { "id": 706256, "postDate": "2019-12-30T06:32:21.650Z", "content": "Resnet with imagenet weights
", "rawMarkdown": "Resnet with imagenet weights" }, { "id": 705633, "postDate": "2019-12-29T07:55:01.617Z", "content": "Xception Imagenet Keras/Pytorch\nMobilenet SSD based on (https://github.com/yeephycho/tensorflow-face-detection)\nEfficientnet Models
", "rawMarkdown": "Xception Imagenet Keras/Pytorch\nMobilenet SSD based on (https://github.com/yeephycho/tensorflow-face-detection)\nEfficientnet Models" }, { "id": 701554, "postDate": "2019-12-23T15:55:13.707Z", "content": "@juliaelliott , Are we allowed to add external libraries(Open Source Initiative licenses) to our submission notebook as files? \nAnd install them in the notebook when it's being executed.
", "rawMarkdown": "@juliaelliott , Are we allowed to add external libraries(Open Source Initiative licenses) to our submission notebook as files? \nAnd install them in the notebook when it's being executed." }, { "id": 700725, "postDate": "2019-12-22T13:57:23.317Z", "content": "Real and Fake faces dataset:\nhttps://www.kaggle.com/ciplab/real-and-fake-face-detection
", "rawMarkdown": "Real and Fake faces dataset:\nhttps://www.kaggle.com/ciplab/real-and-fake-face-detection" }, { "id": 694595, "postDate": "2019-12-13T20:23:31.467Z", "content": "http://www.cbsr.ia.ac.cn/users/xiangyuzhu/projects/3DDFA/main.htm
", "rawMarkdown": "http://www.cbsr.ia.ac.cn/users/xiangyuzhu/projects/3DDFA/main.htm" }, { "id": 693839, "postDate": "2019-12-12T20:48:39.793Z", "content": "Are we allowed to use pre-trained models available on free licenses, for example in order to perform face detection?\nI can't find anything about it in the rules here , so it should be allowed. On the other hand on the webpage of the competition here I can read:
\n\n\n\n\nwe ask that you make the following commitments: (...)\n - You are submitting information and models for the Deepfake Detection Challenge which belong to you and are a result of your own work.
\n
So... Can I get a freely available model for face recognition built by somebody else, upload it as data, announce it in this thread and use it freely? Or is it prohibited and people must reinvent the wheel and prepare their own models for every aspect of the submission instead of focusing on the real challenge? I think the rules should be unified.
", "rawMarkdown": "Are we allowed to use pre-trained models available on free licenses, for example in order to perform face detection?\nI can't find anything about it in the rules [here](https://www.kaggle.com/c/deepfake-detection-challenge/rules) , so it should be allowed. On the other hand on the webpage of the competition [here](https://deepfakedetectionchallenge.ai/terms) I can read:\n> we ask that you make the following commitments: (...)\n- You are submitting information and models for the Deepfake Detection Challenge which belong to you and are a result of your own work.\n\nSo... Can I get a freely available model for face recognition built by somebody else, upload it as data, announce it in this thread and use it freely? Or is it prohibited and people must reinvent the wheel and prepare their own models for every aspect of the submission instead of focusing on the real challenge? I think the rules should be unified.", "replies": [ { "id": 694466, "postDate": "2019-12-13T16:45:35.267Z", "content": "Yes, you can use pretrained models.\nYes, you can use an externally obtained model, so long as you have rights to use it and declare it publicly here.
", "rawMarkdown": "Yes, you can use pretrained models.\nYes, you can use an externally obtained model, so long as you have rights to use it and declare it publicly here.", "votes": 1 }, { "id": 707519, "postDate": "2020-01-01T01:28:17.237Z", "rawMarkdown": "", "isDeleted": true } ] }, { "id": 772665, "postDate": "2020-03-15T18:34:19.613Z", "rawMarkdown": "", "isDeleted": true }, { "id": 762857, "postDate": "2020-03-03T21:52:09.783Z", "content": "https://github.com/thiago1080/SphereFace\nhttps://github.com/vlad3996/FaceDetection-DSFD
", "rawMarkdown": "https://github.com/thiago1080/SphereFace\nhttps://github.com/vlad3996/FaceDetection-DSFD\n", "isDeleted": true }, { "id": 762821, "postDate": "2020-03-03T20:43:36.393Z", "rawMarkdown": "", "isDeleted": true }, { "id": 762358, "postDate": "2020-03-03T13:12:22.580Z", "rawMarkdown": "", "isDeleted": true }, { "id": 762148, "postDate": "2020-03-03T08:53:31.800Z", "rawMarkdown": "", "isDeleted": true }, { "id": 761903, "postDate": "2020-03-03T02:52:12.577Z", "rawMarkdown": "", "votes": -2, "isDeleted": true, "replies": [ { "id": 762423, "postDate": "2020-03-03T13:54:54.043Z", "content": "No Python3??? 👎
", "rawMarkdown": "No Python3??? 👎 " } ] }, { "id": 761861, "postDate": "2020-03-03T01:54:07.090Z", "rawMarkdown": "", "isDeleted": true }, { "id": 761558, "postDate": "2020-03-02T16:55:12.550Z", "rawMarkdown": "", "isDeleted": true }, { "id": 761511, "postDate": "2020-03-02T15:56:48.663Z", "rawMarkdown": "", "isDeleted": true }, { "id": 761299, "postDate": "2020-03-02T11:17:26.747Z", "content": "External datasets:\nDeepfake150: https://www.kaggle.com/unkownhihi/deepfake\nFFHQ: https://www.kaggle.com/greatgamedota/ffhq-face-data-set\nReal and Fake datasets: https://www.kaggle.com/ciplab/real-and-fake-face-detection
\n\nMy datasets:\nhomemade deepfake faces: https://www.kaggle.com/muerbingsha/99430-faces\npretrained models: \nhttps://www.kaggle.com/muerbingsha/deepfake-my-resnet\nhttps://www.kaggle.com/muerbingsha/deepfake-my-inception
", "rawMarkdown": "External datasets:\nDeepfake150: https://www.kaggle.com/unkownhihi/deepfake\nFFHQ: https://www.kaggle.com/greatgamedota/ffhq-face-data-set\nReal and Fake datasets: https://www.kaggle.com/ciplab/real-and-fake-face-detection\n\nMy datasets:\nhomemade deepfake faces: https://www.kaggle.com/muerbingsha/99430-faces\npretrained models: \nhttps://www.kaggle.com/muerbingsha/deepfake-my-resnet\nhttps://www.kaggle.com/muerbingsha/deepfake-my-inception", "isDeleted": true }, { "id": 760173, "postDate": "2020-02-29T23:48:12.127Z", "rawMarkdown": "", "isDeleted": true }, { "id": 759727, "postDate": "2020-02-29T11:44:49.160Z", "rawMarkdown": "", "isDeleted": true }, { "id": 755942, "postDate": "2020-02-25T09:41:03.900Z", "rawMarkdown": "", "isDeleted": true }, { "id": 747766, "postDate": "2020-02-16T20:20:02.913Z", "rawMarkdown": "", "isDeleted": true }, { "id": 746490, "postDate": "2020-02-15T04:59:09.903Z", "rawMarkdown": "", "isDeleted": true }, { "id": 742576, "postDate": "2020-02-11T11:42:59.427Z", "rawMarkdown": "", "isDeleted": true }, { "id": 728722, "postDate": "2020-01-25T06:32:00.533Z", "rawMarkdown": "", "isDeleted": true }, { "id": 726419, "postDate": "2020-01-23T02:14:18.013Z", "rawMarkdown": "", "isDeleted": true }, { "id": 725676, "postDate": "2020-01-22T11:21:39.723Z", "rawMarkdown": "", "isDeleted": true }, { "id": 721000, "postDate": "2020-01-16T23:32:53.960Z", "rawMarkdown": "", "isDeleted": true }, { "id": 719344, "postDate": "2020-01-15T12:02:42.243Z", "rawMarkdown": "", "isDeleted": true }, { "id": 716584, "postDate": "2020-01-11T23:08:48.210Z", "rawMarkdown": "", "isDeleted": true }, { "id": 694027, "postDate": "2019-12-13T04:48:12.473Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 758820, "author_name": "Selim Seferbekov", "author_url": "", "post_date": "2020-02-28T07:35:42.460000", "content": "\n\nAs external data might be very helpful to obtain better scores on public/private leaderboards how are you going to validate solutions' compliance with the rules?\nThere is a possibility to cheat by using non allowed external data (or not posted here before merge deadline). I'm strongly against that but we are not in a perfect world and cheaters will always exist. \nIdeally winning solutions must be reproduced end-to-end: retrain, test - check that scores are quite similar to public/private leaderboards. This will ensure that winners used only allowed datasets.
", "votes": 25, "replies": [ { "id": 758877, "author_name": "Evgeny Nizhibitsky", "author_url": "", "post_date": "2020-02-28T09:27:16.747000", "content": "Also it would be nice to have a pinned post from organizers summarizing the approved datasets from all the comments here. Like \"Ok guys, the merge deadline is a thing now, here are the datasets that we approve: <...>. All the others, including <...> are not compatible with the license rule after the review.
\n\nAs for now some dataset links posted here can be just misleading and not all of them obviously got organizers attention to reply something like \"notice that this one will not be allowed\".
", "votes": 17, "replies": [] }, { "id": 758910, "author_name": "Janne Karttunen", "author_url": "", "post_date": "2020-02-28T10:25:07.640000", "content": "I agree with both. It is still unclear which external data is allowed and what isn't, since it seems like many of the declared datasets are under non-commercial licences and thus not allowed.
\n\nI would also like to know that are we allowed to use our self-collected videos from Youtube, if we post links to these videos here?
", "votes": 9, "replies": [] }, { "id": 759060, "author_name": "Emre Bayram", "author_url": "", "post_date": "2020-02-28T14:19:53.500000", "content": "\"Ideally winning solutions must be reproduced end-to-end: retrain, test - check that scores are quite similar to public/private leaderboards.\"
\n\nAlthough I agree with the concern and repeatibility is really important; I am not sure if repeatibility is always guaranteed for all training frameworks(maybe it is; i am an ML noob so i might be mistaken).\nIf at the end the scores are really close; this validation step can create a lot of confusion. What if one gets 3% worse results during retrain/test and the difference between 1st and 6th is less than 3%? does that mean the top one cheated? Also there is ensembling; so things will get really confusing. You have to reproduce all the single models in similar quality and the ensembled model should also produce the same good results.
\nIt is like \"if you win you have to prove you are not guilty\".
I agree the concern should be addressed but not sure retraining / reproducing the test results in a satisfactory way is always easily doable.
", "votes": 0, "replies": [] }, { "id": 759744, "author_name": "MPWARE", "author_url": "", "post_date": "2020-02-29T12:12:03.100000", "content": "+1 I still don't understand if we can use or not these datasets hosted on Kaggle:\nhttps://www.kaggle.com/tunguz/70000-real-faces-1\nhttps://www.kaggle.com/tunguz/70000-real-faces-2\nhttps://www.kaggle.com/tunguz/70000-real-faces-3\nhttps://www.kaggle.com/tunguz/70000-real-faces-4\nhttps://www.kaggle.com/tunguz/70000-real-faces-5\nhttps://www.kaggle.com/tunguz/70000-real-faces-6
\n\n@juliaelliott Data is hosted on Kaggle, but it does not mean we can use it, correct? We have to find out underlying license but is someone at Kaggle will review and give a green light? It requires a bit of legal skills to be sure.
", "votes": 3, "replies": [] }, { "id": 761672, "author_name": "Julia Elliott", "author_url": "", "post_date": "2020-03-02T20:45:02.410000", "content": "If your model makes use of any external data that is prohibited by the rules, then you are subject to disqualification by the host upon review of your solution, in particular if you are a prospective winner. As stated many times previously, if there are any restrictions imposed on the dataset's use (including non-commercial use only or restriction on those who have access to the dataset), that is considered in violation of the requirement that the data be \"available to use by all participants of the competition\" and therefore prohibited. This should be quite clear at this point.
", "votes": 0, "replies": [] }, { "id": 761681, "author_name": "Selim Seferbekov", "author_url": "", "post_date": "2020-03-02T20:57:26.280000", "content": "@juliaelliott That makes sense. Thank you!
\n\nSo basically we may use data that:\n- is available for commercial usage - this was not clear from the rules\n- and adheres to competition rules
\n\nWhat about datasets collected from youtube videos? Or just youtube videos? \nI think there are a lot of questions regarding them as they actually represent \"organic videos\".
", "votes": 8, "replies": [] }, { "id": 761735, "author_name": "Nuno Ferreira", "author_url": "", "post_date": "2020-03-02T22:10:07.373000", "content": "I think it makes sense, as this is not a commercial use. This is a competition with the goal of improving detection of deep fakes detection, which has prizes for the best models.\nSo if I understood the terms correctly as long as the data is available to all participants at no cost, it is disclosed in this thread and it's allowed to be used for this purpose, then it is ok.
\n\nFor most of the datasets you need to apply to have access to it and mentioned what you are going to use it for. So if you mention the kaggle competition, and they send you the link, then they are allowing you to use it for this competition.
\n\n@juliaelliott Can you just confirm or not, my statements above? This would help answering a lot of the questions present in this thread.
", "votes": 0, "replies": [] }, { "id": 761741, "author_name": "Selim Seferbekov", "author_url": "", "post_date": "2020-03-02T22:17:48.320000", "content": "Seems like you got it wrong. What @juliaelliott says is that we cannot use datasets if they are for research only.
", "votes": 0, "replies": [] }, { "id": 761752, "author_name": "Emre Bayram", "author_url": "", "post_date": "2020-03-02T22:29:29.977000", "content": "@selimsef does that make imagenet pretrained models(including model zoo of many libraries) + 90% of what is posted on this topic unusable?
", "votes": -1, "replies": [] }, { "id": 761756, "author_name": "Selim Seferbekov", "author_url": "", "post_date": "2020-03-02T22:39:01.407000", "content": "AFAIK we cannot use ImageNet dataset directly, but we can use public pretrained weights with open license.
", "votes": 0, "replies": [] }, { "id": 761757, "author_name": "Nuno Ferreira", "author_url": "", "post_date": "2020-03-02T22:40:50.010000", "content": "@emrebayram Yeah, if Selim is right that would be the case. Would be great if the organizers (@juliaelliott @addisonhoward) would give a final answer regarding this, cause at the moment I'm very confused regarding what we can, and what we cannot use. And I'm betting I'm not the only one
", "votes": 1, "replies": [] } ] }, { "id": 716551, "author_name": "ZFTurbo", "author_url": "", "post_date": "2020-01-11T21:25:02.777000", "content": "There are several DeepFake datasets avaialble:\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/
\n\nAll of them avaiable by request. Can we use them for developing solution or not? I want to give it a try, but don't want to waste time on them if they are forbidden.
", "votes": 11, "replies": [ { "id": 717128, "author_name": "", "author_url": "", "post_date": "2020-01-12T19:22:58.937000", "content": "DeeperForensics-1.0 also : https://github.com/EndlessSora/DeeperForensics-1.0 .
\n\n@juliaelliott @addisonhoward
\nAnd what about CC BY-NC 4.0 licence ? For example @tunguz shared 1 Million Fake Faces dataset.
\nCan we use pre-trained models and/or datasets under this (or similar for non-commercial research purposes only) license and expect to receive prize money?
I'm new to kaggle, so sorry for this possibly dummy question.
", "votes": 10, "replies": [] }, { "id": 718257, "author_name": "ZFTurbo", "author_url": "", "post_date": "2020-01-14T08:17:22", "content": "Is there any way to get DeeperForensics-1.0? As I can see there only \"Coming Soon\" and nothing more.
", "votes": 3, "replies": [] }, { "id": 721524, "author_name": "Nikolay Kovachev", "author_url": "", "post_date": "2020-01-17T13:03:25.260000", "content": "So are we allowed to use them or?
", "votes": 0, "replies": [] }, { "id": 721699, "author_name": "deepware.ai", "author_url": "", "post_date": "2020-01-17T16:12:50.043000", "content": "@zfturbo I've contacted them to get the publish date at least. However, it is not defined yet.
\n\nTheir response:\nCurrently, I cannot decide when to release the dataset due to some approval process. But I believe it will be soon. Once it is ready, we will make a notification on our GitHub page.
", "votes": 1, "replies": [] }, { "id": 721707, "author_name": "", "author_url": "", "post_date": "2020-01-17T16:21:57.233000", "content": "I think that we are not allowed to use them. But organizers have not answered yet...
\n\nAs for me, I haven't used any external datasets/models yet except ImageNet pretrained models in some experiments. I think that it can be a waste of time.
", "votes": 0, "replies": [] }, { "id": 721812, "author_name": "Julia Elliott", "author_url": "", "post_date": "2020-01-17T18:15:44.640000", "content": "@vladislavleketush and others - Your use of external data should conform to the external data specification in the competition's rules:
\n\n\n\n\nyou will (i) ensure the External Data is available to use by all participants of the competition for purposes of the competition at no cost to the other participants and (ii) post such access to the External Data for the participants to the official competition forum prior to the Entry Deadline.
\n
Therefore, licenses which place restrictions on datasets' use (whether by purpose, affiliation, cost or other restrictive means) would be in violation of this rule.
", "votes": 8, "replies": [] }, { "id": 728872, "author_name": "SeshuRaju 🧘♂️", "author_url": "", "post_date": "2020-01-25T12:17:52.423000", "content": "March 3, 2020 - Entry deadline. You must accept the competition rules before this date in order to compete. (is it fair sharing external data at end of competition?)\n- we treat these external store are magic features till March 3rd, 2020?
\n\nPlease clarify if i miss understood
", "votes": 1, "replies": [] }, { "id": 730874, "author_name": "Julia Elliott", "author_url": "", "post_date": "2020-01-28T03:36:32.123000", "content": "@seshurajup - yes, the external data declaration deadline on this forum thread is the entry deadline, March 3, 2020.
", "votes": 1, "replies": [] } ] }, { "id": 694492, "author_name": "Julia Elliott", "author_url": "", "post_date": "2019-12-13T17:40:14.113000", "content": "For clarity on the datasets that need to be declared here: You do not need to declare your own self-trained/original work models. But if you are using a dataset or pre-trained model obtained externally which is licensed with the right for you to use, then that dataset needs to be declared.
\n\nSo, for example: if you are using the imagenet dataset and efficientnet to train your model(s), you must declare imagenet and efficientnet on this thread, but you don’t need to make your trained model public.
\n\nYou also do not need to declare things that have already been declared.
", "votes": 9, "replies": [ { "id": 702696, "author_name": "David Bonn", "author_url": "", "post_date": "2019-12-25T02:58:03.390000", "content": "Just for clarity, would that mean if I wanted to use one of the Keras pre-trained models as part of this competition I would need to declare it here?
\n\nSorry if that is a noob question.
", "votes": 0, "replies": [] }, { "id": 712382, "author_name": "Shitai Li", "author_url": "", "post_date": "2020-01-07T07:02:49.667000", "content": "Sorry I'm still a bit confused.
\n\n1) you mentioned use imagenet as external data source. But isn't imagenet larger than 1Gb?\n2) can i create my own deepfake dataset and use it to train my model? do i have to declare the dataset\n3) can i create my own deepfake dataset and use it to PRE-train my model? do i have to just decalre the pretrained model or the original dataset? does the 1 Gb rule apply to the whole dataset or just the pretrained model?
", "votes": 0, "replies": [] }, { "id": 713022, "author_name": "Julia Elliott", "author_url": "", "post_date": "2020-01-07T20:21:49.673000", "content": "@davidbonn Yes, the pre-trained model should be declared.
", "votes": 0, "replies": [] }, { "id": 713041, "author_name": "Julia Elliott", "author_url": "", "post_date": "2020-01-07T20:38:57.670000", "content": "@shitaili \n1) The 1GB external data requirement is a systematic constraint on any notebook you submit. So, you can use large external datasets (like imagenet) that exceed 1GB if you train offline, and then upload that trained model into your submission notebook. The total of external datasets uploaded to the submission notebook you commit in Kaggle must not exceed 1GB.\n2) Yes, you can create your own deepfake dataset if you have the right to use and share the videos/dataset that you are using. You must declare the source datasets that are used.\n3) The original dataset. See #1 about the 1GB constraint.
", "votes": 0, "replies": [] }, { "id": 725200, "author_name": "", "author_url": "", "post_date": "2020-01-21T22:36:26.943000", "content": "", "votes": 0, "replies": [] }, { "id": 725241, "author_name": "Julia Elliott", "author_url": "", "post_date": "2020-01-21T23:34:13.180000", "content": "@mindcool The notebook in Kaggle that you are submitted is not included. But any trained models or other external data that is being used in your submission notebook is included in the 1GB.
", "votes": 0, "replies": [] }, { "id": 753674, "author_name": "Anton Pimenov", "author_url": "", "post_date": "2020-02-22T14:35:17.847000", "content": "Imagenet is non-commercial usage only. Below in this thread, you texted:
\n\n\n\n\n\"I’ve answered the question about BY-NC not being available for use by all (non-commercial use) and therefore violating the requirement that external data be available for use by all participants.\"
\n
and
\n\n\n\n\n\"You raise a fair question. The rule indicates to use by all participants -- Inherent in data licensing terms that constrain the data's use is the restriction that it is not available to be used by all participants. In many cases with research/academic-use licensing, only those in a research profession may gain access to the dataset. Sure, in some cases, everyone may be able to physically download it, but if a license prohibits its use for non-research uses, then it can be interpreted as also prohibiting use by users who are non-researching. So the issue here is one of legally-compliant accessibility to use of the dataset by all.\"
\n
Does it mean, that we can not use the imagenet as well?
", "votes": 2, "replies": [] }, { "id": 753681, "author_name": "", "author_url": "", "post_date": "2020-02-22T14:43:43.273000", "content": "You're asking just to ask or you've already read this ?
", "votes": 0, "replies": [] }, { "id": 753701, "author_name": "Anton Pimenov", "author_url": "", "post_date": "2020-02-22T15:11:10.370000", "content": "Yes, I've been read this, but I really confused about this. \nPoint 1 for agreement to download Imagenet:
\n\n\n\n\n\n
\n- Researcher shall use the Database only for non-commercial research and educational purposes.
\n
Ok, but we can download it by direct link, is it enough to this contest? How we can get proof, that the pre-trained model was not used original images. Typically, code for pre-trained models does not contain a part for download by direct links.
", "votes": 2, "replies": [] }, { "id": 753806, "author_name": "Human Analog", "author_url": "", "post_date": "2020-02-22T16:58:51.543000", "content": "\n\n\nthat we can not use the imagenet as well
\n
ImageNet is available for download on Kaggle.
", "votes": 0, "replies": [] }, { "id": 753831, "author_name": "Emre Bayram", "author_url": "", "post_date": "2020-02-22T17:41:18.667000", "content": "@i7p9h9 I have tried to find answer to the same question. Searched internet and checked similar concerns/questions ( see this and that\n ). None of them have clear answer or no answer at all.
\n\nFirst of all I am not a lawyer.\nBut obviously; there is no clear answer. The reason why it is not clear is there is no court decided for a similar issue.\nImagenet FAQ says the database is for non-commercial; but what is the definition of the database there? Using the images as it is? Using the original images? Or any derived information from these images? Then one would claim that google street view is violating many laws(maybe this will be the case in the future) because they did not get my permission to take a photo of my apartment and use it in a commercial app.
\n\nSo it is hard for any organization to make a clear comment about this. I am sure many commercial apps as of today are using imagenet pretrained models because it is part of many library's model zoo.
\n\nIt is also hard for a judge to decide in my opinion. Because it is a complex topic.\nAs a human being, if I look at an image and learn something from it; can I use this knowledge for commercial gain(for example for teaching)? Probably yes. But probably it is not ok to copy-paste the material as it is.\nIn my opinion we are still in an early era for things like this to settle. It is like 15-20 years ago noone cared about \"personal data\"; but as of today there is GDPR and many other regulations.\nProbably for deep learning 5-15 years from now the regulations will be there about data usage but for now it is just kind of impossible to directly say \"this is ok\" or \"this is not\".
\n\nFor this competition at the moment I think every competitor need to use their own judgement about this.
", "votes": 1, "replies": [] }, { "id": 753902, "author_name": "Anton Pimenov", "author_url": "", "post_date": "2020-02-22T19:55:05.773000", "content": "@emrebayram, thank you for the detailed answer. I hope participants will not any have any problem due to using imagenet pre-trained models in this competition.
", "votes": 1, "replies": [] } ] }, { "id": 760710, "author_name": "Yassine Alouini", "author_url": "", "post_date": "2020-03-01T16:33:10.880000", "content": "Some people have surfaced this already but let me ask again: it would be nice to have the final list of allowed external resources under the post. It probably requires some effort to create this list and check it but will be very useful to everyone since the deadline is approaching. \nWhat do you think @addisonhoward?
", "votes": 7, "replies": [] }, { "id": 706023, "author_name": "Carlo", "author_url": "", "post_date": "2019-12-29T20:24:05.257000", "content": "@tunguz' datasets for real faces and fake faces.
\n\nReal Faces:\nhttps://www.kaggle.com/tunguz/70000-real-faces-1
\n\nFake Faces:\nhttps://www.kaggle.com/tunguz/1-million-fake-faces\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-2\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-3\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-4\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-5\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-6\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-7
", "votes": 7, "replies": [ { "id": 707297, "author_name": "Carlo", "author_url": "", "post_date": "2019-12-31T15:38:15.370000", "content": "Also, The Flickr-Faces-HQ (FFHQ) Dataset of 70000 real faces:\nhttps://github.com/NVlabs/ffhq-dataset
", "votes": 2, "replies": [] }, { "id": 745157, "author_name": "Emre Bayram", "author_url": "", "post_date": "2020-02-13T15:05:51.690000", "content": "@carlolepelaars careful the competition rules state that the models should be available for commercial use. 1-million-fake-faces dataset is under \"Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)\" license so probable is not ok. @juliaelliott am I correct?
", "votes": 2, "replies": [] }, { "id": 745167, "author_name": "", "author_url": "", "post_date": "2020-02-13T15:21:32.127000", "content": "\n\nI've asked the same question. Check this
", "votes": 2, "replies": [] }, { "id": 745173, "author_name": "Carlo", "author_url": "", "post_date": "2020-02-13T15:26:05.667000", "content": "Ok, thanks for the heads-up guys! I'm not using these datasets in my current modelling so its fine.
", "votes": 1, "replies": [] } ] }, { "id": 763235, "author_name": "Sirish Somanchi", "author_url": "", "post_date": "2020-03-04T09:21:52.830000", "content": "By the way, a list of permitted external data and allowed pre-trained models (using approved licenses) should be maintained as a \"Kaggle-level\" topic because we have the same questions repeated across competitions.
", "votes": 6, "replies": [] }, { "id": 701432, "author_name": "Darragh", "author_url": "", "post_date": "2019-12-23T13:37:49.860000", "content": "```\nUADFV: The UADFV dataset - Xin Yang, Yuezun Li, and Siwei Lyu. Exposing deep fakes\nusing inconsistent head poses. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),\n2019.
\n\nDF-TIMIT: Pavel Korshunov and Sebastien Marcel. Deepfakes: a new ´\nthreat to face recognition? assessment and detection. arXiv\npreprint arXiv:1812.08685, 2018.
\n\nFF-DF: Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Chris- ¨\ntian Riess, Justus Thies, and Matthias Nießner. FaceForensics++: Learning to detect manipulated facial images. In\nICCV, 2019.
\n\nDFD: Nicholas Dufour, Andrew Gully, Per Karlsson, Alexey Victor Vorbyov, Thomas Leung, Jeremiah Childs, and Christoph Bregler. Deepfakes detection dataset by google & jigsaw.\n```
", "votes": 5, "replies": [ { "id": 737828, "author_name": "Nikolay Kovachev", "author_url": "", "post_date": "2020-02-05T20:25:30.737000", "content": "@juliaelliott @addisonhoward That dataset is not publicly available. I've apply few weeks ago and didn't got approved. In other hands the pretrained models are free to download. Can we use them in such case?
", "votes": 0, "replies": [] }, { "id": 737830, "author_name": "Nikolay Kovachev", "author_url": "", "post_date": "2020-02-05T20:29:37.170000", "content": "I just apply again and got approved immediatelly. Apologies for the false alert!
", "votes": 0, "replies": [] } ] }, { "id": 694433, "author_name": "btk1", "author_url": "", "post_date": "2019-12-13T15:35:10.093000", "content": "Static build of FFMPEG: https://johnvansickle.com/ffmpeg/
", "votes": 6, "replies": [ { "id": 695069, "author_name": "btk1", "author_url": "", "post_date": "2019-12-14T14:13:28.150000", "content": "@juliaelliott Are we allowed to use the Static Build of FFMPEG from a dataset like https://www.kaggle.com/rakibilly/ffmpeg-static-build ?
", "votes": 4, "replies": [] }, { "id": 713043, "author_name": "Julia Elliott", "author_url": "", "post_date": "2020-01-07T20:39:45.733000", "content": "@rakibilly I don't see any problem here if it's publicly available.
", "votes": 2, "replies": [] }, { "id": 725528, "author_name": "Zeta", "author_url": "", "post_date": "2020-01-22T08:03:46.497000", "content": "I'm using it to extract the audio into a wav file from the original mp4 file.
", "votes": 1, "replies": [] } ] }, { "id": 748358, "author_name": "YoonSoo", "author_url": "", "post_date": "2020-02-17T12:35:31.360000", "content": "Edit: Nevermind; the metadata has individual license info.
", "votes": 0, "replies": [] }, { "id": 750285, "author_name": "Neuron Engineer", "author_url": "", "post_date": "2020-02-19T09:03:35.493000", "content": "@addisonhoward @juliaelliott Could you please clarify the following ?
\n\nThe metadata itself is said to be \"CC-by-NC\" on Nvidia website, so using the metadata to select \"public domain\" pictures are not allowed ?
", "votes": 2, "replies": [] } ] }, { "id": 767591, "author_name": "Ian Pan", "author_url": "", "post_date": "2020-03-09T21:05:36.780000", "content": "@juliaelliott In the rules, it says:
\n\n\n\n\nNo using other Kaggle notebooks or utility scripts as inputs to your submission notebook. External data also cannot be externally-referenced (i.e. via BigQuery, Github, external URL). Instead, load external models or datasets directly as an external data source.
\n
I upload my own scripts as a Kaggle dataset so I can import functions from them and not have an excessively long notebook. Is that allowed?
", "votes": 4, "replies": [ { "id": 779756, "author_name": "Julia Elliott", "author_url": "", "post_date": "2020-03-19T16:57:22.937000", "content": "Yes. This stipulation in the Code Requirements is a systematic one - in that you can't daisy-chain kernels/notebooks into your submission notebook. But you can definitely download those notebooks and upload them as inputs into your submission. Hence the Instead, load external models or datasets directly as an external data source.
InceptionV3 imagenet weights from Keras:\nhttps://keras.io/applications/#inceptionv3
\n\nUFC101 Khurram Soomro, Amir Roshan Zamir and Mubarak Shah, UCF101: A Dataset of 101 Human Action Classes From Videos in The Wild., CRCV-TR-12-01, November, 2012.
\n\nhttp://crcv.ucf.edu/data/UCF101/
", "votes": 3, "replies": [ { "id": 727959, "author_name": "Zeta", "author_url": "", "post_date": "2020-01-24T09:04:47.367000", "content": "I'm going to use this too.
", "votes": 1, "replies": [] } ] }, { "id": 716317, "author_name": "Emre Bayram", "author_url": "", "post_date": "2020-01-11T14:43:08.087000", "content": "facenet-pytorch pre-trained models: https://github.com/timesler/facenet-pytorch\n(MIT license)
", "votes": 3, "replies": [] }, { "id": 693664, "author_name": "ZFTurbo", "author_url": "", "post_date": "2019-12-12T16:49:41.610000", "content": "Open Images Dataset: https://storage.googleapis.com/openimages/web/index.html\nObjects 365 Dataset: https://www.objects365.org/overview.html\nCOCO Dataset: http://cocodataset.org/#home
", "votes": 3, "replies": [] }, { "id": 692907, "author_name": "Bruno G. do Amaral", "author_url": "", "post_date": "2019-12-11T20:53:11.190000", "content": "“External data is allowed up to 1 GB in size. External data must be freely & publicly available, including pre-trained models”
\n\nDoes it count during training time? Or just for predict time?
", "votes": 3, "replies": [ { "id": 692929, "author_name": "Julia Elliott", "author_url": "", "post_date": "2019-12-11T21:16:43.323000", "content": "Good question! The 1 GB restriction is only on your submission notebook in Kaggle, so since you should be training elsewhere, this only applies to your inference/predictions notebook.
", "votes": 8, "replies": [] }, { "id": 693976, "author_name": "Janne Karttunen", "author_url": "", "post_date": "2019-12-13T02:19:09.333000", "content": "To make sure, are we allowed to use private datasets (and larger than 1 GB) during the training phase? Or is the training limited only on the provided dataset?
", "votes": 0, "replies": [] }, { "id": 701058, "author_name": "hirviö", "author_url": "", "post_date": "2019-12-23T03:07:14.303000", "content": "the 1GB limit is only on the data used for final inference. Training can be done on any dataset as long as you mention it here (and have the right to use/mention it).
", "votes": 1, "replies": [] }, { "id": 706022, "author_name": "Carlo", "author_url": "", "post_date": "2019-12-29T20:22:53.757000", "content": "Awesome!
", "votes": 1, "replies": [] } ] }, { "id": 762894, "author_name": "Denis Timoshenko", "author_url": "", "post_date": "2020-03-03T23:15:21.877000", "content": "https://www.youtube.com/ Creative Commons videos\nhttps://drive.google.com/file/d/1NeyTFAwdJSjxA9ZtdviwdUjdptEVjM_i\nhttps://www.kaggle.com/tunguz/1-million-fake-faces\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-2\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-3\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-4\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-5\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-6\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-7\nhttps://github.com/IDRnD/LCC_FASD
", "votes": 2, "replies": [] }, { "id": 762923, "author_name": "Hope46", "author_url": "", "post_date": "2020-03-03T23:58:28.157000", "content": "FACENET, VGG16 https://github.com/davidsandberg/facenet\nMTCNN\nhttps://github.com/ipazc/mtcnn
", "votes": 1, "replies": [] }, { "id": 762654, "author_name": "Mohsin hasan", "author_url": "", "post_date": "2020-03-03T17:28:41.200000", "content": "https://github.com/huawei-noah/ghostnet\nhttps://github.com/iamhankai/ghostnet.pytorch\nhttps://github.com/kuan-wang/pytorch-mobilenet-v3
", "votes": 1, "replies": [] }, { "id": 762339, "author_name": "Artyom Lyan", "author_url": "", "post_date": "2020-03-03T12:52:13.157000", "content": "https://www.thispersondoesnotexist.com/
", "votes": 1, "replies": [] }, { "id": 762337, "author_name": "Dmitry Abulkhanov", "author_url": "", "post_date": "2020-03-03T12:49:36.790000", "content": "All media with permitting CC license from:\nhttps://badoo.com/\nhttps://www.instagram.com/\nhttps://www.youtube.com/\nhttps://www.flickr.com/\nhttps://www.google.com/imghp\nhttps://yandex.ru/images/\nFrom this thread: https://www.kaggle.com/c/sp-society-camera-model-identification/discussion/47235
\n\nIf any source violates the rules, I will not use it. List it here just in case :)
", "votes": 1, "replies": [] }, { "id": 762257, "author_name": "Deressa", "author_url": "", "post_date": "2020-03-03T11:26:55.257000", "content": "pytorch vgg16 model\nwheight https://download.pytorch.org/models/vgg16-397923af.pth\npytorch vgg19 model\nvgg19: https://download.pytorch.org/models/vgg19-dcbb9e9d.pth\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\npretrained model: https://drive.google.com/drive/folders/1zvl89AgFAApbH0At-gMuZSeQB_LpNP-M\nhttps://github.com/timesler/facenet-pytorch\nmodel 20180408-102900: https://drive.google.com/uc?export=download&id=12DYdlLesBl3Kk51EtJsyPS8qA7fErWDX\nmodel 20180402-114759: https://drive.google.com/uc?export=download&id=1TDZVEBudGaEd5POR5X4ZsMvdsh1h68T1\nhttps://github.com/mit-han-lab/temporal-shift-module\nTSN ResNet50: https://hanlab.mit.edu/projects/tsm/models/TSM_kinetics_RGB_resnet50_avg_segment5_e50.pth\nTSM ResNet50: https://hanlab.mit.edu/projects/tsm/models/TSM_kinetics_RGB_resnet50_shift8_blockres_avg_segment8_e100_dense.pth\nhttps://github.com/ufoym/imbalanced-dataset-sampler\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch/releases/\nhttps://github.com/NVlabs/ffhq-dataset\nhttps://github.com/deepinsight/insightface/\nhttps://github.com/deepinsight/insightface/#pretrained-models\nhttps://github.com/Sierkinhane/mtcnn-pytorch\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nweight: https://drive.google.com/file/d/1WeXlNYsM6dMP3xQQELI-4gxhwKUQxc3-/view?usp=sharing\nhttps://github.com/lijiannuist/lightDSFD\nweight: https://github.com/lijiannuist/lightDSFD/raw/master/weights/light_DSFD.pth\nhttps://github.com/cydonia999/VGGFace2-pytorch\nhttps://github.com/cydonia999/VGGFace2-pytorch#pretrained-models\nhttps://github.com/albumentations-team/albumentations
", "votes": 1, "replies": [] }, { "id": 762097, "author_name": "Chew Kok Wah", "author_url": "", "post_date": "2020-03-03T07:53:09.040000", "content": "@juliaelliott \nTo ensure Fairness and Transparency, upon Competition ended and preliminary Private leaderboard score and place has been calculated, will you ask the top 10 Kernels to fully disclose the External dataset they used (together with their kernel code) for Public scrutiny BEFORE the confirmation of any award? This will allowed the Kaggle community to help organizer in catching any form of cheating or overlook.
", "votes": 1, "replies": [ { "id": 763233, "author_name": "Sirish Somanchi", "author_url": "", "post_date": "2020-03-04T09:20:43.430000", "content": "As per the rules of this competition, the winners' licensing terms require that the solution be open sourced, and we can all inspect the external data used by the winners along with their complete source code.
\n\nAlso, I believe that the organizer will reproduce the winning solutions end-to-end: retrain, test, and validate scores. During this process, they will validate all external data and licenses as well.
\n\nPS: We can ensure repeatability of scores by properly setting the seeds for random number generators in numpy, Torch, TF/Keras.
", "votes": 0, "replies": [] } ] }, { "id": 762082, "author_name": "Chew Kok Wah", "author_url": "", "post_date": "2020-03-03T07:38:13.020000", "content": "https://github.com/hukkelas/DSFD-Pytorch-Inference\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/sfzhang15/FaceBoxes\nhttps://github.com/TropComplique/FaceBoxes-tensorflow\nhttps://github.com/XiaXuehai/faceboxes\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/peteryuX/retinaface-tf2\nhttps://github.com/OFRIN/Tensorflow_RetinaFace\nhttps://github.com/yeephycho/tensorflow-face-detection\nhttps://v.qq.com/channel/choice\nwww.youtube.com\nwww.youku.com\nwww.tudou.com\nwww.tiktok.com\nwww.douyin.com\nwww.tonton.com.my\nhttp://tv.cctv.com/\nwww.bbc.com\nwww.cnn.com\nwww.foxnews.com\nwww.iqiyi.com\nhttps://github.com/deepmind/kinetics-i3d\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/yaohungt/Gated-Spatio-Temporal-Energy-Graph\nhttps://github.com/dlpbc/keras-kinetics-i3d\nhttps://github.com/huangyangyu/SeqFace\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/ronghuaiyang/arcface-pytorch\nhttps://github.com/happynear/AMSoftmax\nhttps://github.com/foamliu/InsightFace-v2\nhttps://github.com/1996scarlet/ArcFace-Multiplex-Recognition\nhttps://github.com/xiaoboCASIA/SV-X-Softmax\nhttps://github.com/foamliu/InsightFace-PyTorch\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/iloveuu2011/Face-Detection-Library\nhttps://github.com/iloveuu2011/retinaface-tf2\nhttps://github.com/pvskand/DisguiseNet\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/EricZgw/PyramidBox\nhttps://github.com/swghosh/DeepFace\nhttps://github.com/zma-c-137/VarGFaceNet\nhttps://github.com/cvtower/seesawfacenet_pytorch\nhttps://github.com/kk7nc/RMDL\nhttps://github.com/ZhaoJ9014/High-Performance-Face-Recognition\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/rlaengud123/CMC_LRCN\nhttps://github.com/doronharitan/human_activity_recognition_LRCN\nhttps://github.com/piergiaj/representation-flow-cvpr19\nhttps://github.com/piergiaj/evanet-iccv19\nhttps://github.com/piergiaj\nhttps://github.com/piergiaj/mlb-youtube\nhttps://github.com/craston/MARS\nhttps://github.com/clancylian/retinaface\nhttps://github.com/yangfly\nhttps://pan.baidu.com/share/init?surl=P1ypO7VYUbNAezdvLm2m9w\nhttps://github.com/biubug6/Face-Detector-1MB-with-landmark\nhttps://github.com/610265158/DSFD-tensorflow
", "votes": 1, "replies": [] }, { "id": 761474, "author_name": "deepware.ai", "author_url": "", "post_date": "2020-03-02T14:59:06.247000", "content": "A python package to analyze and compare voices with deep learning\nhttps://github.com/resemble-ai/Resemblyzer
Real-Time Voice Cloning\nhttps://github.com/CorentinJ/Real-Time-Voice-Cloning
LibriSpeech ASR corpus\nhttp://www.openslr.org/12/
LibriTTS corpus\nhttp://www.openslr.org/60/
The VoxCeleb Dataset\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1.html\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox2.html
The M-AILABS Speech Dataset\nhttps://www.caito.de/2019/01/the-m-ailabs-speech-dataset/
MrDeepfakes\nhttps://mrdeepfakes.com/terms
English Multi-speaker Corpus for CSTR Voice Cloning Toolkit\nhttps://homepages.inf.ed.ac.uk/jyamagis/page3/page58/page58.html\nhttps://datashare.is.ed.ac.uk/handle/10283/2651
FaceForensics original repository\nhttps://github.com/ondyari/FaceForensics/tree/original
IEEE's Signal Processing Society - Camera Model Identification\nhttps://www.kaggle.com/c/sp-society-camera-model-identification
BlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch
", "votes": 1, "replies": [ { "id": 763741, "author_name": "Sky", "author_url": "", "post_date": "2020-03-04T20:01:29.143000", "content": "https://github.com/AlexanderParkin/ChaLearn_liveness_challenge
", "votes": 0, "replies": [] } ] }, { "id": 760744, "author_name": "Dashti", "author_url": "", "post_date": "2020-03-01T17:28:03.687000", "content": "dlib: https://github.com/davisking/dlib\ndlib-models: https://github.com/davisking/dlib-models/blob/master/shape_predictor_68_face_landmarks.dat.bz2\nnumpy: https://github.com/numpy/numpy\nopencv: https://github.com/opencv/opencv\nscipy: https://github.com/scipy/scipy\nimutils: https://pypi.org/project/imutils/\nripser: https://github.com/scikit-tda/ripser.py\nglob: https://docs.python.org/3/library/glob.html\nos: https://docs.python.org/3/library/os.html\nmultiprocessing: https://docs.python.org/3/library/multiprocessing.html\nscikit-learn: https://scikit-learn.org/stable/index.html\nScikit-TDA: https://github.com/scikit-tda\ncsv: https://docs.python.org/3/library/csv.html
", "votes": 1, "replies": [] }, { "id": 758759, "author_name": "idcore", "author_url": "", "post_date": "2020-02-28T05:41:37.520000", "content": "opencv extra modules\nhttps://github.com/opencv/opencv_contrib
\n\npytables\nhttps://pypi.org/project/tables/
\n\nh5py\nhttps://pypi.org/project/h5py/
\n\npandas\nhttps://pypi.org/project/pandas/
\n\nalbumetations\nhttps://pypi.org/project/albumentations/
\n\nTacotron2\nhttps://github.com/NVIDIA/tacotron2
\n\nWaveglow\nhttps://github.com/NVIDIA/waveglow
\n\nParallelWaveGAN\nhttps://github.com/kan-bayashi/ParallelWaveGAN
\n\nMozilla TTS\nhttps://github.com/mozilla/TTS
\n\nCatalyst framework\nhttps://github.com/catalyst-team/catalyst
\n\nPytorch-toolbelt & Examples\nhttps://github.com/BloodAxe/pytorch-toolbelt\nhttps://github.com/BloodAxe/Kaggle-2019-Blindness-Detection
\n\nKornia package\nhttps://github.com/kornia/kornia
", "votes": 1, "replies": [] }, { "id": 758018, "author_name": "Human Analog", "author_url": "", "post_date": "2020-02-27T11:12:15.163000", "content": "https://github.com/yxlijun/S3FD.pytorch
", "votes": 1, "replies": [] }, { "id": 757847, "author_name": "Aashish Ghosh", "author_url": "", "post_date": "2020-02-27T07:02:35.427000", "content": "Pretrained models disclosure from two repositories:\nhttps://pytorch.org/docs/stable/torchvision/models.html\nand\nhttps://modelzoo.co/framework/pytorch
", "votes": 1, "replies": [] }, { "id": 757253, "author_name": "vecxoz", "author_url": "", "post_date": "2020-02-26T15:16:09.447000", "content": "https://keras.io/applications\nhttps://www.tensorflow.org/lite/models\nhttps://github.com/NVIDIA/tensorflow-determinism\nhttps://github.com/tensorflow/models/blob/master/official/README.md\nhttps://github.com/tensorflow/models/blob/master/official/README-TPU.md
", "votes": 1, "replies": [ { "id": 757807, "author_name": "Zeta", "author_url": "", "post_date": "2020-02-27T05:54:40.543000", "content": "+1
", "votes": 0, "replies": [] } ] }, { "id": 756800, "author_name": "", "author_url": "", "post_date": "2020-02-26T05:02:26.607000", "content": "", "votes": 1, "replies": [] }, { "id": 756063, "author_name": "", "author_url": "", "post_date": "2020-02-25T11:30:07.583000", "content": "", "votes": 1, "replies": [] }, { "id": 755549, "author_name": "", "author_url": "", "post_date": "2020-02-24T21:57:53.050000", "content": "", "votes": 1, "replies": [] }, { "id": 753252, "author_name": "", "author_url": "", "post_date": "2020-02-21T23:38:53.847000", "content": "", "votes": 1, "replies": [] }, { "id": 750901, "author_name": "", "author_url": "", "post_date": "2020-02-19T19:56:42.073000", "content": "", "votes": 1, "replies": [] }, { "id": 750278, "author_name": "", "author_url": "", "post_date": "2020-02-19T08:48:53.387000", "content": "", "votes": 1, "replies": [ { "id": 750898, "author_name": "", "author_url": "", "post_date": "2020-02-19T19:51:11.923000", "content": "", "votes": 0, "replies": [] }, { "id": 752690, "author_name": "", "author_url": "", "post_date": "2020-02-21T09:59:47.333000", "content": "", "votes": 1, "replies": [] } ] }, { "id": 749115, "author_name": "", "author_url": "", "post_date": "2020-02-18T10:07:37.107000", "content": "", "votes": 1, "replies": [] }, { "id": 748620, "author_name": "", "author_url": "", "post_date": "2020-02-17T18:51:56.490000", "content": "", "votes": 1, "replies": [] }, { "id": 740961, "author_name": "", "author_url": "", "post_date": "2020-02-10T02:30:58.627000", "content": "", "votes": 1, "replies": [] }, { "id": 737834, "author_name": "", "author_url": "", "post_date": "2020-02-05T20:40:10.610000", "content": "", "votes": 1, "replies": [] }, { "id": 736429, "author_name": "", "author_url": "", "post_date": "2020-02-04T07:04:34.133000", "content": "", "votes": 1, "replies": [] }, { "id": 730492, "author_name": "", "author_url": "", "post_date": "2020-01-27T14:53:11.950000", "content": "", "votes": 1, "replies": [] }, { "id": 707331, "author_name": "", "author_url": "", "post_date": "2019-12-31T16:52:23.010000", "content": "", "votes": 1, "replies": [] }, { "id": 705166, "author_name": "", "author_url": "", "post_date": "2019-12-28T15:20:46.887000", "content": "", "votes": 1, "replies": [] }, { "id": 704584, "author_name": "", "author_url": "", "post_date": "2019-12-27T16:50:18.093000", "content": "", "votes": 1, "replies": [] }, { "id": 701639, "author_name": "", "author_url": "", "post_date": "2019-12-23T17:49:52.270000", "content": "", "votes": 1, "replies": [] }, { "id": 695988, "author_name": "", "author_url": "", "post_date": "2019-12-16T01:02:16.160000", "content": "", "votes": 1, "replies": [ { "id": 743205, "author_name": "", "author_url": "", "post_date": "2020-02-11T22:02:57.940000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 695274, "author_name": "", "author_url": "", "post_date": "2019-12-14T22:13:24.190000", "content": "", "votes": 1, "replies": [] }, { "id": 762816, "author_name": "", "author_url": "", "post_date": "2020-03-03T20:30:55.987000", "content": "", "votes": 2, "replies": [] }, { "id": 761637, "author_name": "", "author_url": "", "post_date": "2020-03-02T19:44:19.143000", "content": "", "votes": 2, "replies": [] }, { "id": 761391, "author_name": "", "author_url": "", "post_date": "2020-03-02T13:15:24.470000", "content": "", "votes": 2, "replies": [ { "id": 762013, "author_name": "", "author_url": "", "post_date": "2020-03-03T05:47:11.137000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 760263, "author_name": "", "author_url": "", "post_date": "2020-03-01T03:35:52.287000", "content": "", "votes": 2, "replies": [] }, { "id": 759185, "author_name": "", "author_url": "", "post_date": "2020-02-28T17:22:28.390000", "content": "", "votes": 2, "replies": [] }, { "id": 758640, "author_name": "", "author_url": "", "post_date": "2020-02-28T01:59:33.917000", "content": "", "votes": 2, "replies": [] }, { "id": 742139, "author_name": "", "author_url": "", "post_date": "2020-02-11T05:05:10.307000", "content": "", "votes": 2, "replies": [] }, { "id": 762386, "author_name": "", "author_url": "", "post_date": "2020-03-03T13:38:51.120000", "content": "", "votes": 0, "replies": [] }, { "id": 760274, "author_name": "", "author_url": "", "post_date": "2020-03-01T03:59:05.320000", "content": "", "votes": 0, "replies": [] }, { "id": 732874, "author_name": "", "author_url": "", "post_date": "2020-01-30T12:08:56.300000", "content": "", "votes": 2, "replies": [] }, { "id": 723979, "author_name": "", "author_url": "", "post_date": "2020-01-20T17:14:24.320000", "content": "", "votes": 2, "replies": [ { "id": 728400, "author_name": "", "author_url": "", "post_date": "2020-01-24T17:22:59.320000", "content": "", "votes": 2, "replies": [] }, { "id": 732964, "author_name": "", "author_url": "", "post_date": "2020-01-30T13:47:42.893000", "content": "", "votes": 2, "replies": [] } ] }, { "id": 723311, "author_name": "", "author_url": "", "post_date": "2020-01-19T20:35:53.563000", "content": "", "votes": 2, "replies": [ { "id": 728468, "author_name": "", "author_url": "", "post_date": "2020-01-24T18:52:12.643000", "content": "", "votes": 1, "replies": [] } ] }, { "id": 722928, "author_name": "", "author_url": "", "post_date": "2020-01-19T10:15:58.280000", "content": "", "votes": 2, "replies": [ { "id": 727343, "author_name": "", "author_url": "", "post_date": "2020-01-23T16:37:15.350000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 719539, "author_name": "", "author_url": "", "post_date": "2020-01-15T15:24:11", "content": "", "votes": 2, "replies": [] }, { "id": 705785, "author_name": "", "author_url": "", "post_date": "2019-12-29T13:20:11.673000", "content": "", "votes": 2, "replies": [] }, { "id": 693452, "author_name": "", "author_url": "", "post_date": "2019-12-12T11:39:02.057000", "content": "", "votes": 2, "replies": [ { "id": 693550, "author_name": "", "author_url": "", "post_date": "2019-12-12T13:55:36.580000", "content": "", "votes": 2, "replies": [] }, { "id": 693874, "author_name": "", "author_url": "", "post_date": "2019-12-12T21:58:24.897000", "content": "", "votes": 2, "replies": [] }, { "id": 694702, "author_name": "", "author_url": "", "post_date": "2019-12-14T01:37:15.203000", "content": "", "votes": 0, "replies": [] }, { "id": 712381, "author_name": "", "author_url": "", "post_date": "2020-01-07T07:01:54.423000", "content": "", "votes": 0, "replies": [] }, { "id": 717173, "author_name": "", "author_url": "", "post_date": "2020-01-12T20:28:03.380000", "content": "", "votes": 0, "replies": [] }, { "id": 717190, "author_name": "", "author_url": "", "post_date": "2020-01-12T21:05:49.643000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 769124, "author_name": "", "author_url": "", "post_date": "2020-03-11T14:50:46.837000", "content": "", "votes": 1, "replies": [] }, { "id": 759242, "author_name": "", "author_url": "", "post_date": "2020-02-28T19:17:16.817000", "content": "", "votes": -1, "replies": [] }, { "id": 703168, "author_name": "", "author_url": "", "post_date": "2019-12-25T18:07:41.013000", "content": "", "votes": 1, "replies": [] }, { "id": 761667, "author_name": "", "author_url": "", "post_date": "2020-03-02T20:39:51.413000", "content": "", "votes": -1, "replies": [] }, { "id": 756837, "author_name": "", "author_url": "", "post_date": "2020-02-26T06:09:04.353000", "content": "", "votes": -1, "replies": [] }, { "id": 823973, "author_name": "", "author_url": "", "post_date": "2020-04-28T04:22:42.973000", "content": "", "votes": 0, "replies": [ { "id": 826858, "author_name": "", "author_url": "", "post_date": "2020-04-29T22:51:43.620000", "content": "", "votes": 0, "replies": [] }, { "id": 826891, "author_name": "", "author_url": "", "post_date": "2020-04-29T23:29:18.617000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 792237, "author_name": "", "author_url": "", "post_date": "2020-03-31T02:21:51.710000", "content": "", "votes": 0, "replies": [] }, { "id": 791438, "author_name": "", "author_url": "", "post_date": "2020-03-30T12:25:48.600000", "content": "", "votes": 0, "replies": [] }, { "id": 790959, "author_name": "", "author_url": "", "post_date": "2020-03-30T01:46:11.507000", "content": "", "votes": 0, "replies": [] }, { "id": 788654, "author_name": "", "author_url": "", "post_date": "2020-03-27T21:00:20.390000", "content": "", "votes": 0, "replies": [] }, { "id": 784766, "author_name": "", "author_url": "", "post_date": "2020-03-24T13:46:09.753000", "content": "", "votes": 0, "replies": [] }, { "id": 780671, "author_name": "", "author_url": "", "post_date": "2020-03-20T14:02:14.180000", "content": "", "votes": 0, "replies": [] }, { "id": 780440, "author_name": "", "author_url": "", "post_date": "2020-03-20T09:14:46.300000", "content": "", "votes": 0, "replies": [] }, { "id": 776871, "author_name": "", "author_url": "", "post_date": "2020-03-17T17:33:57.170000", "content": "", "votes": 0, "replies": [] }, { "id": 772680, "author_name": "", "author_url": "", "post_date": "2020-03-15T19:01:12.953000", "content": "", "votes": 0, "replies": [] }, { "id": 770519, "author_name": "", "author_url": "", "post_date": "2020-03-13T03:30:26.840000", "content": "", "votes": 0, "replies": [] }, { "id": 768928, "author_name": "", "author_url": "", "post_date": "2020-03-11T11:08:28.543000", "content": "", "votes": 0, "replies": [] }, { "id": 768701, "author_name": "", "author_url": "", "post_date": "2020-03-11T05:19:20.367000", "content": "", "votes": 0, "replies": [] }, { "id": 768482, "author_name": "", "author_url": "", "post_date": "2020-03-10T21:24:54.230000", "content": "", "votes": 0, "replies": [] }, { "id": 768119, "author_name": "", "author_url": "", "post_date": "2020-03-10T13:37:56.950000", "content": "", "votes": 0, "replies": [] }, { "id": 767031, "author_name": "", "author_url": "", "post_date": "2020-03-09T04:09:52.210000", "content": "", "votes": 0, "replies": [] }, { "id": 766490, "author_name": "", "author_url": "", "post_date": "2020-03-08T08:36:11.447000", "content": "", "votes": 0, "replies": [] }, { "id": 766360, "author_name": "", "author_url": "", "post_date": "2020-03-08T03:49:38.140000", "content": "", "votes": 0, "replies": [ { "id": 767505, "author_name": "", "author_url": "", "post_date": "2020-03-09T18:11:06.933000", "content": "", "votes": 1, "replies": [] }, { "id": 772029, "author_name": "", "author_url": "", "post_date": "2020-03-15T00:00:37.370000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 764307, "author_name": "", "author_url": "", "post_date": "2020-03-05T10:20:37.790000", "content": "", "votes": 0, "replies": [] }, { "id": 763911, "author_name": "", "author_url": "", "post_date": "2020-03-05T01:01:06.137000", "content": "", "votes": 0, "replies": [] }, { "id": 763841, "author_name": "", "author_url": "", "post_date": "2020-03-04T22:30:38.743000", "content": "", "votes": 0, "replies": [] }, { "id": 763804, "author_name": "", "author_url": "", "post_date": "2020-03-04T21:59:42.457000", "content": "", "votes": 0, "replies": [] }, { "id": 763701, "author_name": "", "author_url": "", "post_date": "2020-03-04T18:58:05.020000", "content": "", "votes": 0, "replies": [] }, { "id": 763256, "author_name": "", "author_url": "", "post_date": "2020-03-04T09:52:39.797000", "content": "", "votes": 0, "replies": [] }, { "id": 762986, "author_name": "", "author_url": "", "post_date": "2020-03-04T01:57:30.093000", "content": "", "votes": 0, "replies": [] }, { "id": 762922, "author_name": "", "author_url": "", "post_date": "2020-03-03T23:58:25.580000", "content": "", "votes": 0, "replies": [] }, { "id": 762919, "author_name": "", "author_url": "", "post_date": "2020-03-03T23:56:54.580000", "content": "", "votes": 0, "replies": [] }, { "id": 762905, "author_name": "", "author_url": "", "post_date": "2020-03-03T23:39:24.733000", "content": "", "votes": 0, "replies": [] }, { "id": 762880, "author_name": "", "author_url": "", "post_date": "2020-03-03T22:30:13.247000", "content": "", "votes": 0, "replies": [] }, { "id": 762875, "author_name": "", "author_url": "", "post_date": "2020-03-03T22:13:30.223000", "content": "", "votes": 0, "replies": [] }, { "id": 762862, "author_name": "", "author_url": "", "post_date": "2020-03-03T21:59:35.143000", "content": "", "votes": 0, "replies": [] }, { "id": 762853, "author_name": "", "author_url": "", "post_date": "2020-03-03T21:40:19.220000", "content": "", "votes": 0, "replies": [] }, { "id": 762850, "author_name": "", "author_url": "", "post_date": "2020-03-03T21:29:59.770000", "content": "", "votes": 0, "replies": [] }, { "id": 762847, "author_name": "", "author_url": "", "post_date": "2020-03-03T21:17:38.373000", "content": "", "votes": 0, "replies": [] }, { "id": 762841, "author_name": "", "author_url": "", "post_date": "2020-03-03T21:12:12.317000", "content": "", "votes": 0, "replies": [] }, { "id": 762833, "author_name": "", "author_url": "", "post_date": "2020-03-03T20:57:29.867000", "content": "", "votes": 0, "replies": [] }, { "id": 762824, "author_name": "", "author_url": "", "post_date": "2020-03-03T20:46:18.437000", "content": "", "votes": 0, "replies": [] }, { "id": 762811, "author_name": "", "author_url": "", "post_date": "2020-03-03T20:28:17.103000", "content": "", "votes": 0, "replies": [] }, { "id": 762805, "author_name": "", "author_url": "", "post_date": "2020-03-03T20:23:04.003000", "content": "", "votes": 0, "replies": [] }, { "id": 762783, "author_name": "", "author_url": "", "post_date": "2020-03-03T19:55:51.537000", "content": "", "votes": 0, "replies": [] }, { "id": 762780, "author_name": "", "author_url": "", "post_date": "2020-03-03T19:53:58.747000", "content": "", "votes": 0, "replies": [] }, { "id": 762713, "author_name": "", "author_url": "", "post_date": "2020-03-03T18:46:03.850000", "content": "", "votes": 0, "replies": [] }, { "id": 762701, "author_name": "", "author_url": "", "post_date": "2020-03-03T18:25:08.683000", "content": "", "votes": 0, "replies": [] }, { "id": 762697, "author_name": "", "author_url": "", "post_date": "2020-03-03T18:19:00.840000", "content": "", "votes": 0, "replies": [] }, { "id": 762609, "author_name": "", "author_url": "", "post_date": "2020-03-03T16:36:34.983000", "content": "", "votes": 0, "replies": [] }, { "id": 762602, "author_name": "", "author_url": "", "post_date": "2020-03-03T16:26:37.120000", "content": "", "votes": 0, "replies": [] }, { "id": 762589, "author_name": "", "author_url": "", "post_date": "2020-03-03T16:17:17.930000", "content": "", "votes": 0, "replies": [] }, { "id": 762562, "author_name": "", "author_url": "", "post_date": "2020-03-03T16:01:56.320000", "content": "", "votes": 0, "replies": [] }, { "id": 762538, "author_name": "", "author_url": "", "post_date": "2020-03-03T15:38:36.023000", "content": "", "votes": 0, "replies": [] }, { "id": 762510, "author_name": "", "author_url": "", "post_date": "2020-03-03T15:12:41.367000", "content": "", "votes": 0, "replies": [] }, { "id": 762498, "author_name": "", "author_url": "", "post_date": "2020-03-03T15:04:40.873000", "content": "", "votes": 0, "replies": [] }, { "id": 762467, "author_name": "", "author_url": "", "post_date": "2020-03-03T14:34:47.330000", "content": "", "votes": 0, "replies": [] }, { "id": 762458, "author_name": "", "author_url": "", "post_date": "2020-03-03T14:29:18.623000", "content": "", "votes": 0, "replies": [] }, { "id": 762417, "author_name": "", "author_url": "", "post_date": "2020-03-03T13:52:10.317000", "content": "", "votes": 0, "replies": [] }, { "id": 762354, "author_name": "", "author_url": "", "post_date": "2020-03-03T13:09:02", "content": "", "votes": 0, "replies": [] }, { "id": 762353, "author_name": "", "author_url": "", "post_date": "2020-03-03T13:05:47.763000", "content": "", "votes": 0, "replies": [] }, { "id": 762335, "author_name": "", "author_url": "", "post_date": "2020-03-03T12:47:28.973000", "content": "", "votes": 0, "replies": [] }, { "id": 762333, "author_name": "", "author_url": "", "post_date": "2020-03-03T12:46:10.140000", "content": "", "votes": 0, "replies": [] }, { "id": 762313, "author_name": "", "author_url": "", "post_date": "2020-03-03T12:23:17.177000", "content": "", "votes": 0, "replies": [] }, { "id": 762302, "author_name": "", "author_url": "", "post_date": "2020-03-03T12:15:31.640000", "content": "", "votes": 0, "replies": [] }, { "id": 762290, "author_name": "", "author_url": "", "post_date": "2020-03-03T12:02:29.633000", "content": "", "votes": 0, "replies": [] }, { "id": 762282, "author_name": "", "author_url": "", "post_date": "2020-03-03T11:55:28.947000", "content": "", "votes": 0, "replies": [] }, { "id": 762276, "author_name": "", "author_url": "", "post_date": "2020-03-03T11:51:03.487000", "content": "", "votes": 0, "replies": [] }, { "id": 762274, "author_name": "", "author_url": "", "post_date": "2020-03-03T11:49:12.610000", "content": "", "votes": 0, "replies": [] }, { "id": 762251, "author_name": "", "author_url": "", "post_date": "2020-03-03T11:16:37.300000", "content": "", "votes": 0, "replies": [] }, { "id": 762249, "author_name": "", "author_url": "", "post_date": "2020-03-03T11:12:39.897000", "content": "", "votes": 0, "replies": [] }, { "id": 762180, "author_name": "", "author_url": "", "post_date": "2020-03-03T09:36:13.007000", "content": "", "votes": 0, "replies": [] }, { "id": 762178, "author_name": "", "author_url": "", "post_date": "2020-03-03T09:33:58.483000", "content": "", "votes": 0, "replies": [] }, { "id": 762150, "author_name": "", "author_url": "", "post_date": "2020-03-03T08:55:40.197000", "content": "", "votes": 0, "replies": [] }, { "id": 762114, "author_name": "", "author_url": "", "post_date": "2020-03-03T08:08:25.040000", "content": "", "votes": 0, "replies": [] }, { "id": 762094, "author_name": "", "author_url": "", "post_date": "2020-03-03T07:50:40.513000", "content": "", "votes": 0, "replies": [] }, { "id": 762055, "author_name": "", "author_url": "", "post_date": "2020-03-03T06:58:24.527000", "content": "", "votes": 0, "replies": [] }, { "id": 762048, "author_name": "", "author_url": "", "post_date": "2020-03-03T06:45:24.620000", "content": "", "votes": 0, "replies": [] }, { "id": 762038, "author_name": "", "author_url": "", "post_date": "2020-03-03T06:34:02.810000", "content": "", "votes": 0, "replies": [] }, { "id": 762025, "author_name": "", "author_url": "", "post_date": "2020-03-03T06:00:06.837000", "content": "", "votes": 0, "replies": [] }, { "id": 761998, "author_name": "", "author_url": "", "post_date": "2020-03-03T05:29:16.550000", "content": "", "votes": 0, "replies": [] }, { "id": 761995, "author_name": "", "author_url": "", "post_date": "2020-03-03T05:27:11.200000", "content": "", "votes": 0, "replies": [] }, { "id": 761987, "author_name": "", "author_url": "", "post_date": "2020-03-03T05:23:58.090000", "content": "", "votes": 0, "replies": [] }, { "id": 761933, "author_name": "", "author_url": "", "post_date": "2020-03-03T03:41:23.907000", "content": "", "votes": 0, "replies": [] }, { "id": 761891, "author_name": "", "author_url": "", "post_date": "2020-03-03T02:37:51.870000", "content": "", "votes": 0, "replies": [] }, { "id": 761868, "author_name": "", "author_url": "", "post_date": "2020-03-03T02:14:17.253000", "content": "", "votes": 0, "replies": [] }, { "id": 761863, "author_name": "", "author_url": "", "post_date": "2020-03-03T02:00:53.497000", "content": "", "votes": 0, "replies": [] }, { "id": 761806, "author_name": "", "author_url": "", "post_date": "2020-03-03T00:27:06.483000", "content": "", "votes": 0, "replies": [] }, { "id": 761799, "author_name": "", "author_url": "", "post_date": "2020-03-03T00:21:59.920000", "content": "", "votes": 0, "replies": [] }, { "id": 761796, "author_name": "", "author_url": "", "post_date": "2020-03-03T00:19:44.877000", "content": "", "votes": 0, "replies": [] }, { "id": 761784, "author_name": "", "author_url": "", "post_date": "2020-03-02T23:39:11.873000", "content": "", "votes": 0, "replies": [] }, { "id": 761755, "author_name": "", "author_url": "", "post_date": "2020-03-02T22:36:55.567000", "content": "", "votes": 0, "replies": [] }, { "id": 761675, "author_name": "", "author_url": "", "post_date": "2020-03-02T20:49:52.827000", "content": "", "votes": 0, "replies": [] }, { "id": 761655, "author_name": "", "author_url": "", "post_date": "2020-03-02T20:29:08.713000", "content": "", "votes": 0, "replies": [] }, { "id": 761639, "author_name": "", "author_url": "", "post_date": "2020-03-02T19:50:02.973000", "content": "", "votes": 0, "replies": [] }, { "id": 761633, "author_name": "", "author_url": "", "post_date": "2020-03-02T19:37:39.497000", "content": "", "votes": 0, "replies": [] }, { "id": 761627, "author_name": "", "author_url": "", "post_date": "2020-03-02T19:19:12.137000", "content": "", "votes": 0, "replies": [] }, { "id": 761621, "author_name": "", "author_url": "", "post_date": "2020-03-02T19:03:24.473000", "content": "", "votes": 0, "replies": [] }, { "id": 761587, "author_name": "", "author_url": "", "post_date": "2020-03-02T17:46:24.400000", "content": "", "votes": 0, "replies": [] }, { "id": 761547, "author_name": "", "author_url": "", "post_date": "2020-03-02T16:41:16.063000", "content": "", "votes": 0, "replies": [] }, { "id": 761528, "author_name": "", "author_url": "", "post_date": "2020-03-02T16:18:26.130000", "content": "", "votes": 0, "replies": [] }, { "id": 761522, "author_name": "", "author_url": "", "post_date": "2020-03-02T16:12:49.260000", "content": "", "votes": 0, "replies": [] }, { "id": 761501, "author_name": "", "author_url": "", "post_date": "2020-03-02T15:48:07.443000", "content": "", "votes": 0, "replies": [] }, { "id": 761494, "author_name": "", "author_url": "", "post_date": "2020-03-02T15:33:40.333000", "content": "", "votes": 0, "replies": [] }, { "id": 761488, "author_name": "", "author_url": "", "post_date": "2020-03-02T15:24:27.453000", "content": "", "votes": 0, "replies": [] }, { "id": 761479, "author_name": "", "author_url": "", "post_date": "2020-03-02T15:07:59.640000", "content": "", "votes": 0, "replies": [ { "id": 761539, "author_name": "", "author_url": "", "post_date": "2020-03-02T16:33:08.670000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 761446, "author_name": "", "author_url": "", "post_date": "2020-03-02T14:25:13.323000", "content": "", "votes": 0, "replies": [ { "id": 761658, "author_name": "", "author_url": "", "post_date": "2020-03-02T20:33:46.317000", "content": "", "votes": 2, "replies": [] } ] }, { "id": 761381, "author_name": "", "author_url": "", "post_date": "2020-03-02T12:58:12.697000", "content": "", "votes": 0, "replies": [] }, { "id": 761369, "author_name": "", "author_url": "", "post_date": "2020-03-02T12:48:58.250000", "content": "", "votes": 0, "replies": [] }, { "id": 761356, "author_name": "", "author_url": "", "post_date": "2020-03-02T12:28:49.053000", "content": "", "votes": 0, "replies": [] }, { "id": 761319, "author_name": "", "author_url": "", "post_date": "2020-03-02T11:39:45.333000", "content": "", "votes": 0, "replies": [] }, { "id": 761283, "author_name": "", "author_url": "", "post_date": "2020-03-02T11:01:50.500000", "content": "", "votes": 0, "replies": [] }, { "id": 761264, "author_name": "", "author_url": "", "post_date": "2020-03-02T10:27:28.543000", "content": "", "votes": 0, "replies": [] }, { "id": 761254, "author_name": "", "author_url": "", "post_date": "2020-03-02T10:14:46.817000", "content": "", "votes": 0, "replies": [] }, { "id": 761249, "author_name": "", "author_url": "", "post_date": "2020-03-02T10:04:49.760000", "content": "", "votes": 0, "replies": [] }, { "id": 761239, "author_name": "", "author_url": "", "post_date": "2020-03-02T09:56:40.237000", "content": "", "votes": 0, "replies": [] }, { "id": 761228, "author_name": "", "author_url": "", "post_date": "2020-03-02T09:39:42.783000", "content": "", "votes": 0, "replies": [] }, { "id": 761182, "author_name": "", "author_url": "", "post_date": "2020-03-02T08:21:27.970000", "content": "", "votes": 0, "replies": [] }, { "id": 761162, "author_name": "", "author_url": "", "post_date": "2020-03-02T07:49:41.773000", "content": "", "votes": 0, "replies": [] }, { "id": 761119, "author_name": "", "author_url": "", "post_date": "2020-03-02T06:54:58.390000", "content": "", "votes": 0, "replies": [] }, { "id": 761088, "author_name": "", "author_url": "", "post_date": "2020-03-02T05:51:49.510000", "content": "", "votes": 0, "replies": [] }, { "id": 761044, "author_name": "", "author_url": "", "post_date": "2020-03-02T04:21:16.217000", "content": "", "votes": 0, "replies": [] }, { "id": 761034, "author_name": "", "author_url": "", "post_date": "2020-03-02T04:03:11.603000", "content": "", "votes": 0, "replies": [] }, { "id": 761027, "author_name": "", "author_url": "", "post_date": "2020-03-02T03:42:36.157000", "content": "", "votes": 0, "replies": [] }, { "id": 760961, "author_name": "", "author_url": "", "post_date": "2020-03-02T01:06:36.393000", "content": "", "votes": 0, "replies": [ { "id": 762890, "author_name": "", "author_url": "", "post_date": "2020-03-03T22:59:12.873000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 760957, "author_name": "", "author_url": "", "post_date": "2020-03-02T00:59:10.010000", "content": "", "votes": 0, "replies": [] }, { "id": 760852, "author_name": "", "author_url": "", "post_date": "2020-03-01T19:54:23.753000", "content": "", "votes": 0, "replies": [] }, { "id": 760848, "author_name": "", "author_url": "", "post_date": "2020-03-01T19:50:37.957000", "content": "", "votes": 0, "replies": [] }, { "id": 760846, "author_name": "", "author_url": "", "post_date": "2020-03-01T19:47:35.943000", "content": "", "votes": 0, "replies": [] }, { "id": 760715, "author_name": "", "author_url": "", "post_date": "2020-03-01T16:41:28.880000", "content": "", "votes": 0, "replies": [] }, { "id": 760692, "author_name": "", "author_url": "", "post_date": "2020-03-01T16:07:59.940000", "content": "", "votes": 0, "replies": [] }, { "id": 760663, "author_name": "", "author_url": "", "post_date": "2020-03-01T15:33:23.397000", "content": "", "votes": 0, "replies": [] }, { "id": 760636, "author_name": "", "author_url": "", "post_date": "2020-03-01T14:57:46.060000", "content": "", "votes": 0, "replies": [] }, { "id": 760612, "author_name": "", "author_url": "", "post_date": "2020-03-01T14:23:22.847000", "content": "", "votes": 0, "replies": [] }, { "id": 760609, "author_name": "", "author_url": "", "post_date": "2020-03-01T14:21:41.273000", "content": "", "votes": 0, "replies": [] }, { "id": 760552, "author_name": "", "author_url": "", "post_date": "2020-03-01T13:07:49.110000", "content": "", "votes": 0, "replies": [] }, { "id": 760540, "author_name": "", "author_url": "", "post_date": "2020-03-01T12:42:09.517000", "content": "", "votes": 0, "replies": [ { "id": 761666, "author_name": "", "author_url": "", "post_date": "2020-03-02T20:39:08.270000", "content": "", "votes": 0, "replies": [] }, { "id": 761677, "author_name": "", "author_url": "", "post_date": "2020-03-02T20:50:38.167000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 760514, "author_name": "", "author_url": "", "post_date": "2020-03-01T11:58:56.367000", "content": "", "votes": 0, "replies": [] }, { "id": 760509, "author_name": "", "author_url": "", "post_date": "2020-03-01T11:42:33.390000", "content": "", "votes": 0, "replies": [] }, { "id": 760429, "author_name": "", "author_url": "", "post_date": "2020-03-01T09:23:19.487000", "content": "", "votes": 0, "replies": [ { "id": 762190, "author_name": "", "author_url": "", "post_date": "2020-03-03T09:53:02.107000", "content": "", "votes": 0, "replies": [] }, { "id": 776775, "author_name": "", "author_url": "", "post_date": "2020-03-17T16:17:57.487000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 760421, "author_name": "", "author_url": "", "post_date": "2020-03-01T09:12:05.900000", "content": "", "votes": 0, "replies": [] }, { "id": 760396, "author_name": "", "author_url": "", "post_date": "2020-03-01T08:17:48.400000", "content": "", "votes": 0, "replies": [] }, { "id": 760349, "author_name": "", "author_url": "", "post_date": "2020-03-01T06:45:08.953000", "content": "", "votes": 0, "replies": [] }, { "id": 760324, "author_name": "", "author_url": "", "post_date": "2020-03-01T06:05:26.870000", "content": "", "votes": 0, "replies": [] }, { "id": 760167, "author_name": "", "author_url": "", "post_date": "2020-02-29T23:36:46.230000", "content": "", "votes": 0, "replies": [] }, { "id": 760166, "author_name": "", "author_url": "", "post_date": "2020-02-29T23:36:41.510000", "content": "", "votes": 0, "replies": [ { "id": 760414, "author_name": "", "author_url": "", "post_date": "2020-03-01T08:57:45.667000", "content": "", "votes": 2, "replies": [] }, { "id": 760766, "author_name": "", "author_url": "", "post_date": "2020-03-01T17:46:50.913000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 760147, "author_name": "", "author_url": "", "post_date": "2020-02-29T22:40:09.413000", "content": "", "votes": 0, "replies": [] }, { "id": 760125, "author_name": "", "author_url": "", "post_date": "2020-02-29T22:11:51.337000", "content": "", "votes": 0, "replies": [] }, { "id": 759754, "author_name": "", "author_url": "", "post_date": "2020-02-29T12:24:16.960000", "content": "", "votes": 0, "replies": [] }, { "id": 759745, "author_name": "", "author_url": "", "post_date": "2020-02-29T12:13:17.927000", "content": "", "votes": 0, "replies": [] }, { "id": 759719, "author_name": "", "author_url": "", "post_date": "2020-02-29T11:34:30.520000", "content": "", "votes": 0, "replies": [] }, { "id": 759715, "author_name": "", "author_url": "", "post_date": "2020-02-29T11:27:49.223000", "content": "", "votes": 0, "replies": [] }, { "id": 759491, "author_name": "", "author_url": "", "post_date": "2020-02-29T05:25:46.653000", "content": "", "votes": 0, "replies": [] }, { "id": 759455, "author_name": "", "author_url": "", "post_date": "2020-02-29T03:42:58.507000", "content": "", "votes": 0, "replies": [] }, { "id": 759380, "author_name": "", "author_url": "", "post_date": "2020-02-29T00:38:43.653000", "content": "", "votes": 0, "replies": [] }, { "id": 759222, "author_name": "", "author_url": "", "post_date": "2020-02-28T18:41:35.553000", "content": "", "votes": 0, "replies": [] }, { "id": 759221, "author_name": "", "author_url": "", "post_date": "2020-02-28T18:39:51.563000", "content": "", "votes": 0, "replies": [] }, { "id": 759216, "author_name": "", "author_url": "", "post_date": "2020-02-28T18:32:53.193000", "content": "", "votes": 0, "replies": [] }, { "id": 759213, "author_name": "", "author_url": "", "post_date": "2020-02-28T18:23:59.687000", "content": "", "votes": 0, "replies": [] }, { "id": 759172, "author_name": "", "author_url": "", "post_date": "2020-02-28T16:55:09.753000", "content": "", "votes": 0, "replies": [] }, { "id": 759035, "author_name": "", "author_url": "", "post_date": "2020-02-28T13:39:17.433000", "content": "", "votes": 0, "replies": [] }, { "id": 758791, "author_name": "", "author_url": "", "post_date": "2020-02-28T06:52:00.083000", "content": "", "votes": 0, "replies": [] }, { "id": 758668, "author_name": "", "author_url": "", "post_date": "2020-02-28T02:42:23.680000", "content": "", "votes": 0, "replies": [] }, { "id": 758256, "author_name": "", "author_url": "", "post_date": "2020-02-27T15:34:43.540000", "content": "", "votes": 0, "replies": [ { "id": 759420, "author_name": "", "author_url": "", "post_date": "2020-02-29T02:29:31.650000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 757983, "author_name": "", "author_url": "", "post_date": "2020-02-27T10:00:52.410000", "content": "", "votes": 0, "replies": [] }, { "id": 757863, "author_name": "", "author_url": "", "post_date": "2020-02-27T07:32:13.603000", "content": "", "votes": 0, "replies": [] }, { "id": 757194, "author_name": "", "author_url": "", "post_date": "2020-02-26T14:16:57.170000", "content": "", "votes": 0, "replies": [] }, { "id": 756818, "author_name": "", "author_url": "", "post_date": "2020-02-26T05:40:21.330000", "content": "", "votes": 0, "replies": [] }, { "id": 756553, "author_name": "", "author_url": "", "post_date": "2020-02-25T20:39:38.823000", "content": "", "votes": 0, "replies": [] }, { "id": 756518, "author_name": "", "author_url": "", "post_date": "2020-02-25T19:45:48.570000", "content": "", "votes": 0, "replies": [] }, { "id": 756079, "author_name": "", "author_url": "", "post_date": "2020-02-25T11:47:29.447000", "content": "", "votes": 0, "replies": [] }, { "id": 756071, "author_name": "", "author_url": "", "post_date": "2020-02-25T11:40:07.963000", "content": "", "votes": 0, "replies": [] }, { "id": 755919, "author_name": "", "author_url": "", "post_date": "2020-02-25T09:10:23.913000", "content": "", "votes": 0, "replies": [] }, { "id": 755756, "author_name": "", "author_url": "", "post_date": "2020-02-25T05:13:32.400000", "content": "", "votes": 0, "replies": [] }, { "id": 755630, "author_name": "", "author_url": "", "post_date": "2020-02-25T00:58:11.357000", "content": "", "votes": 0, "replies": [] }, { "id": 755565, "author_name": "", "author_url": "", "post_date": "2020-02-24T22:28:25.930000", "content": "", "votes": 0, "replies": [] }, { "id": 755551, "author_name": "", "author_url": "", "post_date": "2020-02-24T21:58:35.217000", "content": "", "votes": 0, "replies": [] }, { "id": 755548, "author_name": "", "author_url": "", "post_date": "2020-02-24T21:54:56.283000", "content": "", "votes": 0, "replies": [] }, { "id": 755439, "author_name": "", "author_url": "", "post_date": "2020-02-24T19:24:20.573000", "content": "", "votes": 0, "replies": [] }, { "id": 755244, "author_name": "", "author_url": "", "post_date": "2020-02-24T15:36:59.573000", "content": "", "votes": 0, "replies": [ { "id": 756459, "author_name": "", "author_url": "", "post_date": "2020-02-25T18:25:47.800000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 754869, "author_name": "", "author_url": "", "post_date": "2020-02-24T06:40:38.323000", "content": "", "votes": 0, "replies": [] }, { "id": 754851, "author_name": "", "author_url": "", "post_date": "2020-02-24T05:56:20.267000", "content": "", "votes": 0, "replies": [] }, { "id": 754743, "author_name": "", "author_url": "", "post_date": "2020-02-24T02:04:37.703000", "content": "", "votes": 0, "replies": [] }, { "id": 754488, "author_name": "", "author_url": "", "post_date": "2020-02-23T16:03:24.810000", "content": "", "votes": 0, "replies": [] }, { "id": 754480, "author_name": "", "author_url": "", "post_date": "2020-02-23T15:50:15.697000", "content": "", "votes": 0, "replies": [] }, { "id": 754477, "author_name": "", "author_url": "", "post_date": "2020-02-23T15:41:36.513000", "content": "", "votes": 0, "replies": [] }, { "id": 754464, "author_name": "", "author_url": "", "post_date": "2020-02-23T15:22:19.917000", "content": "", "votes": 0, "replies": [] }, { "id": 754279, "author_name": "", "author_url": "", "post_date": "2020-02-23T10:41:52.643000", "content": "", "votes": 0, "replies": [ { "id": 754698, "author_name": "", "author_url": "", "post_date": "2020-02-24T00:12:25.873000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 754273, "author_name": "", "author_url": "", "post_date": "2020-02-23T10:25:59.347000", "content": "", "votes": 0, "replies": [] }, { "id": 753980, "author_name": "", "author_url": "", "post_date": "2020-02-22T22:18:47.200000", "content": "", "votes": 0, "replies": [] }, { "id": 753887, "author_name": "", "author_url": "", "post_date": "2020-02-22T19:27:06.580000", "content": "", "votes": 0, "replies": [] }, { "id": 753172, "author_name": "", "author_url": "", "post_date": "2020-02-21T19:51:55.260000", "content": "", "votes": 0, "replies": [] }, { "id": 752948, "author_name": "", "author_url": "", "post_date": "2020-02-21T15:10:05.030000", "content": "", "votes": 0, "replies": [] }, { "id": 752820, "author_name": "", "author_url": "", "post_date": "2020-02-21T12:51:21.533000", "content": "", "votes": 0, "replies": [] }, { "id": 752661, "author_name": "", "author_url": "", "post_date": "2020-02-21T09:21:41.140000", "content": "", "votes": 0, "replies": [] }, { "id": 752609, "author_name": "", "author_url": "", "post_date": "2020-02-21T08:00:18.870000", "content": "", "votes": 0, "replies": [] }, { "id": 752490, "author_name": "", "author_url": "", "post_date": "2020-02-21T04:27:48.747000", "content": "", "votes": 0, "replies": [] }, { "id": 752306, "author_name": "", "author_url": "", "post_date": "2020-02-20T21:26:45.937000", "content": "", "votes": 0, "replies": [] }, { "id": 751682, "author_name": "", "author_url": "", "post_date": "2020-02-20T12:04:10.730000", "content": "", "votes": 0, "replies": [] }, { "id": 751588, "author_name": "", "author_url": "", "post_date": "2020-02-20T10:05:54.800000", "content": "", "votes": 0, "replies": [] }, { "id": 751556, "author_name": "", "author_url": "", "post_date": "2020-02-20T09:20:40.487000", "content": "", "votes": 0, "replies": [] }, { "id": 750470, "author_name": "", "author_url": "", "post_date": "2020-02-19T12:19:23.397000", "content": "", "votes": 0, "replies": [] }, { "id": 749074, "author_name": "", "author_url": "", "post_date": "2020-02-18T09:12:57.320000", "content": "", "votes": 0, "replies": [] }, { "id": 747037, "author_name": "", "author_url": "", "post_date": "2020-02-15T21:51:39.923000", "content": "", "votes": 0, "replies": [ { "id": 751858, "author_name": "", "author_url": "", "post_date": "2020-02-20T15:22:10.433000", "content": "", "votes": 3, "replies": [] }, { "id": 752959, "author_name": "", "author_url": "", "post_date": "2020-02-21T15:21:39.780000", "content": "", "votes": 1, "replies": [] } ] }, { "id": 744776, "author_name": "", "author_url": "", "post_date": "2020-02-13T06:10:35.637000", "content": "", "votes": 0, "replies": [] }, { "id": 744586, "author_name": "", "author_url": "", "post_date": "2020-02-13T00:24:54.367000", "content": "", "votes": 0, "replies": [] }, { "id": 744330, "author_name": "", "author_url": "", "post_date": "2020-02-12T18:40:36.117000", "content": "", "votes": 0, "replies": [] }, { "id": 743643, "author_name": "", "author_url": "", "post_date": "2020-02-12T06:48:27.283000", "content": "", "votes": 0, "replies": [ { "id": 745598, "author_name": "", "author_url": "", "post_date": "2020-02-14T01:21:57.217000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 743361, "author_name": "", "author_url": "", "post_date": "2020-02-12T02:26:00.633000", "content": "", "votes": 0, "replies": [ { "id": 743403, "author_name": "", "author_url": "", "post_date": "2020-02-12T03:21:13.957000", "content": "", "votes": 2, "replies": [] } ] }, { "id": 738129, "author_name": "", "author_url": "", "post_date": "2020-02-06T07:07:35.427000", "content": "", "votes": 0, "replies": [ { "id": 741220, "author_name": "", "author_url": "", "post_date": "2020-02-10T10:30:26.457000", "content": "", "votes": 0, "replies": [] }, { "id": 746506, "author_name": "", "author_url": "", "post_date": "2020-02-15T05:49:07.920000", "content": "", "votes": 0, "replies": [] }, { "id": 746521, "author_name": "", "author_url": "", "post_date": "2020-02-15T06:24:30.840000", "content": "", "votes": 1, "replies": [] }, { "id": 746623, "author_name": "", "author_url": "", "post_date": "2020-02-15T09:28:59.987000", "content": "", "votes": 0, "replies": [] }, { "id": 763407, "author_name": "", "author_url": "", "post_date": "2020-03-04T13:00:15.693000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 737844, "author_name": "", "author_url": "", "post_date": "2020-02-05T20:55:41.117000", "content": "", "votes": 0, "replies": [] }, { "id": 732513, "author_name": "", "author_url": "", "post_date": "2020-01-29T22:48:30.150000", "content": "", "votes": 0, "replies": [ { "id": 732865, "author_name": "", "author_url": "", "post_date": "2020-01-30T11:53:32.763000", "content": "", "votes": 1, "replies": [] }, { "id": 732954, "author_name": "", "author_url": "", "post_date": "2020-01-30T13:41:31.710000", "content": "", "votes": 1, "replies": [] }, { "id": 733044, "author_name": "", "author_url": "", "post_date": "2020-01-30T15:22:35.820000", "content": "", "votes": 0, "replies": [] }, { "id": 738853, "author_name": "", "author_url": "", "post_date": "2020-02-07T05:12:02.583000", "content": "", "votes": 1, "replies": [] }, { "id": 744264, "author_name": "", "author_url": "", "post_date": "2020-02-12T17:30:17.567000", "content": "", "votes": 1, "replies": [] }, { "id": 744289, "author_name": "", "author_url": "", "post_date": "2020-02-12T17:45:16.917000", "content": "", "votes": 7, "replies": [] }, { "id": 747438, "author_name": "", "author_url": "", "post_date": "2020-02-16T12:47:44.287000", "content": "", "votes": 0, "replies": [] }, { "id": 749071, "author_name": "", "author_url": "", "post_date": "2020-02-18T09:08:54.013000", "content": "", "votes": 2, "replies": [] }, { "id": 749095, "author_name": "", "author_url": "", "post_date": "2020-02-18T09:41:57.930000", "content": "", "votes": 5, "replies": [] } ] }, { "id": 731744, "author_name": "", "author_url": "", "post_date": "2020-01-29T01:53:38.003000", "content": "", "votes": 0, "replies": [] }, { "id": 728072, "author_name": "", "author_url": "", "post_date": "2020-01-24T11:24:11.520000", "content": "", "votes": 0, "replies": [] }, { "id": 726941, "author_name": "", "author_url": "", "post_date": "2020-01-23T10:25:05.517000", "content": "", "votes": 0, "replies": [] }, { "id": 718340, "author_name": "", "author_url": "", "post_date": "2020-01-14T10:01:02.863000", "content": "", "votes": 0, "replies": [] }, { "id": 715444, "author_name": "", "author_url": "", "post_date": "2020-01-10T14:45:02.097000", "content": "", "votes": 0, "replies": [] }, { "id": 715442, "author_name": "", "author_url": "", "post_date": "2020-01-10T14:43:08.773000", "content": "", "votes": 0, "replies": [] }, { "id": 712780, "author_name": "", "author_url": "", "post_date": "2020-01-07T15:16:15.257000", "content": "", "votes": 0, "replies": [] }, { "id": 712699, "author_name": "", "author_url": "", "post_date": "2020-01-07T14:13:55.257000", "content": "", "votes": 0, "replies": [] }, { "id": 711150, "author_name": "", "author_url": "", "post_date": "2020-01-05T18:38:28.630000", "content": "", "votes": 0, "replies": [] }, { "id": 707231, "author_name": "", "author_url": "", "post_date": "2019-12-31T12:57:46.800000", "content": "", "votes": 0, "replies": [] }, { "id": 706256, "author_name": "", "author_url": "", "post_date": "2019-12-30T06:32:21.650000", "content": "", "votes": 0, "replies": [] }, { "id": 705633, "author_name": "", "author_url": "", "post_date": "2019-12-29T07:55:01.617000", "content": "", "votes": 0, "replies": [] }, { "id": 701554, "author_name": "", "author_url": "", "post_date": "2019-12-23T15:55:13.707000", "content": "", "votes": 0, "replies": [] }, { "id": 700725, "author_name": "", "author_url": "", "post_date": "2019-12-22T13:57:23.317000", "content": "", "votes": 0, "replies": [] }, { "id": 694595, "author_name": "", "author_url": "", "post_date": "2019-12-13T20:23:31.467000", "content": "", "votes": 0, "replies": [] }, { "id": 693839, "author_name": "", "author_url": "", "post_date": "2019-12-12T20:48:39.793000", "content": "", "votes": 0, "replies": [ { "id": 694466, "author_name": "", "author_url": "", "post_date": "2019-12-13T16:45:35.267000", "content": "", "votes": 1, "replies": [] }, { "id": 707519, "author_name": "", "author_url": "", "post_date": "2020-01-01T01:28:17.237000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 772665, "author_name": "", "author_url": "", "post_date": "2020-03-15T18:34:19.613000", "content": "", "votes": 0, "replies": [] }, { "id": 762857, "author_name": "", "author_url": "", "post_date": "2020-03-03T21:52:09.783000", "content": "", "votes": 0, "replies": [] }, { "id": 762821, "author_name": "", "author_url": "", "post_date": "2020-03-03T20:43:36.393000", "content": "", "votes": 0, "replies": [] }, { "id": 762358, "author_name": "", "author_url": "", "post_date": "2020-03-03T13:12:22.580000", "content": "", "votes": 0, "replies": [] }, { "id": 762148, "author_name": "", "author_url": "", "post_date": "2020-03-03T08:53:31.800000", "content": "", "votes": 0, "replies": [] }, { "id": 761903, "author_name": "", "author_url": "", "post_date": "2020-03-03T02:52:12.577000", "content": "", "votes": -2, "replies": [ { "id": 762423, "author_name": "", "author_url": "", "post_date": "2020-03-03T13:54:54.043000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 761861, "author_name": "", "author_url": "", "post_date": "2020-03-03T01:54:07.090000", "content": "", "votes": 0, "replies": [] }, { "id": 761558, "author_name": "", "author_url": "", "post_date": "2020-03-02T16:55:12.550000", "content": "", "votes": 0, "replies": [] }, { "id": 761511, "author_name": "", "author_url": "", "post_date": "2020-03-02T15:56:48.663000", "content": "", "votes": 0, "replies": [] }, { "id": 761299, "author_name": "", "author_url": "", "post_date": "2020-03-02T11:17:26.747000", "content": "", "votes": 0, "replies": [] }, { "id": 760173, "author_name": "", "author_url": "", "post_date": "2020-02-29T23:48:12.127000", "content": "", "votes": 0, "replies": [] }, { "id": 759727, "author_name": "", "author_url": "", "post_date": "2020-02-29T11:44:49.160000", "content": "", "votes": 0, "replies": [] }, { "id": 755942, "author_name": "", "author_url": "", "post_date": "2020-02-25T09:41:03.900000", "content": "", "votes": 0, "replies": [] }, { "id": 747766, "author_name": "", "author_url": "", "post_date": "2020-02-16T20:20:02.913000", "content": "", "votes": 0, "replies": [] }, { "id": 746490, "author_name": "", "author_url": "", "post_date": "2020-02-15T04:59:09.903000", "content": "", "votes": 0, "replies": [] }, { "id": 742576, "author_name": "", "author_url": "", "post_date": "2020-02-11T11:42:59.427000", "content": "", "votes": 0, "replies": [] }, { "id": 728722, "author_name": "", "author_url": "", "post_date": "2020-01-25T06:32:00.533000", "content": "", "votes": 0, "replies": [] }, { "id": 726419, "author_name": "", "author_url": "", "post_date": "2020-01-23T02:14:18.013000", "content": "", "votes": 0, "replies": [] }, { "id": 725676, "author_name": "", "author_url": "", "post_date": "2020-01-22T11:21:39.723000", "content": "", "votes": 0, "replies": [] }, { "id": 721000, "author_name": "", "author_url": "", "post_date": "2020-01-16T23:32:53.960000", "content": "", "votes": 0, "replies": [] }, { "id": 719344, "author_name": "", "author_url": "", "post_date": "2020-01-15T12:02:42.243000", "content": "", "votes": 0, "replies": [] }, { "id": 716584, "author_name": "", "author_url": "", "post_date": "2020-01-11T23:08:48.210000", "content": "", "votes": 0, "replies": [] }, { "id": 694027, "author_name": "", "author_url": "", "post_date": "2019-12-13T04:48:12.473000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "758820": "@addisonhoward @juliaelliott\n\nAs external data might be very helpful to obtain better scores on public/private leaderboards how are you going to validate solutions' compliance with the rules?\nThere is a possibility to cheat by using non allowed external data (or not posted here before merge deadline). I'm strongly against that but we are not in a perfect world and cheaters will always exist. \nIdeally winning solutions must be reproduced end-to-end: retrain, test - check that scores are quite similar to public/private leaderboards. This will ensure that winners used only allowed datasets.", "692881": "Post links to your external data sources here before the deadline specified in the rules. Once it has been posted, you do not need to post it again.", "716551": "There are several DeepFake datasets avaialble:\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/\n\nAll of them avaiable by request. Can we use them for developing solution or not? I want to give it a try, but don't want to waste time on them if they are forbidden.", "694492": "For clarity on the datasets that need to be declared here: You do not need to declare your own self-trained/original work models. But if you are using a dataset or pre-trained model obtained externally which is licensed with the right for you to use, then that dataset needs to be declared. \n\nSo, for example: if you are using the imagenet dataset and efficientnet to train your model(s), you must declare imagenet and efficientnet on this thread, but you don’t need to make your trained model public. \n\nYou also do not need to declare things that have already been declared.", "760710": "Some people have surfaced this already but let me ask again: it would be nice to have the final list of allowed external resources under the post. It probably requires some effort to create this list and check it but will be very useful to everyone since the deadline is approaching. \nWhat do you think @addisonhoward? ", "706023": "@tunguz' datasets for real faces and fake faces.\n\nReal Faces:\nhttps://www.kaggle.com/tunguz/70000-real-faces-1\n \nFake Faces:\nhttps://www.kaggle.com/tunguz/1-million-fake-faces\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-2\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-3\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-4\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-5\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-6\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-7", "763235": "By the way, a list of permitted external data and allowed pre-trained models (using approved licenses) should be maintained as a \"Kaggle-level\" topic because we have the same questions repeated across competitions.", "701432": "```\nUADFV: The UADFV dataset - Xin Yang, Yuezun Li, and Siwei Lyu. Exposing deep fakes\nusing inconsistent head poses. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),\n2019.\n\nDF-TIMIT: Pavel Korshunov and Sebastien Marcel. Deepfakes: a new ´\nthreat to face recognition? assessment and detection. arXiv\npreprint arXiv:1812.08685, 2018.\n\nFF-DF: Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Chris- ¨\ntian Riess, Justus Thies, and Matthias Nießner. FaceForensics++: Learning to detect manipulated facial images. In\nICCV, 2019.\n\nDFD: Nicholas Dufour, Andrew Gully, Per Karlsson, Alexey Victor Vorbyov, Thomas Leung, Jeremiah Childs, and Christoph Bregler. Deepfakes detection dataset by google & jigsaw.\n```", "694433": "Static build of FFMPEG: https://johnvansickle.com/ffmpeg/", "748358": "* Pretrained weights from https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet \n* Faces from https://github.com/NVlabs/ffhq-dataset with open licenses (https://creativecommons.org/publicdomain/mark/1.0/, https://creativecommons.org/publicdomain/zero/1.0/, https://creativecommons.org/licenses/by/2.0/)\n* Video Tools to install packages https://www.kaggle.com/harangdev/video-tools\n* faceswap tool https://github.com/deepfakes/faceswap and models https://github.com/deepfakes-models/faceswap-models/releases", "767591": "@juliaelliott In the rules, it says:\n\n> No using other Kaggle notebooks or utility scripts as inputs to your submission notebook. External data also cannot be externally-referenced (i.e. via BigQuery, Github, external URL). Instead, load external models or datasets directly as an external data source.\n\nI upload my own scripts as a Kaggle dataset so I can import functions from them and not have an excessively long notebook. Is that allowed? ", "723253": "InceptionV3 imagenet weights from Keras:\nhttps://keras.io/applications/#inceptionv3\n\nUFC101 Khurram Soomro, Amir Roshan Zamir and Mubarak Shah, UCF101: A Dataset of 101 Human Action Classes From Videos in The Wild., CRCV-TR-12-01, November, 2012.\n\nhttp://crcv.ucf.edu/data/UCF101/", "716317": "facenet-pytorch pre-trained models: https://github.com/timesler/facenet-pytorch\n(MIT license)", "693664": "Open Images Dataset: https://storage.googleapis.com/openimages/web/index.html\nObjects 365 Dataset: https://www.objects365.org/overview.html\nCOCO Dataset: http://cocodataset.org/#home", "692907": "“External data is allowed up to 1 GB in size. External data must be freely & publicly available, including pre-trained models”\n\nDoes it count during training time? Or just for predict time?", "762894": "https://www.youtube.com/ Creative Commons videos\nhttps://drive.google.com/file/d/1NeyTFAwdJSjxA9ZtdviwdUjdptEVjM_i\nhttps://www.kaggle.com/tunguz/1-million-fake-faces\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-2\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-3\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-4\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-5\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-6\nhttps://www.kaggle.com/tunguz/1-million-fake-faces-7\nhttps://github.com/IDRnD/LCC_FASD", "762923": "FACENET, VGG16 https://github.com/davidsandberg/facenet\nMTCNN\nhttps://github.com/ipazc/mtcnn", "762654": "https://github.com/huawei-noah/ghostnet\nhttps://github.com/iamhankai/ghostnet.pytorch\nhttps://github.com/kuan-wang/pytorch-mobilenet-v3", "762339": "https://www.thispersondoesnotexist.com/", "762337": "All media with permitting CC license from:\nhttps://badoo.com/\nhttps://www.instagram.com/\nhttps://www.youtube.com/\nhttps://www.flickr.com/\nhttps://www.google.com/imghp\nhttps://yandex.ru/images/\nFrom this thread: https://www.kaggle.com/c/sp-society-camera-model-identification/discussion/47235\n\nIf any source violates the rules, I will not use it. List it here just in case :)", "762257": "pytorch vgg16 model\nwheight https://download.pytorch.org/models/vgg16-397923af.pth\npytorch vgg19 model\nvgg19: https://download.pytorch.org/models/vgg19-dcbb9e9d.pth\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\npretrained model: https://drive.google.com/drive/folders/1zvl89AgFAApbH0At-gMuZSeQB_LpNP-M\nhttps://github.com/timesler/facenet-pytorch\nmodel 20180408-102900: https://drive.google.com/uc?export=download&id=12DYdlLesBl3Kk51EtJsyPS8qA7fErWDX\nmodel 20180402-114759: https://drive.google.com/uc?export=download&id=1TDZVEBudGaEd5POR5X4ZsMvdsh1h68T1\nhttps://github.com/mit-han-lab/temporal-shift-module\nTSN ResNet50: https://hanlab.mit.edu/projects/tsm/models/TSM_kinetics_RGB_resnet50_avg_segment5_e50.pth\nTSM ResNet50: https://hanlab.mit.edu/projects/tsm/models/TSM_kinetics_RGB_resnet50_shift8_blockres_avg_segment8_e100_dense.pth\nhttps://github.com/ufoym/imbalanced-dataset-sampler\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch/releases/\nhttps://github.com/NVlabs/ffhq-dataset\nhttps://github.com/deepinsight/insightface/\nhttps://github.com/deepinsight/insightface/#pretrained-models\nhttps://github.com/Sierkinhane/mtcnn-pytorch\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nweight: https://drive.google.com/file/d/1WeXlNYsM6dMP3xQQELI-4gxhwKUQxc3-/view?usp=sharing\nhttps://github.com/lijiannuist/lightDSFD\nweight: https://github.com/lijiannuist/lightDSFD/raw/master/weights/light_DSFD.pth\nhttps://github.com/cydonia999/VGGFace2-pytorch\nhttps://github.com/cydonia999/VGGFace2-pytorch#pretrained-models\nhttps://github.com/albumentations-team/albumentations", "762097": "@juliaelliott \nTo ensure Fairness and Transparency, upon Competition ended and preliminary Private leaderboard score and place has been calculated, will you ask the top 10 Kernels to fully disclose the External dataset they used (together with their kernel code) for Public scrutiny BEFORE the confirmation of any award? This will allowed the Kaggle community to help organizer in catching any form of cheating or overlook.", "762082": "https://github.com/hukkelas/DSFD-Pytorch-Inference\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/sfzhang15/FaceBoxes\nhttps://github.com/TropComplique/FaceBoxes-tensorflow\nhttps://github.com/XiaXuehai/faceboxes\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/peteryuX/retinaface-tf2\nhttps://github.com/OFRIN/Tensorflow_RetinaFace\nhttps://github.com/yeephycho/tensorflow-face-detection\nhttps://v.qq.com/channel/choice\nwww.youtube.com\nwww.youku.com\nwww.tudou.com\nwww.tiktok.com\nwww.douyin.com\nwww.tonton.com.my\nhttp://tv.cctv.com/\nwww.bbc.com\nwww.cnn.com\nwww.foxnews.com\nwww.iqiyi.com\nhttps://github.com/deepmind/kinetics-i3d\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/yaohungt/Gated-Spatio-Temporal-Energy-Graph\nhttps://github.com/dlpbc/keras-kinetics-i3d\nhttps://github.com/huangyangyu/SeqFace\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/ronghuaiyang/arcface-pytorch\nhttps://github.com/happynear/AMSoftmax\nhttps://github.com/foamliu/InsightFace-v2\nhttps://github.com/1996scarlet/ArcFace-Multiplex-Recognition\nhttps://github.com/xiaoboCASIA/SV-X-Softmax\nhttps://github.com/foamliu/InsightFace-PyTorch\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/iloveuu2011/Face-Detection-Library\nhttps://github.com/iloveuu2011/retinaface-tf2\nhttps://github.com/pvskand/DisguiseNet\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/EricZgw/PyramidBox\nhttps://github.com/swghosh/DeepFace\nhttps://github.com/zma-c-137/VarGFaceNet\nhttps://github.com/cvtower/seesawfacenet_pytorch\nhttps://github.com/kk7nc/RMDL\nhttps://github.com/ZhaoJ9014/High-Performance-Face-Recognition\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/rlaengud123/CMC_LRCN\nhttps://github.com/doronharitan/human_activity_recognition_LRCN\nhttps://github.com/piergiaj/representation-flow-cvpr19\nhttps://github.com/piergiaj/evanet-iccv19\nhttps://github.com/piergiaj\nhttps://github.com/piergiaj/mlb-youtube\nhttps://github.com/craston/MARS\nhttps://github.com/clancylian/retinaface\nhttps://github.com/yangfly\nhttps://pan.baidu.com/share/init?surl=P1ypO7VYUbNAezdvLm2m9w\nhttps://github.com/biubug6/Face-Detector-1MB-with-landmark\nhttps://github.com/610265158/DSFD-tensorflow", "761474": "- A python package to analyze and compare voices with deep learning\nhttps://github.com/resemble-ai/Resemblyzer\n\n- Real-Time Voice Cloning\nhttps://github.com/CorentinJ/Real-Time-Voice-Cloning\n\n- LibriSpeech ASR corpus\nhttp://www.openslr.org/12/\n\n- LibriTTS corpus\nhttp://www.openslr.org/60/\n\n- The VoxCeleb Dataset\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1.html\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox2.html\n\n- The M-AILABS Speech Dataset\nhttps://www.caito.de/2019/01/the-m-ailabs-speech-dataset/\n\n- MrDeepfakes\nhttps://mrdeepfakes.com/terms\n\n- English Multi-speaker Corpus for CSTR Voice Cloning Toolkit\nhttps://homepages.inf.ed.ac.uk/jyamagis/page3/page58/page58.html\nhttps://datashare.is.ed.ac.uk/handle/10283/2651\n\n- FaceForensics original repository\nhttps://github.com/ondyari/FaceForensics/tree/original\n\n- IEEE's Signal Processing Society - Camera Model Identification\nhttps://www.kaggle.com/c/sp-society-camera-model-identification", "761035": "BlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\n\n\n", "760744": "dlib: https://github.com/davisking/dlib\ndlib-models: https://github.com/davisking/dlib-models/blob/master/shape_predictor_68_face_landmarks.dat.bz2\nnumpy: https://github.com/numpy/numpy\nopencv: https://github.com/opencv/opencv\nscipy: https://github.com/scipy/scipy\nimutils: https://pypi.org/project/imutils/\nripser: https://github.com/scikit-tda/ripser.py\nglob: https://docs.python.org/3/library/glob.html\nos: https://docs.python.org/3/library/os.html\nmultiprocessing: https://docs.python.org/3/library/multiprocessing.html\nscikit-learn: https://scikit-learn.org/stable/index.html\nScikit-TDA: https://github.com/scikit-tda\ncsv: https://docs.python.org/3/library/csv.html", "758759": "opencv extra modules\nhttps://github.com/opencv/opencv_contrib\n\npytables\nhttps://pypi.org/project/tables/\n\nh5py\nhttps://pypi.org/project/h5py/\n\npandas\nhttps://pypi.org/project/pandas/\n\nalbumetations\nhttps://pypi.org/project/albumentations/\n\nTacotron2\nhttps://github.com/NVIDIA/tacotron2\n\nWaveglow\nhttps://github.com/NVIDIA/waveglow\n\nParallelWaveGAN\nhttps://github.com/kan-bayashi/ParallelWaveGAN\n\nMozilla TTS\nhttps://github.com/mozilla/TTS\n\nCatalyst framework\nhttps://github.com/catalyst-team/catalyst\n\nPytorch-toolbelt & Examples\nhttps://github.com/BloodAxe/pytorch-toolbelt\nhttps://github.com/BloodAxe/Kaggle-2019-Blindness-Detection\n\nKornia package\nhttps://github.com/kornia/kornia", "758018": "https://github.com/yxlijun/S3FD.pytorch", "757847": "Pretrained models disclosure from two repositories:\nhttps://pytorch.org/docs/stable/torchvision/models.html\nand\nhttps://modelzoo.co/framework/pytorch", "757253": "https://keras.io/applications\nhttps://www.tensorflow.org/lite/models\nhttps://github.com/NVIDIA/tensorflow-determinism\nhttps://github.com/tensorflow/models/blob/master/official/README.md\nhttps://github.com/tensorflow/models/blob/master/official/README-TPU.md", "756800": "External Data Disclosure\n\nNVlabs ffhq dataset \nhttps://drive.google.com/open?id=1u2xu7bSrWxrbUxk-dT-UvEJq8IjdmNTP\n\nUADFV dataset \nhttps://drive.google.com/drive/folders/1GEk1DSxmlV_61JtpEGzC9Fo_BffvyxpH?usp=sharing\n\nDataset from ASVSpoof 2019 competition \nhttps://datashare.is.ed.ac.uk/handle/10283/3336\n\nKaggle CIPLAB\nhttps://www.kaggle.com/ciplab/real-and-fake-face-detection/download\n\nOpenCV face detector model\n\nDLIB face detector model\n\nVoxCeleb\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/\n\nresnet and xception for learning transfer (I'm testing with both tensorflow and pytorch) using imagenet as baseline.\n\nhttps://github.com/thiago1080/SphereFace\nhttps://github.com/vlad3996/FaceDetection-DSFD\n https://github.com/EndlessSora/DeeperForensics-1.0\n", "756063": "DLIB: https://github.com/davisking/dlib\nfacenet: https://github.com/timesler/facenet-pytorch\nPre-trained models of VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nVGGFace2: https://drive.google.com/uc?export=download&id=1TDZVEBudGaEd5POR5X4ZsMvdsh1h68T1\nface-recognition: https://github.com/ageitgey/face_recognition\nface-recognition-models: https://github.com/ageitgey/face_recognition_models\nautokeras: https://github.com/keras-team/autokeras\nkeras-tuner: https://github.com/keras-team/keras-tuner\nterminaltables: https://github.com/Robpol86/terminaltables", "755549": "keras xception imagenet weights https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xception_weights_tf_dim_ordering_tf_kernels_notop.h5", "753252": "youtube-bb dataset\nhttps://research.google.com/youtube-bb/", "750901": "vggface2 dataset : http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nand also\nhttps://www.kaggle.com/c/imagenet-object-localization-challenge/data\n\nJust a question, are those licenses ok with dataset limitation : \n- Creative Commons Attribution-ShareAlike ?\n- and Creative Commons Attribution-NonCommercial-ShareAlike ?\n", "750278": "http://www.robots.ox.ac.uk/~vgg/data/voxceleb/", "749115": "Pretrained models from:\nhttps://github.com/rwightman/pytorch-image-models\nhttps://github.com/rwightman/gen-efficientnet-pytorch", "748620": "https://github.com/qubvel/segmentation_models.pytorch", "740961": "Keras applications models\nhttps://keras.io/applications/\n", "737834": "https://github.com/NVlabs/ffhq-dataset\nhttps://www.kaggle.com/jessicali9530/celeba-dataset\nhttps://susanqq.github.io/UTKFace/ ", "736429": "Using this set of video tools uploaded by sheldon robinson on kaggle. Didn't find a comment here, so I hope it helps someone else too. Dataset [here](https://www.kaggle.com/sheldonrobinson/video-tools). Tutorial for installation [here](https://www.kaggle.com/sheldonrobinson/starter-video-tools/notebook). Tools include scikit-video, ffmpeg, moviepy etc. I will primarily be using scikit-video and moviepy.", "730492": "Keras ResNet50 (MIT license) and Resnet50V2 (BSD license) models and pre-trained weights from https://keras.io/applications/#resnet", "707331": "[opencvdnnfp16](https://www.kaggle.com/mbmk92/opencvdnnfp16) - Dataset containing opencv face detection Caffe model - courtesy Milandu Keith\n[timesler/facenet-pytorch-vggface2](https://www.kaggle.com/timesler/facenet-pytorch-vggface2) Facenet with pretrained weights - courtesy timesler\nResNet with pre-trained weights", "705166": "Pretrained models on VggFace2: https://github.com/cydonia999/VGGFace2-pytorch#pretrained-models", "704584": "BlazeFace for face and landmark detection: https://sites.google.com/view/perception-cv4arvr/blazeface\n\nEdit: I added this as a [data source](https://www.kaggle.com/humananalog/blazeface-pytorch) and a [demo kernel](https://www.kaggle.com/humananalog/starter-blazeface-pytorch).", "701639": "RetinaFace Pytorch detector: https://github.com/biubug6/Pytorch_Retinaface", "695988": "opencv dnn face detector: \nhttps://raw.githubusercontent.com/opencv/opencv_3rdparty/dnn_samples_face_detector_20180220_uint8/opencv_face_detector_uint8.pb\nhttps://github.com/opencv/opencv/blob/master/samples/dnn/face_detector/opencv_face_detector.pbtxt", "695274": "MTCNN-package\n[https://pypi.org/project/mtcnn/](https://pypi.org/project/mtcnn/)", "762816": "models and weights :\nhttps://github.com/hukkelas/DSFD-Pytorch-Inference\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/sfzhang15/FaceBoxes\nhttps://github.com/TropComplique/FaceBoxes-tensorflow\nhttps://github.com/XiaXuehai/faceboxes\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/peteryuX/retinaface-tf2\nhttps://github.com/OFRIN/Tensorflow_RetinaFace\nhttps://github.com/yeephycho/tensorflow-face-detection\nhttps://github.com/deepmind/kinetics-i3d\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/yaohungt/Gated-Spatio-Temporal-Energy-Graph\nhttps://github.com/dlpbc/keras-kinetics-i3d\nhttps://github.com/huangyangyu/SeqFace\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/ronghuaiyang/arcface-pytorch\nhttps://github.com/happynear/AMSoftmax\nhttps://github.com/foamliu/InsightFace-v2\nhttps://github.com/1996scarlet/ArcFace-Multiplex-Recognition\nhttps://github.com/xiaoboCASIA/SV-X-Softmax\nhttps://github.com/foamliu/InsightFace-PyTorch\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/iloveuu2011/Face-Detection-Library\nhttps://github.com/iloveuu2011/retinaface-tf2\nhttps://github.com/pvskand/DisguiseNet\nhttps://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/EricZgw/PyramidBox\nhttps://github.com/swghosh/DeepFace\nhttps://github.com/HRNet/HRNet-Facial-Landmark-Detection\nhttps://github.com/zma-c-137/VarGFaceNet\nhttps://github.com/cvtower/seesawfacenet_pytorch\nhttps://github.com/kk7nc/RMDL\nhttps://github.com/ZhaoJ9014/High-Performance-Face-Recognition\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/rlaengud123/CMC_LRCN\nhttps://github.com/doronharitan/human_activity_recognition_LRCN\nhttps://github.com/piergiaj/representation-flow-cvpr19\nhttps://github.com/piergiaj/evanet-iccv19\nhttps://github.com/piergiaj\nhttps://github.com/piergiaj/mlb-youtube\nhttps://github.com/craston/MARS\nRetinaFace : https://github.com/deepinsight/insightface/tree/master/RetinaFace\nKeras Xception Imagenet Weights https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xceptionweightstfdimorderingtfkernels_notop.h5\nEfficient Net Weights https://github.com/qubvel/efficientnet\nImgaug: https://imgaug.readthedocs.io/en/latest/\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/albumentations-team/albumentations\nMesoNet - https://github.com/DariusAf/MesoNet\nopencv extra modules\nhttps://github.com/opencv/opencv_contrib\npytables\nhttps://pypi.org/project/tables/\nh5py\nhttps://pypi.org/project/h5py/\npandas\nhttps://pypi.org/project/pandas/\nalbumetations\nhttps://pypi.org/project/albumentations/\nTacotron2\nhttps://github.com/NVIDIA/tacotron2\nWaveglow\nhttps://github.com/NVIDIA/waveglow\nParallelWaveGAN\nhttps://github.com/kan-bayashi/ParallelWaveGAN\nMozilla TTS\nhttps://github.com/mozilla/TTS\nCatalyst framework\nhttps://github.com/catalyst-team/catalyst\nPytorch-toolbelt & Examples\nhttps://github.com/BloodAxe/pytorch-toolbelt\nhttps://github.com/BloodAxe/Kaggle-2019-Blindness-Detection\nKornia package\nhttps://github.com/kornia/kornia\nNetworks and weights in github\neffecientnet ( https://github.com/lukemelas/EfficientNet-PyTorch ) - Google weights\n'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b0-355c32eb.pth',\n'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b1-f1951068.pth',\n'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b2-8bb594d6.pth',\n'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b3-5fb5a3c3.pth',\n'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b4-6ed6700e.pth',\n'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b5-b6417697.pth',\n'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b6-c76e70fd.pth',\n'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b7-dcc49843.pth'\n\n'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b0-b64d5a18.pth',\n'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b1-0f3ce85a.pth',\n'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b2-6e9d97e5.pth',\n'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b3-cdd7c0f4.pth',\n'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b4-44fb3a87.pth',\n'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b5-86493f6b.pth',\n'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b6-ac80338e.pth',\n'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b7-4652b6dd.pth',\n'efficientnet-b8': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b8-22a8fe65.pth'\n\nModels and weights:\nEfficientNet Keras (and TensorFlow Keras)\nhttps://github.com/qubvel/efficientnet\nClassification models Zoo - Keras (and TensorFlow Keras)\nhttps://github.com/qubvel/classification_models\nPython library with Neural Networks for Image Segmentation based on Keras and TensorFlow. \nhttps://github.com/qubvel/segmentation_models\n\nmtccn\npytorch model zoo\nopencv2\nscikit learn\nfastai\nEfficientNet-pyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nresnext101: https://github.com/facebookresearch/WSL-Images\nface.evoLVe.PyTorch: https://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nResidualAttentionNetwork-pytorch: https://github.com/tengshaofeng/ResidualAttentionNetwork-pytorch\nhrnet: https://github.com/HRNet/HRNet-Image-Classification\ntorchvision models: https://github.com/pytorch/vision/tree/master/torchvision/models\nFaceForensics++ dataset: https://github.com/ondyari/FaceForensics\ntrainmsra.tar.gz traincelebrity.tar.gz: http://trillionpairs.deepglint.com/data\n\nwider face datasets: http://shuoyang1213.me/WIDERFACE/\nresnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n'resnext5032x4d': 'https://download.pytorch.org/models/resnext5032x4d-7cdf4587.pth',\n'resnext10132x8d': 'https://download.pytorch.org/models/resnext10132x8d-8ba56ff5.pth',\n'wideresnet502': 'https://download.pytorch.org/models/wideresnet502-95faca4d.pth',\n'wideresnet1012': 'https://download.pytorch.org/models/wideresnet1012-32ee1156.pth',\nhttps://github.com/qubvel/segmentation_models.pytorch\nPretrained weights from https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nKeras applications models\nhttps://keras.io/applications/\nInceptionV3 imagenet weights from Keras:\nhttps://keras.io/applications/#inceptionv3\nfacenet-pytorch pre-trained models: https://github.com/timesler/facenet-pytorch\n(MIT license)\nPre-trained MTCNN model from here: https://github.com/timesler/facenet-pytorch\nimutils package: https://github.com/jrosebr1/imutils\nRetinaFace Pytorch detector: https://github.com/biubug6/Pytorch_Retinaface\nopencv dnn face detector:\nhttps://raw.githubusercontent.com/opencv/opencv_3rdparty/dnn_samples_face_detector_20180220_uint8/opencv_face_detector_uint8.pb\nhttps://github.com/opencv/opencv/blob/master/samples/dnn/face_detector/opencv_face_detector.pbtxt\nhttps://github.com/nyoki-mtl/keras-facenet\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/huawei-noah/ghostnet\nhttps://github.com/iamhankai/ghostnet.pytorch\nhttps://github.com/kuan-wang/pytorch-mobilenet-v3\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nhttps://keras.io/applications/\nDatasets:\nImageNet\nhttps://github.com/Daniil-Osokin/lightweight-human-pose-estimation.pytorch\nhttps://github.com/HRNet\nhttps://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py\n", "761637": "http://www.openslr.org/17\nhttp://www.openslr.org/26\n\nhttp://kaldi-asr.org/models/m4\n\nhttps://github.com/hitachi-speech/EEND\nhttps://github.com/WeidiXie/VGG-Speaker-Recognition\n\nhttps://github.com/mozilla/DeepSpeech\n\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb\n\nhttps://github.com/Fengdalu/LipNet-PyTorch\nhttps://github.com/astorfi/lip-reading-deeplearning\nhttp://www.robots.ox.ac.uk/~vgg/research/deep_lip_reading", "761391": "https://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttps://github.com/rwightman/pytorch-image-models\nhttps://pypi.org/project/gluoncv2/\nhttps://github.com/ondyari/FaceForensics/tree/master/classification\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/mnikitin/EfficientNet\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://github.com/narumiruna/efficientnet-pytorch\nhttps://github.com/zsef123/EfficientNets-PyTorch\nhttps://mxnet.apache.org/api/python/docs/api/gluon/model_zoo/index.html\nhttp://www.cs.albany.edu/~lsw/celeb-deepfakeforensics.html\nhttps://github.com/ox-vgg/vgg_face2\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/grib0ed0v/face_recognition.pytorch\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nhttps://github.com/deepinsight/insightface\nhttp://shuoyang1213.me/WIDERFACE/\nhttps://github.com/lijiannuist/lightDSFD\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/yxlijun/S3FD.pytorch\nhttps://github.com/supernotman/RetinaFace_Pytorch\nhttps://github.com/bogireddytejareddy/retinaface-pytorch\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/clovaai/EXTD_Pytorch\nhttps://github.com/ElvishElvis/68-Retinaface-Pytorch-version\nhttps://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/sfzhang15/SFD\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/hcl14/retinaface-pytorch-inference\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\n", "760263": "RetinaFace : [https://github.com/deepinsight/insightface/tree/master/RetinaFace](https://github.com/deepinsight/insightface/tree/master/RetinaFace)\n\nKeras Xception Imagenet Weights [https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xception_weights_tf_dim_ordering_tf_kernels_notop.h5](https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xception_weights_tf_dim_ordering_tf_kernels_notop.h5)\n\nEfficient Net Weights [https://github.com/qubvel/efficientnet](https://github.com/qubvel/efficientnet)\n\nImgaug: [https://imgaug.readthedocs.io/en/latest/](https://imgaug.readthedocs.io/en/latest/)\n\n\n", "759185": "MesoNet - https://github.com/DariusAf/MesoNet", "758640": "MTCNN - package", "742139": "https://www.kaggle.com/caffeinism/helpers\n\nsome changes in https://www.kaggle.com/humananalog/deepfakes-inference-demo's helpers code", "762386": "https://www.kaggle.com/c/deepfake-detection-challenge/discussion/123665\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/121173\nhttps://github.com/ondyari/FaceForensics \nhttp://kaldir.vc.in.tum.de/FaceForensics/models/faceforensics++_models.zip\nhttps://youtube.com CC videos\n\nhttps://github.com/DariusAf/MesoNet\nhttps://github.com/titu1994/keras-efficientnets \nhttps://keras.io/applications/\nhttps://www.tensorflow.org/api_docs/python/tf/keras/applications/\nhttps://github.com/qubvel/efficientnet\n\nhttps://github.com/KaiyangZhou/deep-person-reid\nhttps://github.com/rwightman/pytorch-image-models\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/pytorch/vision/tree/master/torchvision/models", "760274": "https://github.com/ipazc/mtcnn\n\nhttps://github.com/albumentations-team/albumentations", "732874": "EfficientNet implementation in PyTorch plus pretrained weights found here https://github.com/lukemelas/EfficientNet-PyTorch (Apache 2.0 license)", "723979": "Dual Shot Face Detection - Pytorch Inference Code from https://github.com/hukkelas/DSFD-Pytorch-Inference (Apache 2.0 license)\nWiderFACE pretrained model from https://github.com/TencentYoutuResearch/FaceDetection-DSFD (Apache 2.0 license)\n\n@inproceedings{li2018dsfd,\n title={DSFD: Dual Shot Face Detector},\n author={Li, Jian and Wang, Yabiao and Wang, Changan and Tai, Ying and Qian, Jianjun and Yang, Jian and Wang, Chengjie and Li, Jilin and Huang, Feiyue},\n booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},\n year={2019}\n}", "723311": "not sure if this helps but there is the YouTube 8m challenge videos that are free to use and already on Kaggle.\nhttps://research.google.com/youtube8m/index.html\n\ncurl data.yt8m.org/download.py | partition=3/frame/validate mirror=us python\ncurl data.yt8m.org/download.py | partition=3/frame/test mirror=us python", "722928": "May we use code and models from here: https://github.com/deepinsight/insightface/ ?\nTheir license states:\n> The code of InsightFace is released under the MIT License. There is no limitation for both acadmic and commercial usage.\n>\n> The training data containing the annotation (and the models trained with these data) are available for non-commercial research purposes only.\n\nI'm not entirely sure if this challenge counts as non-commerical research.", "719539": "Pre-trained MTCNN model from here: https://github.com/timesler/facenet-pytorch\nimutils package: https://github.com/jrosebr1/imutils", "705785": "MTCNN-package\nhttps://pypi.org/project/mtcnn/", "693452": "“External data is allowed up to 1 GB in size. External data must be freely & publicly available, including pre-trained models”\n\nSince we must upload our locally trained/pre-trained models as a data source, and all data sources must be made publicly available, won't this make it so that everyone can simply always take the best current model and have everyone score mostly the same? \n\nMaybe I am missing something.", "769124": "A python package to analyze and compare voices with deep learning\nhttps://github.com/resemble-ai/Resemblyzer\nhttps://www.kaggle.com/masterhyj/packagefile\nhttps://www.kaggle.com/sheldonrobinson/video-tools", "759242": "https://pypi.org/, in case I didn't post it. :)", "703168": "youtube.com has a nice dataset of videos", "761667": "All videos from https://youtube.com avialable at this moment.", "756837": "**https://github.com/moabitcoin/ig65m-pytorch**\nNetworks and weights in github\n\n\n**effecientnet ( https://github.com/lukemelas/EfficientNet-PyTorch ) - Google weights**\n'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b0-355c32eb.pth',\n 'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b1-f1951068.pth',\n 'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b2-8bb594d6.pth',\n 'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b3-5fb5a3c3.pth',\n 'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b4-6ed6700e.pth',\n 'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b5-b6417697.pth',\n 'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b6-c76e70fd.pth',\n 'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b7-dcc49843.pth'\n\n 'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b0-b64d5a18.pth',\n 'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b1-0f3ce85a.pth',\n 'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b2-6e9d97e5.pth',\n 'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b3-cdd7c0f4.pth',\n 'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b4-44fb3a87.pth',\n 'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b5-86493f6b.pth',\n 'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b6-ac80338e.pth',\n 'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b7-4652b6dd.pth',\n 'efficientnet-b8': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b8-22a8fe65.pth'\n\n**mtccn**\n**pytroch model zoo**\n**opencv2**\n**scikit learn**\n**fastai**\n\nEfficientNet-pyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nresnext101: https://github.com/facebookresearch/WSL-Images\nface.evoLVe.PyTorch: https://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nResidualAttentionNetwork-pytorch: https://github.com/tengshaofeng/ResidualAttentionNetwork-pytorch\nhrnet: https://github.com/HRNet/HRNet-Image-Classification\ntorchvision models: https://github.com/pytorch/vision/tree/master/torchvision/models\nFaceForensics++ dataset: https://github.com/ondyari/FaceForensics\ntrainmsra.tar.gz traincelebrity.tar.gz: http://trillionpairs.deepglint.com/data\n\n\nwider face datasets: http://shuoyang1213.me/WIDERFACE/\nresnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n'resnext5032x4d': 'https://download.pytorch.org/models/resnext5032x4d-7cdf4587.pth',\n'resnext10132x8d': 'https://download.pytorch.org/models/resnext10132x8d-8ba56ff5.pth',\n'wideresnet502': 'https://download.pytorch.org/models/wideresnet502-95faca4d.pth',\n'wideresnet1012': 'https://download.pytorch.org/models/wideresnet1012-32ee1156.pth',\n\n\n\n\nhttps://www.kaggle.com/tunguz/70000-real-faces-1\nhttps://www.kaggle.com/tunguz/70000-real-faces-2\nhttps://www.kaggle.com/tunguz/70000-real-faces-3\nhttps://www.kaggle.com/tunguz/70000-real-faces-4\nhttps://www.kaggle.com/tunguz/70000-real-faces-5\nhttps://www.kaggle.com/tunguz/70000-real-faces-6\n\n\nmmdetection: https://github.com/open-mmlab/mmdetection\n\n\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\n\n\nhttps://github.com/EndlessSora/DeeperForensics-1.0\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/\n\n\nUADFV: The UADFV dataset - Xin Yang, Yuezun Li, and Siwei Lyu. Exposing deep fakes\nusing inconsistent head poses. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),\n2019.\n\nDF-TIMIT: Pavel Korshunov and Sebastien Marcel. Deepfakes: a new ´\nthreat to face recognition? assessment and detection. arXiv\npreprint arXiv:1812.08685, 2018.\n\nFF-DF: Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Chris- ¨\ntian Riess, Justus Thies, and Matthias Nießner. FaceForensics++: Learning to detect manipulated facial images. In\nICCV, 2019.\n\nDFD: Nicholas Dufour, Andrew Gully, Per Karlsson, Alexey Victor Vorbyov, Thomas Leung, Jeremiah Childs, and Christoph Bregler. Deepfakes detection dataset by google & jigsaw.\n\nhttps://github.com/dessa-research/DeepFake-Detection", "823973": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/facebookresearch/SlowFast\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/elliottzheng/face-detection", "792237": "https://www.kaggle.com/humananalog/deepfakes-inference-demo/download/aioBap8W5sTGGJr16faR%2Fversions%2FvTJB32h13oA5mti8DfTI%2Ffiles%2Fresnext.pth?datasetVersionNumber=1", "791438": "https://github.com/rwightman/pytorch-image-models\nhttps://github.com/danmohaha/celeb-deepfakeforensics\nhttps://github.com/ondyari/FaceForensics\nhttp://mmlab.ie.cuhk.edu.hk/projects/CelebA.html", "790959": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/pytorch/vision/tree/master/torchvision/models\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/hollance/BlazeFace-PyTorch\nhttps://github.com/albumentations-team/albumentations\nhttps://github.com/nerox8664/pytorch2keras\nhttps://github.com/danmohaha/celeb-deepfakeforensics", "788654": "http://parnec.nuaa.edu.cn/xtan/data/ClosedEyeDatabases.html\ndata set of closed eyes", "784766": "https://github.com/cleardusk/MeGlass", "780671": "", "780440": "mtcnn\nhttps://www.kaggle.com/unkownhihi/mtcnn-package\nhttps://github.com/ipazc/mtcnn", "776871": "http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttp://megaface.cs.washington.edu", "772680": "face_recognition https://github.com/ageitgey/face_recognition\nvgg-19 http://www.vlfeat.org/matconvnet/models/beta16/imagenet-vgg-verydeep-19.mat\nhttps://github.com/shekkizh/FCN.tensorflow\n", "770519": "I found a landmark pretrained model today, and I rewrite the postprocess code. Does this case also need to be declared?\nhttps://github.com/justusschock/shapenet", "768928": "https://github.com/tensorflow/models/\nhttps://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md\nhttps://github.com/yeephycho/tensorflow-face-detection\nhttps://github.com/EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10", "768701": "Using Tensorflow 2 latest", "768482": "https://github.com/kpzhang93/MTCNN_face_detection_alignment\n\nhttps://github.com/davisking/dlib\nhttp://dlib.net/files/mmod_human_face_detector.dat.bz2\nhttps://www.kaggle.com/pedromoya/dlibpackage\n\nhttps://github.com/weiliu89/caffe/tree/ssd\nhttps://github.com/opencv/opencv/tree/master/samples/dnn/face_detector\n\nhttps://github.com/nhatthai/opencv-face-recognition/blob/master/src/face_detection_model/res10_300x300_ssd_iter_140000.caffemodel\n\nhttp://dlib.net/files/data/dlib_face_detector_training_data.tar.gz\nhttp://dlib.net/files/data/dlib_faces_5points.tar\nhttp://dlib.net/files/data/dlib_face_detector_training_data.tar.gz\n\nhttps://github.com/blaueck/tf-mtcnn\nhttps://www.kaggle.com/pedromoya/mtcmtflib\n\nfaceRecognition\nhttps://github.com/ydwen/caffe-face\n\nhearth-rate\nhttps://www.kaggle.com/pedromoya/hearthrate", "768119": "imagenet pretrained models resnet,xception ,resnext .. not sure if i will use anything in future ..", "767031": "Test dataset from https://github.com/PeterWang512/CNNDetection\n", "766490": "https://github.com/open-mmlab/mmaction", "766360": "Where it says that we have to disclose pre-trained models, does it mean all pre-trained models, including our original model that we trained outside of Kaggle and then uploaded to Kaggle (as essentially all models are, since you can't really train your models on Kaggle) or is it only referring to open source models that others have built (and possibly pre-trained, so we can use transfer-learning), and we are burrowing to incorporate in our own model?\n\nThanks in advance.", "764307": "All of freely available debian deb packages and shared libs.", "763911": "https://github.com/yxlijun/S3FD.pytorch\nhttps://www.kaggle.com/pedromoya/s3fdpytorch\nhttps://www.kaggle.com/pedromoya/mtcnnbib\nhttps://www.kaggle.com/pedromoya/mtcnnmtcnn", "763841": "https://github.com/pytorch/vision/blob/master/torchvision/models\npytorch Resnext\n 'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n 'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n 'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n 'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n 'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n 'resnext50_32x4d': 'https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pth',\n 'resnext101_32x8d': 'https://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pth',\n 'wide_resnet50_2': 'https://download.pytorch.org/models/wide_resnet50_2-95faca4d.pth',\n 'wide_resnet101_2': 'https://download.pytorch.org/models/wide_resnet101_2-32ee1156.pth'\npytorch inception\n'inception_v3_google': 'https://download.pytorch.org/models/inception_v3_google-1a9a5a14.pth'", "763804": "https://github.com/shijianjian/EfficientNet-PyTorch-3D", "763701": "https://www.kaggle.com/humananalog/blazeface-pytorch\nhttps://github.com/timesler/facenet-pytorch\n", "763256": "Some video classification project\nhttps://github.com/HHTseng/video-classification\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/boyaolin/Video-Classification", "762986": "https://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://pypi.org/project/mtcnn/\nhttps://pypi.org/project/python_speech_features/0.4/\nhttps://pypi.org/project/pytorchcv/\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/ageitgey/face_recognition_models\nwww.youtube.com with permitted license\n", "762922": "Resemblyzer: https://github.com/resemble-ai/Resemblyzer\nmmdetection: https://github.com/open-mmlab/mmdetection\nDual Shot Face Detector\nhttps://keithito.com/LJ-Speech-Dataset/\nhttps://github.com/thiago1080/SphereFace\nVoice Datasets: https://github.com/jim-schwoebel/voice_datasets (Only ones that comply with the rules, dataset in this list with any restrictions will not be used).\nCommonVoice:https://www.kaggle.com/mozillaorg/common-voice/home\nhttps://github.com/nyoki-mtl/keras-facenet\nhttps://github.com/onnx/onnx\nTFFD: https://github.com/yeephycho/tensorflow-face-detection code not training data used (which is not allowed)\nHelen: http://www.ifp.illinois.edu/~vuongle2/helen/\nSFEW: http://cs.anu.edu.au/few\nFacenet: https://github.com/davidsandberg/facenet\nhttps://www.kaggle.com/selfishgene/youtube-faces-with-facial-keypoints\nFace Detection in Images: https://www.kaggle.com/dataturks/face-detection-in-images\nYoutube Faces Dataset: https://www.cs.tau.ac.il/~wolf/ytfaces/\nhttps://research.google.com/audioset/download.html\nFace Detection Data Set & Benchmark http://vis-www.cs.umass.edu/fddb/\nFaces in the wild dataset: http://tamaraberg.com/faceDataset/index.html\nFlickr Faces HQ Dataset: https://github.com/NVlabs/ffhq-dataset\nTufts Face Database: https://www.kaggle.com/kpvisionlab/tufts-face-database\nReal and Fake Face Detection: https://www.kaggle.com/ciplab/real-and-fake-face-detection\nGoogle Facial Expression: https://research.google/tools/datasets/google-facial-expression/\nhttps://www.kaggle.com/drgilermo/face-images-with-marked-landmark-points\nhttps://github.com/thiago1080/SphereFace\nhttps://ai.googleblog.com/2017/08/launching-speech-commands-dataset.html\nLibrispeech: https://www.openslr.org/12\nhttps://homepages.inf.ed.ac.uk/jyamagis/page3/page58/page58.html\nhttps://github.com/ageitgey/face_recognition\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nhttps://github.com/iitzco/faced\nhttps://github.com/qubvel/efficientnet\nhttps://github.com/Star-Clouds/CenterFace\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/NVlabs/ffhq-dataset\nhttps://github.com/matterport/Mask_RCNN\nhttps://github.com/deepfakes/faceswap\nhttps://www.youtube.com/ Creative Commons\nEverything a :https://keras.io/applications/ and\nhttps://github.com/keras-team/keras-applications and any of their pre-trained models and weights that comply with licensing rules of competition.\nhttps://github.com/ageitgey/face_recognition\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/fyr91/facedetection https://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices/tree/master/facedetection\nSome architectures and pretrained models and datasets:\nhttp://www.robots.ox.ac.uk/~vgg/data/\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttp://www.robots.ox.ac.uk/~vgg/data/voxceleb/\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/ondyari/FaceForensics\nhttp://www.robots.ox.ac.uk/~vgg/data/lipreading/ https://github.com/astorfi/lip-reading-deeplearning https://github.com/joseph-zhong/LipReading https://github.com/afourast/deeplip_reading\nhttps://github.com/hassanhub/LipReading\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/qubvel/efficientnet\nhttps://github.com/qubvel/classification_models\nhttps://github.com/1adrianb/face-alignment\nhttps://www.kaggle.com/humananalog/deepfakes-inference-demo\n \nDisclaimer:\nDatasets I post may be used in the final submission. They will be thoroughly vetted before final submission and any ones that do not comply with rules will not be used in the final submission, subject to your further review during verification.\nThank you.", "762919": "pip install ffmpeg-python\nconda install -c conda-forge ffmpeg\npip install librosa\nTensorflow 1.*", "762905": " 1. https://github.com/EndlessSora/DeeperForensics-1.0\n 2. https://github.com/TencentYoutuResearch/FaceDetection-DSFD\n 3. https://github.com/iitzco/faced\n 4. https://github.com/matterport/Mask_RCNN\n 5. https://github.com/timesler/facenet-pytorch\n 6. https://github.com/yeephycho/tensorflow-face-detection\n 7. https://github.com/qubvel/efficientnet\n 8. https://github.com/ipazc/mtcnn\n 9. https://github.com/ondyari/FaceForensics\n 10. https://github.com/Star-Clouds/CenterFace\n 11. http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1.html\n 12. http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox2.html\n 13. https://github.com/NVlabs/ffhq-dataset\n 14. https://github.com/resemble-ai/Resemblyzer\n 15. https://github.com/thiago1080/SphereFace\n 16. https://github.com/HRNet/HRNet-Image-Classification\n 17. https://github.com/deepfakes/faceswap", "762880": "https://github.com/timesler/facenet-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/resemble-ai/Resemblyzer\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/opencv/opencv_contrib\nhttps://pypi.org/project/tables/\nhttps://pypi.org/project/h5py/\nhttps://pypi.org/project/pandas/\nhttps://github.com/ondyari/FaceForensics/\nmtccn\npytorch model zoo\nopencv2\nscikit learn\nfastai\nskvideo\nffmpeg-python\nhttps://www.kaggle.com/rakibilly/ffmpeg-static-build\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/mnikitin/EfficientNet\nhttps://pypi.org/project/librosa/\nhttps://www.kaggle.com/ciplab/real-and-fake-face-detection", "762875": "https://github.com/ondyari/FaceForensics\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/hollance/BlazeFace-PyTorch\nhttps://github.com/thiago1080/SphereFace\nhttps://github.com/vlad3996/FaceDetection-DSFD\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://johnvansickle.com/ffmpeg/\nhttps://www.kaggle.com/rakibilly/ffmpeg-static-build\nhttps://github.com/astorfi/speechpy\n\nPretrained models from:\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttps://github.com/fchollet/deep-learning-models\n", "762862": "https://github.com/ondyari/FaceForensics\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/hollance/BlazeFace-PyTorch\n", "762853": "https://github.com/opencv/opencv/blob/master/data/haarcascades/haarcascade_frontalface_default.xml", "762850": "https://github.com/sampepose/flownet2-tf", "762847": "https://datashare.is.ed.ac.uk/handle/10283/3336", "762841": "https://pytorch.org/hub/research-models\nhttps://github.com/pytorch/examples\nhttps://github.com/taki0112/SPADE-Tensorflow\nhttps://github.com/ajbrock/BigGAN-PyTorch\nhttps://github.com/rosinality/style-based-gan-pytorch\nhttps://github.com/junyanz/pytorch-CycleGAN-and-pix2pix\nhttps://github.com/yunjey/stargan\nhttps://github.com/WojciechMormul/crn\nhttps://github.com/johnathanlouie/crn\nhttps://github.com/zth667/Diverse-Image-Synthesis-from-Semantic-Layout\nhttps://github.com/KupynOrest/DeblurGAN\nhttps://github.com/TAMU-VITA/DeblurGANv2\nhttp://cchen156.web.engr.illinois.edu/SID.html\nhttps://github.com/daitao/SAN\nhttps://modelzoo.co/\n", "762833": "MesoNet\nhttps://github.com/DariusAf/MesoNet\nFaceForensics code, no dataset\nhttps://github.com/ondyari/FaceForensics\nMulti-task\nhttps://github.com/nii-yamagishilab/ClassNSeg\nCapsule\nhttps://github.com/nii-yamagishilab/Capsule-Forensics-v2\npretrainedmodels\nhttps://github.com/cadene/pretrained-models.pytorch\nhttps://data.lip6.fr/cadene/pretrainedmodels/\ntfhub models \nhttps://tfhub.dev/\ndlib\nhttps://github.com/davisking/dlib\nface_recognition\nhttps://github.com/ageitgey/face_recognition\nmtcnn\nhttps://github.com/ipazc/mtcnn\nfacenet-pytorch\nhttps://github.com/timesler/facenet-pytorch\nBlazeFace-PyTorch\nhttps://github.com/hollance/BlazeFace-PyTorch\nffmpeg\nhttps://johnvansickle.com/ffmpeg/\nffmpeg-python\nhttps://pypi.org/project/ffmpeg-python/#files", "762824": "Fakes generated by, and Datasets and pretrained models (where license permits) used by, the following:\n https://github.com/deepfakes/faceswap\n https://github.com/joshua-wu/deepfakes_faceswap\n https://github.com/wuhuikai/FaceSwap\n https://github.com/jinfagang/faceswap_pytorch\n https://github.com/iperov/DeepFaceLab\n https://github.com/shaoanlu/faceswap-GAN\n https://github.com/goberoi/faceit\n https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix\n https://github.com/LynnHo/AttGAN-Tensorflow\n https://github.com/LynnHo/DCGAN-LSGAN-WGAN-GP-DRAGAN-Tensorflow-2\n https://github.com/gsurma/face_generator\n https://github.com/snknitin/DeepfakeCapsuleGAN\n https://github.com/tkarras/progressive_growing_of_gans\n https://github.com/pfnet-research/sngan_projection\n https://github.com/mbinkowski/MMD-GAN\n https://github.com/dfaker/df\n https://github.com/NVlabs/stylegan2\n https://github.com/NVlabs/stylegan\n https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix\n https://github.com/NVIDIA/unsupervised-video-interpolation\n https://github.com/NVIDIA/pix2pixHD\n https://github.com/NVIDIA/flownet2-pytorch\n https://github.com/NVIDIA/OpenSeq2Seq\n https://github.com/andabi/deep-voice-conversion\n https://github.com/keithito/tacotron\n https://github.com/NVIDIA/tacotron2\n https://github.com/MycroftAI/mimic2\n https://github.com/Sharad24/Neural-Voice-Cloning-with-Few-Samples\n https://github.com/r9y9/deepvoice3_pytorch\n https://github.com/CorentinJ/Real-Time-Voice-Cloning\nhttps://keithito.com/LJ-Speech-Dataset/\nUNet3D etc at https://github.com/qubvel/tpu/tree/master/models\nhttps://github.com/Res2Net/Res2Net-PretrainedModels\nhttps://github.com/facebookresearch/semi-supervised-ImageNet1K-models\nModels at https://pytorch.org/hub/\nhttps://github.com/matterport/Mask_RCNN\nSegmentation model Unet with different backbone models (resnet, vgg16 etc)\nFastai model zoo and pretrained weights https://docs.fast.ai/vision.models.html\nhttps://github.com/divamgupta/image-segmentation-keras\nModels at https://github.com/NVIDIA/semantic-segmentation\n such as DeepLabV3+ architecture with different backbones, including WideResNet38, SEResNeXt(50, 101) and ResNet(50,101)\nhttps://github.com/gasvn/Res2Net\nhttps://github.com/ansleliu/LightNet\nhttps://github.com/adobe/antialiased-cnns\nhttps://github.com/Media-Smart/vedaseg\nhttps://github.com/tyiannak/pyAudioAnalysis\nhttps://github.com/amsehili/auditok\nhttps://github.com/ina-foss/inaSpeechSegmenter\nhttps://github.com/albietz/online_hmm\nhttps://github.com/qubvel/ttach", "762811": "https://github.com/1adrianb/face-alignment", "762805": "https://github.com/YuvalNirkin/face_segmentation.git\nhttps://github.com/rwightman/posenet-python\nhttps://github.com/XifengGuo/CapsNet-Keras\nhttps://github.com/tensorflow/tfjs-models/tree/master/body-pix\nhttps://github.com/shamangary/FSA-Net\nhttps://github.com/1adrianb/face-alignment\nhttps://github.com/ipazc/mtcnn\nhttps://github.com/deepinsight/insightface\nhttps://www.npmjs.com/package/@tensorflow-models/blazeface\nhttps://github.com/scikit-image/scikit-image\nhttps://github.com/librosa/librosa\nhttps://github.com/makcedward/nlpaug\nhttps://github.com/tyiannak/pyAudioAnalysis\nhttps://datashare.is.ed.ac.uk/handle/10283/3336\nhttps://github.com/jefflai108/ASSERT\nhttps://github.com/resemble-ai/Resemblyzer", "762783": "**Models**\n\nDarknet (https://github.com/AlexeyAB/darknet#yolo-v3-in-other-frameworks)\nyolov3-tiny-prn :\nhttps://drive.google.com/open?id=1_NnfVgj0EDtb_WLNoXV8Mo7WKgwdYZCc\nyolov3-tiny-prn:\nhttps://raw.githubusercontent.com/AlexeyAB/darknet/master/cfg/csresnext50-panet-spp.cfg\n\nPytorch model zoo\nresnet18: https://download.pytorch.org/models/resnet18-5c106cde.pth\nresnet34: https://download.pytorch.org/models/resnet34-333f7ec4.pth\nresnet50: https://download.pytorch.org/models/resnet50-19c8e357.pth\nresnet101: https://download.pytorch.org/models/resnet101-5d3b4d8f.pth\nresnet152: https://download.pytorch.org/models/resnet152-b121ed2d.pth\nresnext5032x4d:\nhttps://download.pytorch.org/models/resnext5032x4d-7cdf4587.pth\nresnext10132x8d:\nhttps://download.pytorch.org/models/resnext10132x8d-8ba56ff5.pth\nwideresnet502:\n'https://download.pytorch.org/models/wideresnet502-95faca4d.pth\nwideresnet1012:\n'https://download.pytorch.org/models/wideresnet1012-32ee1156.pth\n\n\n**Packages**\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nFastai2: https://github.com/fastai/fastai2\nOpenCV 4.2.0 (https://github.com/opencv/opencv) including all contrib\nmodules (https://github.com/opencv/opencv_contrib)\n", "762780": "Pretrained weights and code:\nhttps://github.com/sthanhng/yoloface\nhttps://github.com/pjreddie/darknet\nhttps://pjreddie.com/media/files/yolov3.weights\nhttps://pjreddie.com/media/files/yolov3-tiny.weights\nhttps://sites.google.com/view/perception-cv4arvr/blazeface\n\nDatasets:\nhttps://pjreddie.com/projects/pascal-voc-dataset-mirror/\nhttp://cocodataset.org", "762713": "https://github.com/nyoki-mtl/keras-facenet\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/davidsandberg/facenet\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/danmohaha/celeb-deepfakeforensics", "762701": "pyopencl\ntimm\nSSIM-PIL", "762697": "Some people mentioned s3fd related resources. However, it's not clear to me if the underlying datasets and libraries are the same, so I posting what exactly we might use:\n'2DFAN-4': 'https://www.adrianbulat.com/downloads/python-fan/2DFAN4-11f355bf06.pth.tar',\n'3DFAN-4': 'https://www.adrianbulat.com/downloads/python-fan/3DFAN4-7835d9f11d.pth.tar',\n 'depth': 'https://www.adrianbulat.com/downloads/python-fan/depth-2a464da4ea.pth.tar',\n's3fd': 'https://www.adrianbulat.com/downloads/python-fan/s3fd-619a316812.pth", "762609": "pytorch, skimage, sklearn, libsvm, opencv\nhttps://github.com/cvondrick/soundnet\nhttps://github.com/keunhong/pytorch-soundnet\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/bukalapak/pybrisque\nhttps://pypi.org/project/pydub/\nhttps://github.com/timesler/facenet-pytorch", "762602": "These are some of the resources that has helped me/is helping me to understand deepFakes and deepFake detection properly!\n\nhttps://arxiv.org/pdf/1909.11573.pdf\nhttps://arxiv.org/pdf/1812.08685.pdf\nhttps://publications.idiap.ch/downloads/papers/2019/Korshunov_ICB_2019.pdf\nhttps://arxiv.org/pdf/2001.00179.pdf\n\n", "762589": "libfacedetection", "762562": "https://github.com/sthanhng/yoloface\nhttps://github.com/eriklindernoren/PyTorch-YOLOv3", "762538": "I don't know if this is allowed. If not, we will not use it.\n\nhttps://www.kaggle.com/greatgamedota/ffhq-face-data-set", "762510": "ImageNet pre-trained models in pytorchcv: https://pypi.org/project/pytorchcv/\nImageNet pre-trained models in PyTorch Hub: https://pytorch.org/hub/\npython_speech_features: https://pypi.org/project/python_speech_features/0.4/\n", "762498": "https://github.com/iitzco/faced\nhttps://github.com/tensorflow/tfjs-models/tree/master/posenet\nhttps://github.com/matterport/Mask_RCNN", "762467": "BlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch", "762458": "facenet-pytorch: https://github.com/timesler/facenet-pytorch\nhttps://keras.io/applications/\nhttps://github.com/qubvel/efficientnet\n\nDatasets:\nImageNet", "762417": "- FastAi and pre-trained models (depends on torchvision): https://github.com/fastai/fastai and https://github.com/fastai/fastai2\n- mtcnn: https://pypi.org/project/mtcnn/\n- https://github.com/daniilidis-group/neural_renderer \n- https://github.com/fbcotter/pytorch_wavelets \n- https://imgaug.readthedocs.io/en/latest/\n- Dlib (https://github.com/davisking/dlib) with pretrained models from http://dlib.net/files/ \n- https://github.com/hollance/BlazeFace-PyTorch\n- OpenCV and OpenCV_contrib with pretrained models downloaded via https://github.com/opencv/opencv/tree/master/samples/dnn and data https://github.com/opencv/opencv/tree/master/data \n- facenet-pytorch: https://github.com/timesler/facenet-pytorch\n- skimage (https://scikit-image.org/), sklearn (https://scikit-learn.org/stable/), filterpy (https://filterpy.readthedocs.io/en/latest/), scipy (https://scipy.org/), Pandas\n- https://github.com/scikit-video/scikit-video \n- https://github.com/google/mediapipe \n- https://github.com/rwightman/pytorch-image-models \n- https://github.com/grib0ed0v/face_recognition.pytorch \n- https://github.com/lukemelas/EfficientNet-PyTorch \n- https://github.com/rwightman/pytorch-image-models \n- https://github.com/lessw2020/Ranger-Deep-Learning-Optimizer\n- https://github.com/TencentYoutuResearch/FaceDetection-DSFD \n- https://github.com/ZhaoJ9014/face.evoLVe.PyTorch \n- https://github.com/marvis/pytorch-caffe\n- https://github.com/NVIDIA/flownet2-pytorch\n- https://github.com/RanhaoKang/PWC-Net_pytorch \n- https://www.kaggle.com/hmendonca/proper-clustering-with-facenet-embeddings-eda/data\n\nDatasets:\n- ImageNet \n- CIFAR https://www.cs.toronto.edu/~kriz/cifar.html \n- http://live.ece.utexas.edu/research/incaptureDatabase/index.html \n- http://shuoyang1213.me/WIDERFACE/ \n- http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/ \n- https://www.umdfaces.io/ \n- http://sintel.is.tue.mpg.de/\n\nApparently, mostly all resources are restricted due to poor licenses (as I understood). So, I’ll write down a list nice to use…\n- https://github.com/lidq92/VSFA\n- http://tamaraberg.com/faceDataset/index.html \n- http://www.helsinki.fi/psychology/groups/visualcognition/ \n- http://vision.eng.shizuoka.ac.jp/course/view.php?id=6\n- https://www.cs.tau.ac.il/~wolf/ytfaces/ \n- http://vis-www.cs.umass.edu/lfw/ \n- https://github.com/1adrianb/face-alignment \n- https://github.com/ondyari/FaceForensics/tree/original \n- https://github.com/ondyari/FaceForensics/tree/master/classification \n- https://github.com/deepinsight/insightface \n", "762354": "DeeperForensics-1.0 also : https://github.com/EndlessSora/DeeperForensics-1.0\n\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nRetinaFace : https://github.com/deepinsight/insightface/tree/master/RetinaFace\n\nKeras Xception Imagenet Weights https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xceptionweightstfdimorderingtfkernels_notop.h5\n\nEfficient Net Weights https://github.com/qubvel/efficientnet\n\nImgaug: https://imgaug.readthedocs.io/en/latest/\nhttps://github.com/ipazc/mtcnn\n\nhttps://github.com/albumentations-team/albumentations\n\nhttps://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\nhttps://github.com/rwightman/pytorch-image-models\nhttps://pypi.org/project/gluoncv2/\nhttps://github.com/ondyari/FaceForensics/tree/master/classification\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/mnikitin/EfficientNet\nhttps://github.com/rwightman/gen-efficientnet-pytorch\nhttps://github.com/narumiruna/efficientnet-pytorch\nhttps://github.com/zsef123/EfficientNets-PyTorch\nhttps://mxnet.apache.org/api/python/docs/api/gluon/model_zoo/index.html\nhttp://www.cs.albany.edu/~lsw/celeb-deepfakeforensics.html\nhttps://github.com/ox-vgg/vgg_face2\nhttp://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/TreB1eN/InsightFace_Pytorch\nhttps://github.com/grib0ed0v/face_recognition.pytorch\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nhttps://github.com/deepinsight/insightface\nhttp://shuoyang1213.me/WIDERFACE/\nhttps://github.com/lijiannuist/lightDSFD\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\nhttps://github.com/ChiCheng123/SRN\nhttps://github.com/yxlijun/S3FD.pytorch\nhttps://github.com/supernotman/RetinaFace_Pytorch\nhttps://github.com/bogireddytejareddy/retinaface-pytorch\nhttps://github.com/zisianw/FaceBoxes.PyTorch\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/clovaai/EXTD_Pytorch\nhttps://github.com/ElvishElvis/68-Retinaface-Pytorch-version\nhttps://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices\nhttps://github.com/bairdzhang/smallhardface\nhttps://github.com/sfzhang15/SFD\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/hcl14/retinaface-pytorch-inference\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/TencentYoutuResearch/FaceDetection-DSFD\n\nPretrained weights from https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nFaces from https://github.com/NVlabs/ffhq-dataset with open licenses (https://creativecommons.org/publicdomain/mark/1.0/, https://creativecommons.org/publicdomain/zero/1.0/, https://creativecommons.org/licenses/by/2.0/)\nVideo Tools to install packages https://www.kaggle.com/harangdev/video-tools\nfaceswap tool https://github.com/deepfakes/faceswap and models https://github.com/deepfakes-models/faceswap-models/releases\n\n", "762353": "classy vision: https://github.com/facebookresearch/ClassyVision", "762335": "Pytorch and Dlib models: [http://dlib.net/files/](http://dlib.net/files/), [https://pytorch.org/docs/stable/torchvision/models.html](https://pytorch.org/docs/stable/torchvision/models.html)", "762333": "https://github.com/nesl/asvspoof2019 (MIT)", "762313": "MTCNN package :- https://github.com/ipazc/mtcnn\nDSFD from tencent\nPytorch pre-trained models", "762302": "https://www.deepfaker.xyz/", "762290": "https://github.com/qubvel/efficientnet\nhttps://github.com/Callidior/keras-applications/releases/tag/efficientnet", "762282": "https://pypi.org/project/ffmpeg-python/\nhttps://johnvansickle.com/ffmpeg/", "762276": "https://pypi.org/project/mtcnn/", "762274": "Xception Imagenet weights (no top): https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xceptionweightstfdimorderingtfkernels_notop.h5", "762251": "- Bob’s library of image-quality feature-extractors\nhttps://pypi.org/project/bob.ip.qualitymeasure/\t\n\n- Scikit image\nhttps://pypi.org/project/scikit-image/\t\n\n- Lightgbm \nhttps://pypi.org/project/lightgbm/\t\n\n- Scikit learn\nhttps://pypi.org/project/scikit-learn/\t", "762249": "https://www.kaggle.com/xhlulu/flickrfaceshq-dataset-nvidia-resized-256px\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/123665\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/121173\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://ffmpeg.org/ffmpeg-all.html\nhttps://github.com/yeephycho/tensorflow-face-detection\nDCFC preview dataset\nyoutube.com\nCeleb-DF-v2(https://arxiv.org/abs/1909.12962)\n\n", "762180": "https://www.kaggle.com/josecarmona/ffmpeg-python-example-to-extract-audio-from-mp4", "762178": "1. https://github.com/ondyari/FaceForensics \n2. https://github.com/davidsandberg/facenet\n3. https://github.com/timesler/facenet-pytorch\n4. https://github.com/lukemelas/EfficientNet-PyTorch\n5. https://github.com/Cadene/pretrained-models.pytorch", "762150": "\nfacenet: https://github.com/timesler/facenet-pytorch\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nVgg19: https://download.pytorch.org/models/vgg19-dcbb9e9d.pth\nPre-trained models of VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\n\nface-recognition: https://github.com/ageitgey/face_recognition\nface-recognition-models: https://github.com/ageitgey/face_recognition_models\n\nblazeFace: https://github.com/hollance/BlazeFace-PyTorch\nFfmpeg: Static Build - https://www.kaggle.com/rakibilly/ffmpeg-static-build \n\nFake Audio dataset: https://datashare.is.ed.ac.uk/handle/10283/3336\n\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch \nhttps://github.com/hollance/BlazeFace-PyTorch\nRetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nPytorch Retinaface: https://github.com/biubug6/Pytorch_Retinaface\nWIDER FACE Dataset: http://shuoyang1213.me/WIDERFACE/\n\nResemblyzer: https://github.com/resemble-ai/Resemblyzer\nReal-Time Voice Cloning: https://github.com/CorentinJ/Real-Time-Voice-Cloning\nThe VoxCeleb Dataset\n http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1.html\n http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox2.html\nThe M-AILABS Speech Dataset\n https://www.caito.de/2019/01/the-m-ailabs-speech-dataset/\nhttps://librosa.github.io/librosa/", "762114": "https://librosa.github.io/librosa/", "762094": "Blazeface TensorFlow Lite model from https://github.com/google/mediapipe ", "762055": "https://www.kaggle.com/chopinforest1986/efficientnetb0b7-keras-weights", "762048": "External Data\n\n1.https://www.kaggle.com/prashantkikani/efficientnet-pytorch(pretrainedmodel)\n2.https://www.kaggle.com/rishabhiitbhu/pretrainedmodels(pretrained model)\n3.torchvision.models (pretrained model)\n4.https://www.kaggle.com/timesler/facenet-pytorch-vggface2(face extract)\n5.https://www.kaggle.com/humananalog/blazeface-pytorch and https://www.kaggle.com/humananalog/deepfakes-inference-demo (face extract)\n6.https://github.com/ondyari/FaceForensics （FaceForensics++ dataset）\n7.https://github.com/tensorflow/tensorflow/blob/3d86d8ce14989ca65a59ad4cf37f690694bf6267/tensorflow/contrib/factorization/python/ops/gmm.py (gmm model code)", "762038": "https://github.com/danmohaha/DSP-FWA", "762025": "https://github.com/calmisential/TensorFlow2.0_ResNet\nhttps://www.kaggle.com/xiaofengmao/retinaface\nhttps://www.kaggle.com/daddyjin/scikitvideo1111\nhttps://files.pythonhosted.org/packages/31/d4/bcdbad92101430ff9a5161eb7612fc0e66cae96daa81953edb17aa3d7c37/librosa-0.7.0rc1-py3-none-any.whl\nhttps://files.pythonhosted.org/packages/73/63/ebf4332964fba68f72cd78723a2722d77cf78f988b13461c6fbf30fc96bc/SoundFile-0.10.3.post1-py2.py3.cp26.cp27.cp32.cp33.cp34.cp35.cp36.pp27.pp32.pp33-none-win32.whl", "761998": "https://github.com/ondyari/FaceForensics", "761995": "https://pypi.org/project/pytorchcv/0.0.13/\nhttps://github.com/yeephycho/tensorflow-face-detection", "761987": "BlazeFace PyTorch: https://github.com/hollance/BlazeFace-PyTorch\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nAgeGenderDeepLearning: https://github.com/GilLevi/AgeGenderDeepLearning/tree/master/models\nKeras Pretrained Xception: https://keras.io/applications/#xception\nKeras Pretrained Inception: https://keras.io/applications/#inceptionv3\nKeras Pretrained Resnet: https://keras.io/applications/#resnet\nPublic Kaggle Kernel: \nFaceRecognition: https://github.com/grib0ed0v/face_recognition.pytorch\nPytorch Retinaface: https://github.com/biubug6/Pytorch_Retinaface\nEfficient Net Weights https://github.com/qubvel/efficientnet\nhttps://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md\nhttps://github.com/opencv/opencv/tree/master/data/haarcascades", "761933": "FaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nDeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nRetinaFace Pytorch detector: https://github.com/biubug6/Pytorch_Retinaface\nhrnet: https://github.com/HRNet/HRNet-Image-Classification\nFFMPEG: https://johnvansickle.com/ffmpeg/", "761891": "keras inception_v3 imagenet weights https://github.com/fchollet/deep-learning-models/releases/download/v0.5/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5\n\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\n\nyou tube faces dataset - http://www.cs.tau.ac.il/~wolf/ytfaces/\n(I don't see anywhere on this site/page that it cannot be used for commercial purposes. So I must assume the opposite to be true.)\n\nHoha Dataset - https://www.di.ens.fr/~laptev/actions/ (currently cant download but may use if commercially allowed)\n\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/\n", "761868": "https://drive.google.com/drive/folders/10AVxwiSundwHoJxONKYF5U9ieZHPkXXR?usp=sharing\ncontains 4 publicly shared files: yolo.weights, yolo.cfg, labels.txt & model.h5", "761863": "https://github.com/foamliu/InsightFace-v2\nhttps://github.com/ondyari/FaceForensics\nhttps://github.com/facebookresearch/SlowFast\nhttps://github.com/biubug6/Pytorch_Retinaface", "761806": "1) ffmpeg: (all declared)\n\n https://pypi.org/project/ffmpeg/ (offline pip installation)\n https://johnvansickle.com/ffmpeg/\n https://github.com/kkroening/ffmpeg-python\n https://www.kaggle.com/sheldonrobinson/video-tools\n\n2) OpenCV pre-trained weight XMLs: (declared)\n\n https://github.com/opencv/opencv/tree/master/data/haarcascades\n \n3) pre-trained models and weights listed these TF-hub official sites below\n\n [Mobilenet] https://tfhub.dev/s?fine-tunable=yes&module-type=image-classification&tf-version=tf2\n [Efficientnet] https://tfhub.dev/google/collections/efficientnet/1\n\n note: these models above are pre-trained using ImageNet (according to GGL)\n\n4) additional Python modules to be installed (offline) using pip:\n\n pydub: https://pypi.org/project/pydub/\n moviepy: https://pypi.org/project/moviepy/\n pandas: https://pypi.org/project/pandas/\n seaborn: https://pypi.org/project/seaborn/\n\n note: dependent modules on these above are included if not installed in default Kaggle Notebook platform\n\nQ: asking this just in case. we do NOT need to declare Python pip packages that can be installed additionally with '!pip install xxxx' even offline? am I understanding declare rule correct?", "761799": "A light and fast face detector (lffd):\n[https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB](https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB)\n[https://github.com/fyr91/face_detection](https://github.com/fyr91/face_detection)\n[https://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices/tree/master/face_detection](https://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices/tree/master/face_detection)\n\nSome architectures and pretrained models and datasets:\n[http://www.robots.ox.ac.uk/~vgg/data/ ](http://www.robots.ox.ac.uk/~vgg/data/ )\n[http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/](http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/)\n[http://www.robots.ox.ac.uk/~vgg/data/voxceleb/](http://www.robots.ox.ac.uk/~vgg/data/voxceleb/)\n[https://github.com/davidsandberg/facenet](https://github.com/davidsandberg/facenet)\n[https://github.com/ondyari/FaceForensics](https://github.com/ondyari/FaceForensics)\n\nLip Reading:\n[http://www.robots.ox.ac.uk/~vgg/data/lip_reading/](http://www.robots.ox.ac.uk/~vgg/data/lip_reading/)\n[https://github.com/astorfi/lip-reading-deeplearning (Apache License)](https://github.com/astorfi/lip-reading-deeplearning)\n[https://github.com/joseph-zhong/LipReading](https://github.com/joseph-zhong/LipReading)\n[https://github.com/afourast/deep_lip_reading](https://github.com/afourast/deep_lip_reading)\n[https://github.com/hassanhub/LipReading](https://github.com/hassanhub/LipReading)\n\nEverything at:\n[https://keras.io/applications/](https://keras.io/applications/)\n[https://github.com/keras-team/keras-applications](https://github.com/keras-team/keras-applications)\n\nOnnx: https://github.com/onnx/onnx\nONNXRuntime: https://www.wheelodex.org/projects/onnxruntime/ (and in general, any wheel at https://www.wheelodex.org/projects/)\nFaceRecognition: https://github.com/ageitgey/face_recognition\nBlazeFace: https://www.kaggle.com/humananalog/blazeface-pytorch", "761796": "Pretrained XCeption, ResNet", "761784": "Pytorchcv: https://pypi.org/project/pytorchcv/\nBlazeface: https://www.kaggle.com/humananalog/blazeface-pytorch\nResnext pretrained: https://www.kaggle.com/humananalog/deepfakes-inference-demo", "761755": "https://github.com/Cadene/pretrained-models.pytorch https://github.com/timesler/facenet-pytorch https://github.com/biubug6/Pytorch_Retinaface https://github.com/lukemelas/EfficientNet-PyTorch https://github.com/HRNet http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/ https://storage.googleapis.com/openimages/web/index.html https://www.cs.tau.ac.il/~wolf/ytfaces https://github.com/NVlabs/ffhq-dataset\n", "761675": "Pytorchcv: https://pypi.org/project/pytorchcv/0.0.13/\nA mobilenet SSD based face detector: https://github.com/yeephycho/tensorflow-face-detection\n", "761655": "http://www.robots.ox.ac.uk/~vgg/data/voxceleb/\nDataset and pre-trained models from VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/deepfakes/faceswap\nhttps://github.com/YuvalNirkin/face_swap\nhttps://github.com/ipazc/mtcnn", "761639": "\nFace Detection and other packages:\nhttps://github.com/ipazc/mtcnn (wheel: https://www.kaggle.com/unkownhihi/mtcnn-package)\nhttps://www.kaggle.com/robikscube/deepfakemodelspackages\n\nModel implementations and pretrained model weights:\nhttps://github.com/ondyari/FaceForensics\nhttps://data.lip6.fr\nhttps://github.com/ondyari/FaceForensics\nhttps://data.lip6.fr/cadene/pretrainedmodels/xception-b5690688.pth\nhttp://kaldir.vc.in.tum.de:/FaceForensics/models/faceforensics++_models.zip\nhttps://download.pytorch.org/models/r3d_18-b3b3357e.pth\nhttps://download.pytorch.org/models/r2plus1d_18-91a641e6.pth\n", "761633": "https://www.kaggle.com/rakibilly/ffmpeg-static-build\nhttps://sites.google.com/view/perception-cv4arvr/blazeface\nhttps://github.com/keras-team/keras-applications", "761627": "https://github.com/marl/crepe\nhttps://github.com/santi-pdp/segan", "761621": "https://github.com/sthanhng/yoloface\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Linzaer/Face-Track-Detect-Extract\nhttps://github.com/cc-hpc-itwm/DeepFakeDetection\nhttps://github.com/Xtra-Computing/thundersvm\nhttps://github.com/iperov/DeepFaceLab\nhttps://github.com/deepfakes/faceswap\nhttps://github.com/dessa-research/DeepFake-Detection\nhttps://github.com/VainF/DeepLabV3Plus-Pytorch\nhttps://pytorch.org/hub/pytorch_vision_deeplabv3_resnet101/\nhttps://pytorch.org/hub/pytorch_vision_fcn_resnet101/\nhttps://github.com/SConsul/Global_Convolutional_Network\nhttps://github.com/cutz-j/FDFtNet\nhttps://research.google.com/youtube-bb/\nhttps://research.google.com/youtube8m/index.html\nhttps://www.kaggle.com/c/deepfake-detection-challenge/discussion/121173\nhttps://github.com/DariusAf/MesoNet\nhttps://github.com/nii-yamagishilab/Capsule-Forensics-v2\nhttps://github.com/danmohaha/DSP-FWA\nhttps://github.com/mvaleriani/Shallow\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch\nhttps://github.com/grib0ed0v/face_recognition.pytorch\nhttps://github.com/nii-yamagishilab/ClassNSeg\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pypi.org/project/face-recognition/\nhttps://github.com/microsoft/LightGBM\nhttps://github.com/dmlc/xgboost\nhttps://github.com/adobe/antialiased-cnns\nhttps://github.com/amilworks/GanDetection\nhttps://github.com/tensorflow/models", "761587": "Playing around with hybrid quantum-classical computations: \nhttps://github.com/XanaduAI/PennyLane\nhttps://github.com/XanaduAI/pennylane-qiskit (edited)", "761547": "https://github.com/stanfordnlp/mac-network", "761528": "https://github.com/PeterWang512/CNNDetection", "761522": "Xception: https://www.dropbox.com/s/1hplpzet9d7dv29/xception-c0a72b38.pth.tar?dl=1\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nDeepfakes Inference Demo: https://www.kaggle.com/humananalog/deepfakes-inference-demo\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nEDIT. Not used: RetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nEDIT. Not used:FaceDetection-DSFD: https://github.com/TencentYoutuResearch/FaceDetection-DSFD\nEDIT. Not used:lightDSFD: https://github.com/lijiannuist/lightDSFD\nEDIT. Not used:FaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nEDIT. Not used:DeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html\n\nImageNet pre-trained models in pytorchcv: https://pypi.org/project/pytorchcv/\nImageNet pre-trained models in PyTorch Hub: https://pytorch.org/hub/\npythonspeechfeatures: https://pypi.org/project/python_speech_features/0.4/\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/tengshaofeng/ResidualAttentionNetwork-pytorch\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch", "761501": "Youtube-df\nhttp://cs.uef.fi/deepfake_dataset/", "761494": "ImageNet\nJFT dataset\nWIDERFace\nhttps://github.com/tensorflow/models/ and corresponding training datasets\nhttps://pytorch.org/docs/stable/torchvision/models.html and corresponding training datasets\nhttps://github.com/rwightman/pytorch-image-models and corresponding training datasets\nhttps://github.com/rwightman/gen-efficientnet-pytorch and corresponding training datasets\nhttps://github.com/iitzco/faced and corresponding training datasets\nRetinaFace : https://github.com/deepinsight/insightface/tree/master/RetinaFace,\nhttps://github.com/biubug6/Pytorch_Retinaface and corresponding training datasets\nFacenet: https://github.com/davidsandberg/facenet, https://github.com/timesler/facenet-pytorch and corresponding training datasets\nhttps://github.com/open-mmlab/mmdetection", "761488": "Stingray: https://github.com/StingraySoftware/stingray (MIT license)", "761479": "AudioSet: https://research.google.com/audioset/dataset/index.html\nUAFDV: https://drive.google.com/drive/folders/1GEk1DSxmlV_61JtpEGzC9Fo_BffvyxpH\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics/blob/master/Celeb-DF-v2/README.md\nFaceForensics++: https://github.com/ondyari/FaceForensics\n300W: https://ibug.doc.ic.ac.uk/resources/300-W/", "761446": "* [UMDFaces](http://umdfaces.io/)\n* [MS-Celeb-1M](https://academictorrents.com/details/9e67eb7cc23c9417f39778a8e06cca5e26196a97/tech&hit=1&filelist=1)\n* [YouTube Faces DB](https://www.cs.tau.ac.il/~wolf/ytfaces/)\n* https://github.com/deepfakeinthewild/deepfake_in_the_wild\n* https://github.com/NVIDIA/DeepLearningExamples\n", "761381": "mtcnn package:\n[https://www.kaggle.com/unkownhihi/mtcnn-package](https://www.kaggle.com/unkownhihi/mtcnn-package)\nffmpeg static build:\n[https://www.kaggle.com/rakibilly/ffmpeg-static-build](https://www.kaggle.com/rakibilly/ffmpeg-static-build)\ninception_resnet_v2:\n[https://github.com/tensorflow/models/blob/master/research/slim/nets/inception_resnet_v2.py](https://github.com/tensorflow/models/blob/master/research/slim/nets/inception_resnet_v2.py)\nDCGAN:\n[https://github.com/tensorflow/models/blob/master/research/slim/nets/dcgan.py](https://github.com/tensorflow/models/blob/master/research/slim/nets/dcgan.py)\nCycleGAN:\n[https://github.com/tensorflow/models/blob/master/research/slim/nets/cyclegan.py](https://github.com/tensorflow/models/blob/master/research/slim/nets/cyclegan.py)\nOpenCV Haarcascades:\n[https://github.com/opencv/opencv/tree/master/data/haarcascades](https://github.com/opencv/opencv/tree/master/data/haarcascades)\nRealtime Glasses Detection:\n[https://github.com/TianxingWu/realtime-glasses-detection](https://github.com/TianxingWu/realtime-glasses-detection)\nKeras Applications:\n[https://www.tensorflow.org/api_docs/python/tf/keras/applications/](https://www.tensorflow.org/api_docs/python/tf/keras/applications/)\nMobileNet face extractor:\nhttps://github.com/yeephycho/tensorflow-face-detection/tree/master/model", "761369": "librosa: https://github.com/librosa/librosa (ISC license)", "761356": "FaceForensics+: https://github.com/ondyari/FaceForensics\nmodel\": https://data.lip6.fr/cadene/pretrainedmodels/xception-b5690688.pth\nmodel weights: : http://kaldir.vc.in.tum.de:/FaceForensics/models/faceforensics++_models.zip", "761319": "The TIMIT dataset, reportedly used by the elite in this competition, is not available without a '.edu' or similar email address.\nAnd thus, should NOT be eligible to this competition.\nBTW : What's all this mess about availability of the data from the authors ???", "761283": "https://github.com/biubug6/Pytorch_Retinaface\n", "761264": "https://github.com/open-mmlab/mmcv\nhttps://github.com/opencv/dldt - OpenVINO\nhttps://github.com/opencv/open_model_zoo", "761254": "https://github.com/rcmalli/keras-vggface/ \n(with all models)\n\nhttps://github.com/the-house-of-black-and-white/hall-of-faces", "761249": "TensorFlow Speech Recognition Challenge data: https://www.kaggle.com/c/tensorflow-speech-recognition-challenge/data\nRetinaFace MobileNet0.25: https://github.com/deepinsight/insightface/issues/669", "761239": "face_recognition: https://github.com/ageitgey/face_recognition\nopenface: https://github.com/cmusatyalab/openface\nDeep Residual Learning for Image Recognition: deep-residuahttps://github.com/KaimingHe/deep-residual-networks\nLFW dataset: http://vis-www.cs.umass.edu/lfw/\n", "761228": "vggface : https://github.com/rcmalli/keras-vggface", "761182": "EfficientNet-pyTorch: https://github.com/lukemelas/EfficientNet-PyTorch \nresnext101: https://github.com/facebookresearch/WSL-Images \nface.evoLVe.PyTorch: https://github.com/ZhaoJ9014/face.evoLVe.PyTorch \nInsightface:https://github.com/deepinsight/insightface\nhrnet: https://github.com/HRNet/HRNet-Image-Classification \ntorchvision models: https://github.com/pytorch/vision/tree/master/torchvision/models \nCennternet:https://github.com/xingyizhou/CenterNet \nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nRetinafacePytorch:https://github.com/biubug6/Pytorch_Retinaface\nFaceForensics++ dataset: https://github.com/ondyari/FaceForensics \nFFMPEG: https://johnvansickle.com/ffmpeg/\nTensorRT: https://github.com/NVIDIA/TensorRT \nOnnx: https://github.com/onnx/onnx\nOnnx-tensorRT: https://github.com/onnx/onnx-tensorrt\n", "761162": "TensorRT: [https://github.com/NVIDIA/TensorRT](https://github.com/NVIDIA/TensorRT) TensorRt-6.0.1.5.Ubuntu-14.04.x86_64-gnu.cuda-10.0.cudnn7.6.tar.gz\nOnnx: [https://github.com/onnx/onnx](https://github.com/onnx/onnx)\nOnnx-tensorRT: [https://github.com/onnx/onnx-tensorrt](https://github.com/onnx/onnx-tensorrt) ", "761119": "MTCNN: https://pypi.org/project/mtcnn/\nFacenet: https://github.com/nyoki-mtl/keras-facenet\n", "761088": "FFMPEG: https://johnvansickle.com/ffmpeg/\n\nFacenet: https://github.com/timesler/facenet-pytorch\n\nResnext: https://pytorch.org/hub/facebookresearch_WSL-Images_resnext/\n\nPytorch torch vision models https://pytorch.org/docs/stable/torchvision/models.html\n\nNvidia DALI https://github.com/NVIDIA/DALI", "761044": "https://gluon-cv.mxnet.io/ and their model zoo https://gluon-cv.mxnet.io/model_zoo/index.html including ImageNet (still not sure if this is allowed) pre-trained models\nfilterpy http://github.com/rlabbe/filterpy\n", "761034": "youtube videos ds: https://deepfake-detection.s3.amazonaws.com/augment_deepfake.tar.gz", "761027": "pre-trained pytorch YOLO object detection https://github.com/eriklindernoren/PyTorch-YOLOv3", "760961": "MMSkeleton and some dependencies\nhttps://github.com/open-mmlab/mmskeleton\nhttps://pypi.org/project/addict/\nhttps://pypi.org/project/mmcv/\nhttps://pypi.org/project/lazy-import/\nhttps://github.com/cocodataset/cocoapi", "760957": "https://github.com/deepmind/sonnet", "760852": "Models & weights from:\n* https://github.com/keras-team/keras-applications\n* https://github.com/qubvel/efficientnet\n* https://github.com/ipazc/mtcnn\n* https://github.com/Star-Clouds/CenterFace", "760848": "https://www.kaggle.com/sheldonrobinson/video-tools", "760846": "https://www.kaggle.com/pranavpulijala/moviepy", "760715": "dLib Library: http://dlib.net https://www.kaggle.com/carlossouza/dlibpkg\n\nKeras InceptionResNetV2: https://keras.io/applications\n\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\n\nyolov3: https://github.com/ultralytics/yolov3\n\nface_recognition: https://github.com/ageitgey/face_recognition\n\nFaceForensics: https://github.com/ondyari/FaceForensics\n\nreal-and-fake-face-detection: https://www.kaggle.com/ciplab/real-and-fake-face-detection", "760692": "- Keras InceptionResNetV2: https://keras.io/applications -> input: https://www.kaggle.com/keras/inceptionresnetv2\n- dLib Library: http://dlib.net -> input: https://www.kaggle.com/carlossouza/dlibpkg\n- dLib mmod human face detector: http://dlib.net/files/mmod_human_face_detector.dat.bz2 -> input: https://www.kaggle.com/ip4963/dlib-model\n- Deepfake Detection Challenge dataset: https://www.kaggle.com/c/deepfake-detection-challenge", "760663": "EfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nDeepfakes Inference Demo: https://www.kaggle.com/humananalog/deepfakes-inference-demo\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nRetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nFaceDetection-DSFD: https://github.com/TencentYoutuResearch/FaceDetection-DSFD\nlightDSFD: https://github.com/lijiannuist/lightDSFD\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nDeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html\nWIDER FACE Dataset: http://shuoyang1213.me/WIDERFACE/", "760636": "Kaggle Facenet : https://www.kaggle.com/timesler/facenet-pytorch-vggface2", "760612": "https://github.com/foolwood/DaSiamRPN\nhttps://github.com/deepinsight/insightface/tree/master/RetinaFace", "760609": "https://github.com/deepinsight/insightface/tree/master/RetinaFace\nhttps://github.com/foolwood/DaSiamRPN", "760552": "https://github.com/thomasbrandon/mish-cuda\nhttps://developer.nvidia.com/tensorrt\nhttps://opencv.org/", "760540": "@juliaelliott @addisonhoward \nI'm a bit confused about usage of external data. For example there is a public available dataset. It has license which implies only research usage(no commercial). I'm not using this dataset directly in my experiments, but I'm using a trained model on this dataset which is public available on GitHub. It will be violation of the rules or not?", "760514": "MIT License\nhttps://pypi.org/project/mtcnn/\nFacenet: https://github.com/timesler/facenet-pytorch\nFFMPEG: https://johnvansickle.com/ffmpeg/\nResnext: https://pytorch.org/hub/facebookresearch_WSL-Images_resnext/\nBlazeface: https://github.com/hollance/BlazeFace-PyTorch\nhttps://opencv.org/\nEfficientNet-PyTorch: https://github.com/lukemelas/EfficientNet-PyTorch\nEfficientNet pretrained weights: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nalbumentations: https://github.com/albumentations-team/albumentations\nBlazeFace PyTorch: https://www.kaggle.com/humananalog/blazeface-pytorch\nDeepfakes Inference Demo: https://www.kaggle.com/humananalog/deepfakes-inference-demo\nfacenet-pytorch: https://github.com/timesler/facenet-pytorch\nRetinaFace: https://github.com/deepinsight/insightface/tree/master/RetinaFace\nFaceDetection-DSFD: https://github.com/TencentYoutuResearch/FaceDetection-DSFD\nlightDSFD: https://github.com/lijiannuist/lightDSFD\nFaceForensics++ Dataset: http://kaldir.vc.in.tum.de/faceforensics_benchmark/documentation\nDeepFakeDetection Dataset: https://ai.googleblog.com/2019/09/contributing-data-to-deepfake-detection.html\nWIDER FACE Dataset: http://shuoyang1213.me/WIDERFACE/\n\n\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\n\nhttps://github.com/yaojieliu/CVPR2019-DeepTreeLearningForZeroShotFaceAntispoofing\nhttps://github.com/yaojieliu/ECCV2018-FaceDeSpoofing\n\nhttps://github.com/EndlessSora/DeeperForensics-1.0\nFaceForensics++: https://github.com/ondyari/FaceForensics/\nCeleb-DF: https://github.com/danmohaha/celeb-deepfakeforensics\nDeepfakeTIMIT: https://www.idiap.ch/dataset/deepfaketimit/\n\nUADFV: The UADFV dataset - Xin Yang, Yuezun Li, and Siwei Lyu. Exposing deep fakes\nusing inconsistent head poses. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),\n2019.\n\nDF-TIMIT: Pavel Korshunov and Sebastien Marcel. Deepfakes: a new ´\nthreat to face recognition? assessment and detection. arXiv\npreprint arXiv:1812.08685, 2018.\n\nFF-DF: Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Chris- ¨\ntian Riess, Justus Thies, and Matthias Nießner. FaceForensics++: Learning to detect manipulated facial images. In\nICCV, 2019.\n\nDFD: Nicholas Dufour, Andrew Gully, Per Karlsson, Alexey Victor Vorbyov, Thomas Leung, Jeremiah Childs, and Christoph Bregler. Deepfakes detection dataset by google & jigsaw.\n\nhttps://github.com/dessa-research/DeepFake-Detection\n\ndlib: https://github.com/davisking/dlib\ndlib-models: https://github.com/davisking/dlib-models/blob/master/shape_predictor_68_face_landmarks.dat.bz2\nnumpy: https://github.com/numpy/numpy\nopencv: https://github.com/opencv/opencv\nscipy: https://github.com/scipy/scipy\nimutils: https://pypi.org/project/imutils/\nripser: https://github.com/scikit-tda/ripser.py\nglob: https://docs.python.org/3/library/glob.html\nos: https://docs.python.org/3/library/os.html\nmultiprocessing: https://docs.python.org/3/library/multiprocessing.html\nscikit-learn: https://scikit-learn.org/stable/index.html\nScikit-TDA: https://github.com/scikit-tda\ncsv: https://docs.python.org/3/library/csv.html", "760509": "Facenet: https://github.com/timesler/facenet-pytorch\nFFMPEG: https://johnvansickle.com/ffmpeg/\nResnext: https://pytorch.org/hub/facebookresearch_WSL-Images_resnext/\nBlazeface: https://github.com/hollance/BlazeFace-PyTorch\n", "760429": "Pytorch implementation of xception net with ImageNet pretrained weights by tstandley. [MIT License]\nModel: https://github.com/tstandley/Xception-PyTorch\nWeight: https://www.dropbox.com/s/1hplpzet9d7dv29/xception-c0a72b38.pth.tar?dl=1\n\nPytorch VGG16 pretrained weights:\nhttps://download.pytorch.org/models/vgg19-dcbb9e9d.pth", "760421": "https://pypi.org/project/librosa/", "760396": "https://www.cs.tau.ac.il/~wolf/ytfaces/\nhttps://www.kaggle.com/sophatvathana/casia-dataset#Sp_D_CND_A_sec0056_sec0015_0282.jpg\nhttps://github.com/Daniil-Osokin/lightweight-human-pose-estimation.pytorch\nhttps://github.com/HRNet\nhttps://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/MerHS/SASA-pytorch\nhttps://github.com/PeterWang512/CNNDetection\nhttps://github.com/tensorflow/magenta", "760349": "Keras EfficientNet Noisy Student Weights as referred here::\nhttps://www.kaggle.com/c/bengaliai-cv19/discussion/132894\n\nMay try : \nhttps://github.com/chen0040/keras-video-classifier\nhttps://github.com/sagarvegad/Video-Classification-CNN-and-LSTM-", "760324": "https://pypi.org/project/moviepy/", "760167": "https://github.com/nii-yamagishilab/Capsule-Forensics-v2", "760166": "*WIDER FACE: http://shuoyang1213.me/WIDERFACE/ (not allowed, no longer using)*\nmmdetection: https://github.com/open-mmlab/mmdetection\n\nEdited to add (March 1, 2020):\nhttps://github.com/piergiaj/pytorch-i3d\nhttps://github.com/kenshohara/video-classification-3d-cnn-pytorch\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch", "760147": "facenet-pytorch: https://github.com/timesler/facenet-pytorch\nMesoNet: https://github.com/DariusAf/MesoNet\nPytorch xception net with ImageNet pretrained weights by tstandley. [MIT License]\nModel: https://github.com/tstandley/Xception-PyTorch\nWeight: https://www.dropbox.com/s/1hplpzet9d7dv29/xception-c0a72b38.pth.tar?dl=1\nPytorch VGG19 pretrained weights:\nhttps://download.pytorch.org/models/vgg19-dcbb9e9d.pth\n", "760125": "https://serre-lab.clps.brown.edu/resource/hmdb-a-large-human-motion-database/", "759754": "Resnet Weights https://github.com/tensorflow/models/tree/master/official/r1/resnet\nEfficientnet Weights https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\nLFFD Weights https://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices\n", "759745": "[https://github.com/biubug6/Pytorch_Retinaface](https://github.com/biubug6/Pytorch_Retinaface)\n[https://github.com/ShiqiYu/libfacedetection](https://github.com/ShiqiYu/libfacedetection)\n[https://github.com/timesler/facenet-pytorch](https://github.com/timesler/facenet-pytorch)", "759719": "https://github.com/fyu/drn/\nhttps://github.com/biubug6/Pytorch_Retinaface.git\nhttps://github.com/ondyari/FaceForensics/tree/master/classification \nhttps://github.com/TreB1eN/InsightFace_Pytorch.git\n https://github.com/kenshohara/3D-ResNets-PyTorch\n https://github.com/facebookresearch/SlowFast\n https://github.com/open-mmlab/mmaction\n https://github.com/MIT-HAN-LAB/temporal-shift-module\n https://deepmind.com/research/open-source/kinetics\n\n", "759715": "https://github.com/protossw512/AdaptiveWingLoss\nhttps://github.com/TadasBaltrusaitis/OpenFace\nhttps://github.com/NVIDIA/flownet2-pytorch\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/qijiezhao/py-denseflow\nhttps://github.com/wanglimin/dense_flow\nhttps://github.com/yjxiong/tsn-pytorch\nhttps://github.com/metalbubble/TRN-pytorch\nhttps://github.com/mit-han-lab/temporal-shift-module", "759491": "https://github.com/supernotman/RetinaFace_Pytorch", "759455": "https://github.com/open-mmlab/mmdetection/", "759380": "COCO pre trained object detection https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md#coco-trained-models-coco-models\n\npretrained image net models as others mentioned", "759222": "https://pypi.org/project/Keras/", "759221": "https://pypi.org/project/ffmpeg-python/", "759216": "https://www.kaggle.com/rakibilly/ffmpeg-static-build", "759213": "https://www.kaggle.com/dagnelies/deepfake-faces", "759172": "https://github.com/dessa-research", "759035": "https://pytorch.org/hub/huggingface_pytorch-transformers/", "758791": "https://www.kaggle.com/chooyoungjun/deepmtcnn\nhttps://www.kaggle.com/sheldonrobinson/video-tools\nhttps://www.kaggle.com/chooyoungjun/deepfake14\nhttps://github.com/biubug6/Pytorch_Retinaface\nhttps://github.com/ShiqiYu/libfacedetection\nhttps://github.com/timesler/facenet-pytorch\nhttps://github.com/fyu/drn/\nhttps://github.com/biubug6/Pytorch_Retinaface.git\nhttps://github.com/TreB1eN/InsightFace_Pytorch.git\nhttps://github.com/kenshohara/3D-ResNets-PyTorch\nhttps://github.com/facebookresearch/SlowFast\nhttps://github.com/open-mmlab/mmaction\nhttps://github.com/MIT-HAN-LAB/temporal-shift-module\nhttps://deepmind.com/research/open-source/kinetics", "758668": "https://github.com/sthanhng/yoloface\nhttps://github.com/ultralytics/yolov3\nhttps://github.com/sthanhng/yoloface/blob/master/model-weights/get_models.sh\nhttps://pytorch.org/hub/pytorch_vision_resnet/\nhttps://github.com/pytorch/examples/tree/master/imagenet\nhttps://docs.opencv.org/3.4/d4/dee/tutorial_optical_flow.html\nhttps://pytorch.org/tutorials/beginner/audio_classifier_tutorial.html?highlight=audio\nhttp://openaccess.thecvf.com/content_ICCVW_2019/papers/HBU/Amerini_Deepfake_Video_Detection_through_Optical_Flow_Based_CNN_ICCVW_2019_paper.pdf", "758256": "https://github.com/aleju/imgaug\nhttps://github.com/deepmind/kinetics-i3d\nhttps://github.com/OanaIgnat/i3d_keras", "757983": "A Light and Fast Face Detector for Edge Devices\nhttps://github.com/YonghaoHe/A-Light-and-Fast-Face-Detector-for-Edge-Devices\nUltra-Light-Fast-Generic-Face-Detector-1MB\nhttps://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB", "757863": "https://github.com/deepmind/kinetics-i3d/", "757194": "Pretrained models from here: https://github.com/pytorch/fairseq", "756818": "LFW - People (Face Recognition) License GNU Lesser General Public License 3.0\nhttps://www.kaggle.com/atulanandjha/lfwpeople\nor http://vis-www.cs.umass.edu/lfw/", "756553": "https://pypi.org/project/efficientnet/\nimagenet weights", "756518": "MTCNN package - https://www.kaggle.com/diamondsnake/mtcnnpackage\nKeras applications models - https://keras.io/applications/\n", "756079": "https://github.com/ufoym/imbalanced-dataset-sampler", "756071": "Pre-trained models of VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/", "755919": "- Toolchain & pre-trained model: https://github.com/dessa-public/DeepFake-Detection\n- Dataset of Real and Fake Face Detection : https://www.kaggle.com/ciplab/real-and-fake-face-detection\n- Dataset and pre-trained models of VGGFace2: http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/ \n- Pre-trained models from https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet\n- FaceNet \n - Wrapper: https://pypi.org/project/mtcnn/\n - Core library & pre-trained models: https://github.com/davidsandberg/facenet \n- FFMPEG tools: https://github.com/kkroening/ffmpeg-python", "755756": "https://github.com/hollance/BlazeFace-PyTorch\nsome changes in https://www.kaggle.com/humananalog/deepfakes-inference-demo's helpers code\n\n", "755630": "faceforensics++ model\nhttp://kaldir.vc.in.tum.de/FaceForensics/models/faceforensics++_models.zip\npretrainedmodels\nhttps://files.pythonhosted.org/packages/84/0e/be6a0e58447ac16c938799d49bfb5fb7a80ac35e137547fc6cee2c08c4cf/pretrainedmodels-0.7.4.tar.gz\nface_recognition\nhttps://files.pythonhosted.org/packages/3f/ed/ad9a28042f373d4633fc8b49109b623597d6f193d3bbbef7780a5ee8eef2/face_recognition-1.2.3-py2.py3-none-any.whl\nface-recognition-models\nhttps://files.pythonhosted.org/packages/cf/3b/4fd8c534f6c0d1b80ce0973d01331525538045084c73c153ee6df20224cf/face_recognition_models-0.3.0.tar.gz\ndlib\nhttps://files.pythonhosted.org/packages/63/92/05c3b98636661cb80d190a5a777dd94effcc14c0f6893222e5ca81e74fbc/dlib-19.19.0.tar.gz\nhaarcascades\nhttps://github.com/opencv/opencv/tree/master/data/haarcascades\nshape_predictor_68_face_landmarks\nhttps://ja.osdn.net/projects/sfnet_dclib/downloads/dlib/v18.10/shape_predictor_68_face_landmarks.dat.bz2/", "755565": "All pre-trained models referenced on https://tfhub.dev before the entry deadline.", "755551": "I3D with pretrained weights for kinetics in pytorch - https://github.com/piergiaj/pytorch-i3d\n3D ResNets for pytorch https://github.com/kenshohara/3D-ResNets-PyTorch with pretrained weights from the google drive they link to", "755548": "DeepFace trained on VGGFace2\nGoogle facial expression comparison dataset (CC0) https://research.google/tools/datasets/google-facial-expression/", "755439": "https://github.com/protossw512/AdaptiveWingLoss\nhttps://wywu.github.io/projects/LAB/WFLW.html - Wider Facial Landmarks in-the-wild ", "755244": "\"Post links to your external data sources here before the deadline specified in the rules.\"\n\nIs the date March 31st? There's an early March deadline of March 3rd for mergers, etc. I didn't see anywhere in the rules about \"external data disclosure\" deadlines.\n\nPlease let me know and thank you!\n\nRodney\n", "754869": "RetinaFace : [https://github.com/deepinsight/insightface/tree/master/RetinaFace](https://github.com/deepinsight/insightface/tree/master/RetinaFace)", "754851": "https://github.com/scikit-video/scikit-video\nhttps://github.com/facebookresearch/detectron2\nhttps://github.com/pytorch/vision/tree/master/torchvision/models\nhttps://github.com/yhenon/pytorch-retinanet\nhttps://github.com/kenshohara/3D-ResNets-PyTorch\nhttps://github.com/toandaominh1997/EfficientDet.Pytorch", "754743": "Nice repo (MIT license) with pretrained models: https://github.com/dessa-public/DeepFake-Detection\n\nAstronomy library with some useful image analysis tools: https://github.com/keflavich/agpy", "754488": "\ndetectron2 model zoo (https://github.com/facebookresearch/detectron2/blob/master/MODEL_ZOO.md)", "754480": "http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nhttps://pypi.org/project/mtcnn/", "754477": "https://pypi.org/project/face-recognition/", "754464": "BlazeFace Tensorflow light model (.tflite): https://github.com/google/mediapipe/blob/master/mediapipe/models/face_detection_front.tflite", "754279": "ffmpeg-python\n- Github [https://github.com/kkroening/ffmpeg-python](https://github.com/kkroening/ffmpeg-python)\n- ffmpeg-python wheel for offline installation by @phoenix9032 [https://www.kaggle.com/phoenix9032/ffmpegpython](https://www.kaggle.com/phoenix9032/ffmpegpython)\n", "754273": "Do we need to post implementations of specific layers as well (here and in general)? Like for example a special loss function we want to use that is not per default implemented in tensorflow / pytorch?", "753980": "Yolo V3 face detector: https://github.com/sthanhng/yoloface", "753887": "https://github.com/ipazc/mtcnn\nAlso, a blanket declaration for all models under Module: tf.keras.applications (https://github.com/tensorflow/tensorflow)", "753172": "https://github.com/Cadene/pretrained-models.pytorch/blob/master/pretrainedmodels/models/senet.py", "752948": "https://github.com/qubvel/segmentation_models\nhttps://github.com/ipazc/mtcnn", "752820": "Real and fake face dataset: https://www.kaggle.com/ciplab/real-and-fake-face-detection", "752661": "https://github.com/d-li14/mobilenetv3.pytorch", "752609": "Celebrity face dataset: \nhttps://github.com/prateekmehta59/Celebrity-Face-Recognition-Dataset", "752490": "Facenet Implementation by keras2 and model weight\nhttps://github.com/nyoki-mtl/keras-facenet", "752306": "face2 dataset : http://www.robots.ox.ac.uk/~vgg/data/vgg_face2/\nimagenet of kaggle competition dataset : https://www.kaggle.com/c/imagenet-object-localization-challenge/data\n\nI am very new on kaggle so could someone confirm me that Creative Commons Attribution-ShareAlike 4.0 International License is ok ?\n", "751682": "MTCNN package\nhttps://www.kaggle.com/diamondsnake/mtcnnpackage", "751588": "https://github.com/qubvel/efficientnet\nhttps://github.com/qubvel/classification_models\nhttps://github.com/1adrianb/face-alignment\nhttps://www.kaggle.com/humananalog/deepfakes-inference-demo", "751556": "pytorch pretrained model collections:\nhttps://github.com/Cadene/pretrained-models.pytorch", "750470": "pytorch pretrained models:\n 'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n 'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n 'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n 'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n 'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n 'resnext50_32x4d': 'https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pth',\n 'resnext101_32x8d': 'https://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pth',\n 'wide_resnet50_2': 'https://download.pytorch.org/models/wide_resnet50_2-95faca4d.pth',\n 'wide_resnet101_2': 'https://download.pytorch.org/models/wide_resnet101_2-32ee1156.pth',", "749074": "https://keras.io/applications/#inceptionresnetv2", "747037": "https://www.kaggle.com/selfishgene/youtube-faces-with-facial-keypoints\nPretrained models: https://github.com/osmr/imgclsmob", "744776": "https://github.com/qubvel/efficientnet\nefficientnet and pretrained weights ", "744586": "Facenet weights for keras: https://drive.google.com/open?id=1pwQ3H4aJ8a6yyJHZkTwtjcL4wYWQb7bn", "744330": "ASVspoof 2019: The 3rd Automatic Speaker Verification Spoofing and Countermeasures Challenge database\n[https://datashare.is.ed.ac.uk/handle/10283/3336](https://datashare.is.ed.ac.uk/handle/10283/3336)\nlicensed under The Open Data Commons Attribution License\n", "743643": "I'm going to be using posenet and tensorflow lite runtime to run it.\nhttps://www.tensorflow.org/lite/guide/python\nhttps://dl.google.com/coral/python/tflite_runtime-2.1.0-cp36-cp36m-linux_x86_64.whl\nhttps://www.tensorflow.org/lite/models/pose_estimation/overview\nhttps://storage.googleapis.com/download.tensorflow.org/models/tflite/posenet_mobilenet_v1_100_257x257_multi_kpt_stripped.tflite", "743361": "Can we use new external models/datasets after March 3rd ? because the deadline to disclose would have been passed then ?", "738129": "RetinaFace and ArcFace and pre-trained models in https://github.com/deepinsight/insightface", "737844": "Pytorch pretrained models from: https://pytorch.org/docs/stable/torchvision/models.html", "732513": "@tunguz' datasets under CC BY-NC 4.0 licence are allowed, or not allowed in this competition? A saw the question already below, sorry for repeat, but from the answer it is not clear for me.", "731744": "The EfficientNet repository as uploaded by @xhlulu:\n\nhttps://www.kaggle.com/xhlulu/efficientnet-keras-source-code", "728072": "AVA dataset: https://research.google.com/ava/download.html", "726941": "https://www.kaggle.com/carlossouza/dlibpkg\nhttps://www.kaggle.com/minhtam/imageio-ffmpeg\nhttps://www.kaggle.com/minhtam/face-recognition\nhttps://www.kaggle.com/minhtam/fake-detect-basic", "718340": "Open Images\n[https://www.kaggle.com/bigquery/open-images](https://www.kaggle.com/bigquery/open-images)", "715444": "ResNeXt pre-trained weights available via torch.hub or TorchVision:\nhttps://pytorch.org/hub/pytorch_vision_resnext/", "715442": "face.evoLVe: High-Performance Face Recognition Library based on PyTorch\n(including some pre-trained models):\nhttps://github.com/ZhaoJ9014/face.evoLVe.PyTorch", "712780": "External data that must be 1 GB, is the only pre-trained model weights's data or it includes any external model (Including what i have trained with myself)? Question is about size of external data", "712699": "DenseNet pre-trained weights", "711150": "Mesonet pre-trained weights\n[http://vis-www.cs.umass.edu/lfw/index.html#download](http://vis-www.cs.umass.edu/lfw/index.html#download)", "707231": "Keras FaceNet Model: https://www.kaggle.com/nikhil1011/facenet", "706256": "Resnet with imagenet weights", "705633": "Xception Imagenet Keras/Pytorch\nMobilenet SSD based on (https://github.com/yeephycho/tensorflow-face-detection)\nEfficientnet Models", "701554": "@juliaelliott , Are we allowed to add external libraries(Open Source Initiative licenses) to our submission notebook as files? \nAnd install them in the notebook when it's being executed.", "700725": "Real and Fake faces dataset:\nhttps://www.kaggle.com/ciplab/real-and-fake-face-detection", "694595": "http://www.cbsr.ia.ac.cn/users/xiangyuzhu/projects/3DDFA/main.htm", "693839": "Are we allowed to use pre-trained models available on free licenses, for example in order to perform face detection?\nI can't find anything about it in the rules [here](https://www.kaggle.com/c/deepfake-detection-challenge/rules) , so it should be allowed. On the other hand on the webpage of the competition [here](https://deepfakedetectionchallenge.ai/terms) I can read:\n> we ask that you make the following commitments: (...)\n- You are submitting information and models for the Deepfake Detection Challenge which belong to you and are a result of your own work.\n\nSo... Can I get a freely available model for face recognition built by somebody else, upload it as data, announce it in this thread and use it freely? Or is it prohibited and people must reinvent the wheel and prepare their own models for every aspect of the submission instead of focusing on the real challenge? I think the rules should be unified.", "772665": "", "762857": "https://github.com/thiago1080/SphereFace\nhttps://github.com/vlad3996/FaceDetection-DSFD\n", "762821": "", "762358": "", "762148": "", "761903": "", "761861": "", "761558": "", "761511": "", "761299": "External datasets:\nDeepfake150: https://www.kaggle.com/unkownhihi/deepfake\nFFHQ: https://www.kaggle.com/greatgamedota/ffhq-face-data-set\nReal and Fake datasets: https://www.kaggle.com/ciplab/real-and-fake-face-detection\n\nMy datasets:\nhomemade deepfake faces: https://www.kaggle.com/muerbingsha/99430-faces\npretrained models: \nhttps://www.kaggle.com/muerbingsha/deepfake-my-resnet\nhttps://www.kaggle.com/muerbingsha/deepfake-my-inception", "760173": "", "759727": "", "755942": "", "747766": "", "746490": "", "742576": "", "728722": "", "726419": "", "725676": "", "721000": "", "719344": "", "716584": "", "694027": "" } }![https://www.kaggle.com/sophatvathana/casia-dataset#Sp_D_CND_A_sec0056_sec0015_0282.jpg](\"https://www.kaggle.com/sophatvathana/casia-dataset#Sp_D_CND_A_sec0056_sec0015_0282.jpg\"){kind=link}