![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2309090%2F513d3ff2f28f851f28bcf87829fe9ea9%2F017RMS0PH1.jpg?generation=1576155331257406&alt=media\")
A little while ago I have uploaded a dataset of 1 Million Fake Faces to Kaggle Datasets. It has been broken into seven different datasets:
\n\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
\n\nIf you find any of those datasets useful, please remember to upvote them.
", "messages": [ { "id": 692741, "postDate": "2019-12-11T17:17:08.667Z", "content": "A little while ago I have uploaded a dataset of 1 Million Fake Faces to Kaggle Datasets. It has been broken into seven different datasets:
\n\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
\n\nIf you find any of those datasets useful, please remember to upvote them.
", "rawMarkdown": "A little while ago I have uploaded a dataset of 1 Million Fake Faces to Kaggle Datasets. It has been broken into seven different datasets:\n\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\n\nIf you find any of those datasets useful, please remember to upvote them. ", "votes": 216 }, { "id": 693498, "postDate": "2019-12-12T12:55:43.870Z", "content": "Great collection, thanks for sharing.\nI particularly liked the half man, half pizza.\n
It's a Walking Dead π @jaredsavage
", "rawMarkdown": "It's a Walking Dead π @jaredsavage ", "votes": 2 }, { "id": 693522, "postDate": "2019-12-12T13:24:42.440Z", "content": "I still can't believe that's actually a fake face. Really reminds me of someone I know. π
", "rawMarkdown": "I still can't believe that's actually a **fake** face. Really reminds me of someone I know. π ", "votes": 3 }, { "id": 693542, "postDate": "2019-12-12T13:51:21.507Z", "content": "Someone must have left a pack of Doritos in the training set π \n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2309090%2F22bc0614cfe071f625421ee00fc67a6e%2F800px-Nacho_Cheese_Flavor_Doritos_Ready_to_Party.jpg?generation=1576158656216937&alt=media\")
this competition is really interestionπ
", "rawMarkdown": "this competition is really interestionπ ", "votes": 1 }, { "id": 693687, "postDate": "2019-12-12T17:13:16.317Z", "content": "I, as a human like you are, think that is a real face and not a pack of Doritoes.
", "rawMarkdown": "I, as a human like you are, think that is a real face and not a pack of Doritoes.", "votes": 2 }, { "id": 693841, "postDate": "2019-12-12T20:54:54.210Z", "content": "But are they real Doritos or fake ones? π¬
", "rawMarkdown": "But are they real Doritos or fake ones? π¬ ", "votes": 1 }, { "id": 693989, "postDate": "2019-12-13T02:44:20.347Z", "content": "A NN would say those are Cheetos π€
", "rawMarkdown": "A NN would say those are Cheetos π€ ", "votes": 2 }, { "id": 694506, "postDate": "2019-12-13T18:06:57.753Z", "content": "Unless the NN was smart enough to notice the Doritos bag in the background πͺ
", "rawMarkdown": "Unless the NN was smart enough to notice the Doritos bag in the background πͺ", "votes": 2 }, { "id": 696835, "postDate": "2019-12-17T06:01:21.083Z", "content": "That would be data leakage though π€
", "rawMarkdown": "That would be data leakage though π€ ", "votes": 2 }, { "id": 3439641, "postDate": "2026-04-11T03:40:20.367Z", "content": "utterly terrifying
", "rawMarkdown": "utterly terrifying" } ] }, { "id": 692755, "postDate": "2019-12-11T17:44:41.247Z", "content": "π More data = Better!!....in general....in many cases....most of the time. π
", "rawMarkdown": "π More data = Better!!....in general....in many cases....most of the time. π ", "votes": 4, "replies": [ { "id": 693177, "postDate": "2019-12-12T05:33:49.503Z", "content": "To a point.
", "rawMarkdown": "To a point." } ] }, { "id": 844025, "postDate": "2020-05-12T12:17:31.063Z", "content": "Thanks for the collection!\nWhere did all these images come from? GANs or photoshopped or with other applications?
", "rawMarkdown": "Thanks for the collection!\nWhere did all these images come from? GANs or photoshopped or with other applications?", "votes": 1, "replies": [ { "id": 844212, "postDate": "2020-05-12T14:08:21.837Z", "content": "They were all generated using Nvidia's StyleGAN: https://github.com/NVlabs/stylegan
", "rawMarkdown": "They were all generated using Nvidia's StyleGAN: https://github.com/NVlabs/stylegan", "votes": 3 }, { "id": 844573, "postDate": "2020-05-12T17:28:47.517Z", "content": "Thank you very much :D
", "rawMarkdown": "Thank you very much :D", "votes": 1 } ] }, { "id": 791949, "postDate": "2020-03-30T19:49:21.597Z", "content": "Great useful collection!
", "rawMarkdown": "Great useful collection!", "votes": 1 }, { "id": 766889, "postDate": "2020-03-08T22:48:13.403Z", "content": "@tunguz Great collection of fake faces, and I would love to be able to use them for this competition, but unfortunately the license is \"Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)\", which prevents us from using it for any competition with money prizes. πͺ \nDon't worry you got anyhow my up vote for the effort π
", "rawMarkdown": "@tunguz Great collection of fake faces, and I would love to be able to use them for this competition, but unfortunately the license is \"Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)\", which prevents us from using it for any competition with money prizes. πͺ \nDon't worry you got anyhow my up vote for the effort π", "votes": 1, "replies": [ { "id": 778802, "postDate": "2020-03-18T18:40:24.037Z", "content": "Hi Nuno, do you think this dataset by xhlulu is allowed: https://www.kaggle.com/c/deepfake-detection-challenge/discussion/122786
\n\nAlso what are some datasets that are allowed that are useful?
\n\nThank you!
", "rawMarkdown": "Hi Nuno, do you think this dataset by xhlulu is allowed: https://www.kaggle.com/c/deepfake-detection-challenge/discussion/122786 \n\nAlso what are some datasets that are allowed that are useful?\n\nThank you!\n\n" } ] }, { "id": 754080, "postDate": "2020-02-23T02:59:36.180Z", "content": "Great collection ! It's appreciated.
", "rawMarkdown": "Great collection ! It's appreciated.", "votes": 1 }, { "id": 699239, "postDate": "2019-12-20T08:17:43.237Z", "content": "Mark
", "rawMarkdown": "Mark", "votes": 1 }, { "id": 699028, "postDate": "2019-12-20T01:36:26.960Z", "content": "Great work!
", "rawMarkdown": "Great work!\n", "votes": 1 }, { "id": 697905, "postDate": "2019-12-18T14:38:52.107Z", "content": "Thank you for this information, sir! Salute!
", "rawMarkdown": "Thank you for this information, sir! Salute!", "votes": 1 }, { "id": 697750, "postDate": "2019-12-18T10:58:13.280Z", "content": "Good
", "rawMarkdown": "Good", "votes": 1 }, { "id": 695914, "postDate": "2019-12-15T20:47:43.857Z", "content": "Thanks for sharing Bojan, great collection.
", "rawMarkdown": "Thanks for sharing Bojan, great collection.", "votes": 1 }, { "id": 695210, "postDate": "2019-12-14T19:06:38.957Z", "content": "Cool!\nAre they all generated using GANs or some other technique is used?
", "rawMarkdown": "Cool!\nAre they all generated using GANs or some other technique is used?", "votes": 1, "replies": [ { "id": 696752, "postDate": "2019-12-17T02:21:12.850Z", "rawMarkdown": "", "isDeleted": true }, { "id": 697116, "postDate": "2019-12-17T13:54:26.667Z", "content": "Didn't you generate them?
", "rawMarkdown": "Didn't you generate them?", "votes": 1 } ] }, { "id": 694310, "postDate": "2019-12-13T12:20:19.080Z", "content": "Very hot subject
", "rawMarkdown": "Very hot subject", "votes": 1 }, { "id": 694300, "postDate": "2019-12-13T12:09:04.710Z", "content": "ty
", "rawMarkdown": "ty", "votes": 1 }, { "id": 693559, "postDate": "2019-12-12T14:10:05.043Z", "content": "Wow! Awesome!!! Thanks for sharing!π π
", "rawMarkdown": "Wow! Awesome!!! Thanks for sharing!π π ", "votes": 1 }, { "id": 693448, "postDate": "2019-12-12T11:33:27.620Z", "content": "1 Million Fake Faces, 1 Million dollarsπ \ngood job @tunguz
", "rawMarkdown": "1 Million Fake Faces, 1 Million dollarsπ \ngood job @tunguz ", "votes": 1, "replies": [ { "id": 693470, "postDate": "2019-12-12T12:03:12.837Z", "content": "Winning the competition one fake face at the time. π
", "rawMarkdown": "Winning the competition one fake face at the time. π ", "votes": 3 } ] }, { "id": 692793, "postDate": "2019-12-11T18:25:23.467Z", "content": "is it possible to carry out this research with 20-30 GB of data, for those who don't have power powerful workstations?
", "rawMarkdown": "is it possible to carry out this research with 20-30 GB of data, for those who don't have power powerful workstations? ", "votes": 1, "replies": [ { "id": 692870, "postDate": "2019-12-11T20:03:04.590Z", "content": "Unfortunately, my guess is that doing well in this competition will be almost impossible without powerful hardware.
", "rawMarkdown": "Unfortunately, my guess is that doing well in this competition will be almost impossible without powerful hardware.", "votes": 2 }, { "id": 693200, "postDate": "2019-12-12T06:14:35.703Z", "content": "\"Powerful\" is very subject here π
", "rawMarkdown": "\"Powerful\" is very subject here π ", "votes": 1 }, { "id": 694504, "postDate": "2019-12-13T18:03:34.817Z", "content": "I'm in the same boat. My plan is to get something working on a small-ish set of data and then using AWS if I'm confident enough.
", "rawMarkdown": "I'm in the same boat. My plan is to get something working on a small-ish set of data and then using AWS if I'm confident enough.", "votes": 1 } ] }, { "id": 697982, "postDate": "2019-12-18T16:18:25.707Z", "content": "Will fake faces be relevant to deepfakes, which are usually real faces added to the videos? \nI guess we will have our answer in 4 months!
", "rawMarkdown": "Will fake faces be relevant to deepfakes, which are usually real faces added to the videos? \nI guess we will have our answer in 4 months!", "votes": 2, "replies": [ { "id": 700170, "postDate": "2019-12-21T14:44:51.533Z", "rawMarkdown": "", "isDeleted": true } ] }, { "id": 694029, "postDate": "2019-12-13T04:53:22.987Z", "content": "Is there any chance that your profile pic is a fake face? ;)
", "rawMarkdown": "Is there any chance that your profile pic is a fake face? ;)", "votes": 2, "replies": [ { "id": 694240, "postDate": "2019-12-13T10:31:06.357Z", "content": "Build an algorithm and find out. ;-)
", "rawMarkdown": "Build an algorithm and find out. ;-)", "votes": 5 }, { "id": 694721, "postDate": "2019-12-14T02:16:07.323Z", "content": "Now we know where the winning algorithm will be used π
", "rawMarkdown": "Now we know where the winning algorithm will be used π ", "votes": 3 } ] }, { "id": 1638278, "postDate": "2022-01-04T15:30:48.033Z", "content": "hi, I wanna ask. Creating fake faces requires real faces to train on. so where are the real ones using to generate fake faces in this collection come from? Are they from from the Flickr dataset collected by Nvidia?
\nThanks alot for the work u done and the answer
@tunguz Where did all of these fake faces come from? I know they were generated using StyleGAN, and I know to create fake faces, they require real faces to train on. Where are the real faces come from? Are they from Flickr dataset collected by Nvidia?
\nThanks alot for the work u done and the answer
Are these Data sets of Known fake faces or all faces to teach a machine to figure out which ones are fake?
", "rawMarkdown": "Are these Data sets of Known fake faces or all faces to teach a machine to figure out which ones are fake?" }, { "id": 793166, "postDate": "2020-03-31T21:01:10.287Z", "content": "I actually used them in a different project. I just wanted to thank you for this!
", "rawMarkdown": "I actually used them in a different project. I just wanted to thank you for this!" }, { "id": 742203, "postDate": "2020-02-11T06:35:03.837Z", "content": "These images are generated by GAN2?
", "rawMarkdown": "These images are generated by GAN2?" }, { "id": 715997, "postDate": "2020-01-11T05:58:46.017Z", "content": "You are much appreciated for uploading those fake faces!!
", "rawMarkdown": "You are much appreciated for uploading those fake faces!!" }, { "id": 707939, "postDate": "2020-01-01T18:22:48.417Z", "content": "after peeking several images, oh boy, I can't tell why the images are fake
", "rawMarkdown": "after peeking several images, oh boy, I can't tell why the images are fake" }, { "id": 696985, "postDate": "2019-12-17T10:26:06.743Z", "content": "Hi! Just saw in the competition rules that there should not be external data >= 1GB, I take that this does not include the auxiliary data we use to train our model?
", "rawMarkdown": "Hi! Just saw in the competition rules that there should not be external data >= 1GB, I take that this does not include the auxiliary data we use to train our model?", "replies": [ { "id": 698579, "postDate": "2019-12-19T12:16:26.473Z", "content": "Yes I suppose. This data can be used during the training procedure and as the model won't need it during inference time you don't have to upload it to the platform.
", "rawMarkdown": "Yes I suppose. This data can be used during the training procedure and as the model won't need it during inference time you don't have to upload it to the platform." } ] }, { "id": 694503, "postDate": "2019-12-13T18:02:10.353Z", "content": "How would I download a subset of these? For example, the subdir 1m_faces_00/?
", "rawMarkdown": "How would I download a subset of these? For example, the subdir 1m_faces_00/?" }, { "id": 741346, "postDate": "2020-02-10T14:11:51.633Z", "rawMarkdown": "", "isDeleted": true }, { "id": 705406, "postDate": "2019-12-28T22:19:24.740Z", "rawMarkdown": "", "isDeleted": true }, { "id": 698860, "postDate": "2019-12-19T19:33:49.210Z", "rawMarkdown": "", "votes": 1, "isDeleted": true }, { "id": 696751, "postDate": "2019-12-17T02:18:45.360Z", "rawMarkdown": "", "votes": 1, "isDeleted": true }, { "id": 696047, "postDate": "2019-12-16T03:44:50.520Z", "rawMarkdown": "", "votes": 1, "isDeleted": true }, { "id": 693987, "postDate": "2019-12-13T02:39:14.100Z", "rawMarkdown": "", "votes": 1, "isDeleted": true }, { "id": 692767, "postDate": "2019-12-11T17:53:51.913Z", "content": "Awesome thanks!
", "rawMarkdown": "Awesome thanks!", "votes": 4 }, { "id": 700554, "postDate": "2019-12-22T07:42:33.627Z", "content": "Thanks for Sharing!!!
", "rawMarkdown": "Thanks for Sharing!!!", "votes": 1 }, { "id": 698008, "postDate": "2019-12-18T16:58:24.180Z", "content": "Nice ,Thanks for sharing
", "rawMarkdown": "Nice ,Thanks for sharing", "votes": 1 }, { "id": 697908, "postDate": "2019-12-18T14:40:04.643Z", "content": "Thanks!
", "rawMarkdown": "Thanks!", "votes": 1 }, { "id": 697728, "postDate": "2019-12-18T10:24:34.183Z", "content": "thanks for your effort Bojan
", "rawMarkdown": "thanks for your effort Bojan", "votes": 1 }, { "id": 697549, "postDate": "2019-12-18T04:19:10.593Z", "content": "Thanks!
", "rawMarkdown": "Thanks!", "votes": 1 }, { "id": 697383, "postDate": "2019-12-17T20:28:01.900Z", "content": "Thanks much!
", "rawMarkdown": "Thanks much!", "votes": 1 }, { "id": 697321, "postDate": "2019-12-17T18:06:50.153Z", "content": "Thanks!
", "rawMarkdown": "Thanks!", "votes": 1 }, { "id": 696828, "postDate": "2019-12-17T05:36:15.517Z", "content": "Thanks much!
", "rawMarkdown": "Thanks much!", "votes": 1 }, { "id": 696630, "postDate": "2019-12-16T21:20:08.633Z", "content": "Thanks
", "rawMarkdown": "Thanks", "votes": 1 }, { "id": 696524, "postDate": "2019-12-16T18:16:09.203Z", "content": "Thanks
", "rawMarkdown": "Thanks", "votes": 1 }, { "id": 696325, "postDate": "2019-12-16T13:24:13.293Z", "content": "Thanks!
", "rawMarkdown": "Thanks!", "votes": 1 }, { "id": 696177, "postDate": "2019-12-16T08:14:06.950Z", "content": "Thanks!
", "rawMarkdown": "Thanks!", "votes": 1 }, { "id": 696144, "postDate": "2019-12-16T07:14:40.510Z", "content": "Thanks, It's helpful for me.
", "rawMarkdown": "Thanks, It's helpful for me.", "votes": 1 }, { "id": 696137, "postDate": "2019-12-16T07:05:44.410Z", "content": "Thanks
", "rawMarkdown": "Thanks", "votes": 1 }, { "id": 695921, "postDate": "2019-12-15T21:12:12.373Z", "content": "Nice! Thanks.
", "rawMarkdown": "Nice! Thanks.", "votes": 1 }, { "id": 695186, "postDate": "2019-12-14T18:07:17.753Z", "content": "Thanks for sharingπ
", "rawMarkdown": "Thanks for sharingπ ", "votes": 1 }, { "id": 695059, "postDate": "2019-12-14T13:46:44.360Z", "content": "Thanks for Sharing
", "rawMarkdown": "Thanks for Sharing", "votes": 1 }, { "id": 694529, "postDate": "2019-12-13T18:39:15.723Z", "content": "Thanks for sharingπ
", "rawMarkdown": "Thanks for sharingπ ", "votes": 1 }, { "id": 694460, "postDate": "2019-12-13T16:33:55.210Z", "content": "Thanks for sharing
", "rawMarkdown": "Thanks for sharing", "votes": 1 }, { "id": 693445, "postDate": "2019-12-12T11:32:02.900Z", "content": "Thanks for sharing!
", "rawMarkdown": "Thanks for sharing!", "votes": 1 }, { "id": 693286, "postDate": "2019-12-12T08:00:16.950Z", "content": "Thanks for sharing
", "rawMarkdown": "Thanks for sharing", "votes": 1 }, { "id": 693158, "postDate": "2019-12-12T05:00:05.680Z", "content": "Thanks for sharing
", "rawMarkdown": "Thanks for sharing", "votes": 1 }, { "id": 692868, "postDate": "2019-12-11T20:00:00.293Z", "content": "Thanks for sharing!
", "rawMarkdown": "Thanks for sharing!", "votes": 1 }, { "id": 700288, "postDate": "2019-12-21T18:32:25.623Z", "content": "Thanks for sharing !
", "rawMarkdown": "Thanks for sharing !", "votes": 2 }, { "id": 700002, "postDate": "2019-12-21T08:46:51.827Z", "content": "Thanks for sharing !
", "rawMarkdown": "Thanks for sharing !" }, { "id": 939958, "postDate": "2020-07-22T15:36:32.623Z", "content": "Thanks! Very useful dataset.
", "rawMarkdown": "Thanks! Very useful dataset." }, { "id": 713553, "postDate": "2020-01-08T12:13:24.590Z", "content": "Thanks for sharing Bojan ! :)
", "rawMarkdown": "Thanks for sharing Bojan ! :)" }, { "id": 707937, "postDate": "2020-01-01T18:19:06.573Z", "content": "Thanks!
", "rawMarkdown": "Thanks!" }, { "id": 703503, "postDate": "2019-12-26T09:08:03.447Z", "content": "Thanks for sharing.
", "rawMarkdown": "Thanks for sharing." }, { "id": 754073, "postDate": "2020-02-23T02:44:07.287Z", "content": "Thanks for sharing!
", "rawMarkdown": "Thanks for sharing!\n", "isDeleted": true } ], "comments": [ { "id": 693498, "author_name": "Jared Savage", "author_url": "", "post_date": "2019-12-12T12:55:43.870000", "content": "Great collection, thanks for sharing.\nI particularly liked the half man, half pizza.\n
It's a Walking Dead π @jaredsavage
", "votes": 2, "replies": [] }, { "id": 693522, "author_name": "Bojan Tunguz", "author_url": "", "post_date": "2019-12-12T13:24:42.440000", "content": "I still can't believe that's actually a fake face. Really reminds me of someone I know. π
", "votes": 3, "replies": [] }, { "id": 693542, "author_name": "Jared Savage", "author_url": "", "post_date": "2019-12-12T13:51:21.507000", "content": "Someone must have left a pack of Doritos in the training set π \n
this competition is really interestionπ
", "votes": 1, "replies": [] }, { "id": 693687, "author_name": "Raizazel", "author_url": "", "post_date": "2019-12-12T17:13:16.317000", "content": "I, as a human like you are, think that is a real face and not a pack of Doritoes.
", "votes": 2, "replies": [] }, { "id": 693841, "author_name": "wheresmycookie", "author_url": "", "post_date": "2019-12-12T20:54:54.210000", "content": "But are they real Doritos or fake ones? π¬
", "votes": 1, "replies": [] }, { "id": 693989, "author_name": "Sanyam Bhutani", "author_url": "", "post_date": "2019-12-13T02:44:20.347000", "content": "A NN would say those are Cheetos π€
", "votes": 2, "replies": [] }, { "id": 694506, "author_name": "wheresmycookie", "author_url": "", "post_date": "2019-12-13T18:06:57.753000", "content": "Unless the NN was smart enough to notice the Doritos bag in the background πͺ
", "votes": 2, "replies": [] }, { "id": 696835, "author_name": "Sanyam Bhutani", "author_url": "", "post_date": "2019-12-17T06:01:21.083000", "content": "That would be data leakage though π€
", "votes": 2, "replies": [] }, { "id": 3439641, "author_name": "Debajyati Dey", "author_url": "", "post_date": "2026-04-11T03:40:20.367000", "content": "utterly terrifying
", "votes": 0, "replies": [] } ] }, { "id": 692755, "author_name": "Carlo", "author_url": "", "post_date": "2019-12-11T17:44:41.247000", "content": "π More data = Better!!....in general....in many cases....most of the time. π
", "votes": 4, "replies": [ { "id": 693177, "author_name": "NAIN", "author_url": "", "post_date": "2019-12-12T05:33:49.503000", "content": "To a point.
", "votes": 0, "replies": [] } ] }, { "id": 844025, "author_name": "dabarre", "author_url": "", "post_date": "2020-05-12T12:17:31.063000", "content": "Thanks for the collection!\nWhere did all these images come from? GANs or photoshopped or with other applications?
", "votes": 1, "replies": [ { "id": 844212, "author_name": "Bojan Tunguz", "author_url": "", "post_date": "2020-05-12T14:08:21.837000", "content": "They were all generated using Nvidia's StyleGAN: https://github.com/NVlabs/stylegan
", "votes": 3, "replies": [] }, { "id": 844573, "author_name": "dabarre", "author_url": "", "post_date": "2020-05-12T17:28:47.517000", "content": "Thank you very much :D
", "votes": 1, "replies": [] } ] }, { "id": 791949, "author_name": "Aldo Kacorri", "author_url": "", "post_date": "2020-03-30T19:49:21.597000", "content": "Great useful collection!
", "votes": 1, "replies": [] }, { "id": 766889, "author_name": "Nuno Ferreira", "author_url": "", "post_date": "2020-03-08T22:48:13.403000", "content": "@tunguz Great collection of fake faces, and I would love to be able to use them for this competition, but unfortunately the license is \"Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)\", which prevents us from using it for any competition with money prizes. πͺ \nDon't worry you got anyhow my up vote for the effort π
", "votes": 1, "replies": [ { "id": 778802, "author_name": "Bo Peng", "author_url": "", "post_date": "2020-03-18T18:40:24.037000", "content": "Hi Nuno, do you think this dataset by xhlulu is allowed: https://www.kaggle.com/c/deepfake-detection-challenge/discussion/122786
\n\nAlso what are some datasets that are allowed that are useful?
\n\nThank you!
", "votes": 0, "replies": [] } ] }, { "id": 754080, "author_name": "Huu An Hua", "author_url": "", "post_date": "2020-02-23T02:59:36.180000", "content": "Great collection ! It's appreciated.
", "votes": 1, "replies": [] }, { "id": 699239, "author_name": "xxn-xx", "author_url": "", "post_date": "2019-12-20T08:17:43.237000", "content": "Mark
", "votes": 1, "replies": [] }, { "id": 699028, "author_name": "Tom Paes", "author_url": "", "post_date": "2019-12-20T01:36:26.960000", "content": "Great work!
", "votes": 1, "replies": [] }, { "id": 697905, "author_name": "Rangga Kusuma Dinata", "author_url": "", "post_date": "2019-12-18T14:38:52.107000", "content": "Thank you for this information, sir! Salute!
", "votes": 1, "replies": [] }, { "id": 697750, "author_name": "Nikolai Shabalov", "author_url": "", "post_date": "2019-12-18T10:58:13.280000", "content": "Good
", "votes": 1, "replies": [] }, { "id": 695914, "author_name": "Enzo", "author_url": "", "post_date": "2019-12-15T20:47:43.857000", "content": "Thanks for sharing Bojan, great collection.
", "votes": 1, "replies": [] }, { "id": 695210, "author_name": "Ali Mirferdos", "author_url": "", "post_date": "2019-12-14T19:06:38.957000", "content": "Cool!\nAre they all generated using GANs or some other technique is used?
", "votes": 1, "replies": [ { "id": 696752, "author_name": "", "author_url": "", "post_date": "2019-12-17T02:21:12.850000", "content": "", "votes": 0, "replies": [] }, { "id": 697116, "author_name": "Shubham Patil", "author_url": "", "post_date": "2019-12-17T13:54:26.667000", "content": "Didn't you generate them?
", "votes": 1, "replies": [] } ] }, { "id": 694310, "author_name": "Tuan Nguyen-Sy", "author_url": "", "post_date": "2019-12-13T12:20:19.080000", "content": "Very hot subject
", "votes": 1, "replies": [] }, { "id": 694300, "author_name": "xkotix", "author_url": "", "post_date": "2019-12-13T12:09:04.710000", "content": "ty
", "votes": 1, "replies": [] }, { "id": 693559, "author_name": "Miyabon", "author_url": "", "post_date": "2019-12-12T14:10:05.043000", "content": "Wow! Awesome!!! Thanks for sharing!π π
", "votes": 1, "replies": [] }, { "id": 693448, "author_name": "DiegoJohnson", "author_url": "", "post_date": "2019-12-12T11:33:27.620000", "content": "1 Million Fake Faces, 1 Million dollarsπ \ngood job @tunguz
", "votes": 1, "replies": [ { "id": 693470, "author_name": "Bojan Tunguz", "author_url": "", "post_date": "2019-12-12T12:03:12.837000", "content": "Winning the competition one fake face at the time. π
", "votes": 3, "replies": [] } ] }, { "id": 692793, "author_name": "Yash Upadhyay", "author_url": "", "post_date": "2019-12-11T18:25:23.467000", "content": "is it possible to carry out this research with 20-30 GB of data, for those who don't have power powerful workstations?
", "votes": 1, "replies": [ { "id": 692870, "author_name": "Bojan Tunguz", "author_url": "", "post_date": "2019-12-11T20:03:04.590000", "content": "Unfortunately, my guess is that doing well in this competition will be almost impossible without powerful hardware.
", "votes": 2, "replies": [] }, { "id": 693200, "author_name": "Sanyam Bhutani", "author_url": "", "post_date": "2019-12-12T06:14:35.703000", "content": "\"Powerful\" is very subject here π
", "votes": 1, "replies": [] }, { "id": 694504, "author_name": "wheresmycookie", "author_url": "", "post_date": "2019-12-13T18:03:34.817000", "content": "I'm in the same boat. My plan is to get something working on a small-ish set of data and then using AWS if I'm confident enough.
", "votes": 1, "replies": [] } ] }, { "id": 697982, "author_name": "S D", "author_url": "", "post_date": "2019-12-18T16:18:25.707000", "content": "Will fake faces be relevant to deepfakes, which are usually real faces added to the videos? \nI guess we will have our answer in 4 months!
", "votes": 2, "replies": [ { "id": 700170, "author_name": "", "author_url": "", "post_date": "2019-12-21T14:44:51.533000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 694029, "author_name": "pocket", "author_url": "", "post_date": "2019-12-13T04:53:22.987000", "content": "Is there any chance that your profile pic is a fake face? ;)
", "votes": 2, "replies": [ { "id": 694240, "author_name": "Bojan Tunguz", "author_url": "", "post_date": "2019-12-13T10:31:06.357000", "content": "Build an algorithm and find out. ;-)
", "votes": 5, "replies": [] }, { "id": 694721, "author_name": "Sanyam Bhutani", "author_url": "", "post_date": "2019-12-14T02:16:07.323000", "content": "Now we know where the winning algorithm will be used π
", "votes": 3, "replies": [] } ] }, { "id": 1638278, "author_name": "Toxxi.frank", "author_url": "", "post_date": "2022-01-04T15:30:48.033000", "content": "hi, I wanna ask. Creating fake faces requires real faces to train on. so where are the real ones using to generate fake faces in this collection come from? Are they from from the Flickr dataset collected by Nvidia?
\nThanks alot for the work u done and the answer
@tunguz Where did all of these fake faces come from? I know they were generated using StyleGAN, and I know to create fake faces, they require real faces to train on. Where are the real faces come from? Are they from Flickr dataset collected by Nvidia?
\nThanks alot for the work u done and the answer
Are these Data sets of Known fake faces or all faces to teach a machine to figure out which ones are fake?
", "votes": 0, "replies": [] }, { "id": 793166, "author_name": "", "author_url": "", "post_date": "2020-03-31T21:01:10.287000", "content": "I actually used them in a different project. I just wanted to thank you for this!
", "votes": 0, "replies": [] }, { "id": 742203, "author_name": "Clark Albert", "author_url": "", "post_date": "2020-02-11T06:35:03.837000", "content": "These images are generated by GAN2?
", "votes": 0, "replies": [] }, { "id": 715997, "author_name": "Owen Xu", "author_url": "", "post_date": "2020-01-11T05:58:46.017000", "content": "You are much appreciated for uploading those fake faces!!
", "votes": 0, "replies": [] }, { "id": 707939, "author_name": "", "author_url": "", "post_date": "2020-01-01T18:22:48.417000", "content": "after peeking several images, oh boy, I can't tell why the images are fake
", "votes": 0, "replies": [] }, { "id": 696985, "author_name": "Nicholas Lyu", "author_url": "", "post_date": "2019-12-17T10:26:06.743000", "content": "Hi! Just saw in the competition rules that there should not be external data >= 1GB, I take that this does not include the auxiliary data we use to train our model?
", "votes": 0, "replies": [ { "id": 698579, "author_name": "Amey Patil", "author_url": "", "post_date": "2019-12-19T12:16:26.473000", "content": "Yes I suppose. This data can be used during the training procedure and as the model won't need it during inference time you don't have to upload it to the platform.
", "votes": 0, "replies": [] } ] }, { "id": 694503, "author_name": "wheresmycookie", "author_url": "", "post_date": "2019-12-13T18:02:10.353000", "content": "How would I download a subset of these? For example, the subdir 1m_faces_00/?
", "votes": 0, "replies": [] }, { "id": 741346, "author_name": "", "author_url": "", "post_date": "2020-02-10T14:11:51.633000", "content": "", "votes": 0, "replies": [] }, { "id": 705406, "author_name": "", "author_url": "", "post_date": "2019-12-28T22:19:24.740000", "content": "", "votes": 0, "replies": [] }, { "id": 698860, "author_name": "", "author_url": "", "post_date": "2019-12-19T19:33:49.210000", "content": "", "votes": 1, "replies": [] }, { "id": 696751, "author_name": "", "author_url": "", "post_date": "2019-12-17T02:18:45.360000", "content": "", "votes": 1, "replies": [] }, { "id": 696047, "author_name": "", "author_url": "", "post_date": "2019-12-16T03:44:50.520000", "content": "", "votes": 1, "replies": [] }, { "id": 693987, "author_name": "", "author_url": "", "post_date": "2019-12-13T02:39:14.100000", "content": "", "votes": 1, "replies": [] }, { "id": 692767, "author_name": "", "author_url": "", "post_date": "2019-12-11T17:53:51.913000", "content": "", "votes": 4, "replies": [] }, { "id": 700554, "author_name": "", "author_url": "", "post_date": "2019-12-22T07:42:33.627000", "content": "", "votes": 1, "replies": [] }, { "id": 698008, "author_name": "", "author_url": "", "post_date": "2019-12-18T16:58:24.180000", "content": "", "votes": 1, "replies": [] }, { "id": 697908, "author_name": "", "author_url": "", "post_date": "2019-12-18T14:40:04.643000", "content": "", "votes": 1, "replies": [] }, { "id": 697728, "author_name": "", "author_url": "", "post_date": "2019-12-18T10:24:34.183000", "content": "", "votes": 1, "replies": [] }, { "id": 697549, "author_name": "", "author_url": "", "post_date": "2019-12-18T04:19:10.593000", "content": "", "votes": 1, "replies": [] }, { "id": 697383, "author_name": "", "author_url": "", "post_date": "2019-12-17T20:28:01.900000", "content": "", "votes": 1, "replies": [] }, { "id": 697321, "author_name": "", "author_url": "", "post_date": "2019-12-17T18:06:50.153000", "content": "", "votes": 1, "replies": [] }, { "id": 696828, "author_name": "", "author_url": "", "post_date": "2019-12-17T05:36:15.517000", "content": "", "votes": 1, "replies": [] }, { "id": 696630, "author_name": "", "author_url": "", "post_date": "2019-12-16T21:20:08.633000", "content": "", "votes": 1, "replies": [] }, { "id": 696524, "author_name": "", "author_url": "", "post_date": "2019-12-16T18:16:09.203000", "content": "", "votes": 1, "replies": [] }, { "id": 696325, "author_name": "", "author_url": "", "post_date": "2019-12-16T13:24:13.293000", "content": "", "votes": 1, "replies": [] }, { "id": 696177, "author_name": "", "author_url": "", "post_date": "2019-12-16T08:14:06.950000", "content": "", "votes": 1, "replies": [] }, { "id": 696144, "author_name": "", "author_url": "", "post_date": "2019-12-16T07:14:40.510000", "content": "", "votes": 1, "replies": [] }, { "id": 696137, "author_name": "", "author_url": "", "post_date": "2019-12-16T07:05:44.410000", "content": "", "votes": 1, "replies": [] }, { "id": 695921, "author_name": "", "author_url": "", "post_date": "2019-12-15T21:12:12.373000", "content": "", "votes": 1, "replies": [] }, { "id": 695186, "author_name": "", "author_url": "", "post_date": "2019-12-14T18:07:17.753000", "content": "", "votes": 1, "replies": [] }, { "id": 695059, "author_name": "", "author_url": "", "post_date": "2019-12-14T13:46:44.360000", "content": "", "votes": 1, "replies": [] }, { "id": 694529, "author_name": "", "author_url": "", "post_date": "2019-12-13T18:39:15.723000", "content": "", "votes": 1, "replies": [] }, { "id": 694460, "author_name": "", "author_url": "", "post_date": "2019-12-13T16:33:55.210000", "content": "", "votes": 1, "replies": [] }, { "id": 693445, "author_name": "", "author_url": "", "post_date": "2019-12-12T11:32:02.900000", "content": "", "votes": 1, "replies": [] }, { "id": 693286, "author_name": "", "author_url": "", "post_date": "2019-12-12T08:00:16.950000", "content": "", "votes": 1, "replies": [] }, { "id": 693158, "author_name": "", "author_url": "", "post_date": "2019-12-12T05:00:05.680000", "content": "", "votes": 1, "replies": [] }, { "id": 692868, "author_name": "", "author_url": "", "post_date": "2019-12-11T20:00:00.293000", "content": "", "votes": 1, "replies": [] }, { "id": 700288, "author_name": "", "author_url": "", "post_date": "2019-12-21T18:32:25.623000", "content": "", "votes": 2, "replies": [] }, { "id": 700002, "author_name": "", "author_url": "", "post_date": "2019-12-21T08:46:51.827000", "content": "", "votes": 0, "replies": [] }, { "id": 939958, "author_name": "", "author_url": "", "post_date": "2020-07-22T15:36:32.623000", "content": "", "votes": 0, "replies": [] }, { "id": 713553, "author_name": "", "author_url": "", "post_date": "2020-01-08T12:13:24.590000", "content": "", "votes": 0, "replies": [] }, { "id": 707937, "author_name": "", "author_url": "", "post_date": "2020-01-01T18:19:06.573000", "content": "", "votes": 0, "replies": [] }, { "id": 703503, "author_name": "", "author_url": "", "post_date": "2019-12-26T09:08:03.447000", "content": "", "votes": 0, "replies": [] }, { "id": 754073, "author_name": "", "author_url": "", "post_date": "2020-02-23T02:44:07.287000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "692741": "A little while ago I have uploaded a dataset of 1 Million Fake Faces to Kaggle Datasets. It has been broken into seven different datasets:\n\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\n\nIf you find any of those datasets useful, please remember to upvote them. ", "693498": "Great collection, thanks for sharing.\nI particularly liked the half man, half pizza.\n\n", "692755": "π More data = Better!!....in general....in many cases....most of the time. π ", "844025": "Thanks for the collection!\nWhere did all these images come from? GANs or photoshopped or with other applications?", "791949": "Great useful collection!", "766889": "@tunguz Great collection of fake faces, and I would love to be able to use them for this competition, but unfortunately the license is \"Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)\", which prevents us from using it for any competition with money prizes. πͺ \nDon't worry you got anyhow my up vote for the effort π", "754080": "Great collection ! It's appreciated.", "699239": "Mark", "699028": "Great work!\n", "697905": "Thank you for this information, sir! Salute!", "697750": "Good", "695914": "Thanks for sharing Bojan, great collection.", "695210": "Cool!\nAre they all generated using GANs or some other technique is used?", "694310": "Very hot subject", "694300": "ty", "693559": "Wow! Awesome!!! Thanks for sharing!π π ", "693448": "1 Million Fake Faces, 1 Million dollarsπ \ngood job @tunguz ", "692793": "is it possible to carry out this research with 20-30 GB of data, for those who don't have power powerful workstations? ", "697982": "Will fake faces be relevant to deepfakes, which are usually real faces added to the videos? \nI guess we will have our answer in 4 months!", "694029": "Is there any chance that your profile pic is a fake face? ;)", "1638278": "hi, I wanna ask. Creating fake faces requires real faces to train on. so where are the real ones using to generate fake faces in this collection come from? Are they from from the Flickr dataset collected by Nvidia?\nThanks alot for the work u done and the answer", "1637970": "@tunguz Where did all of these fake faces come from? I know they were generated using StyleGAN, and I know to create fake faces, they require real faces to train on. Where are the real faces come from? Are they from Flickr dataset collected by Nvidia?\nThanks alot for the work u done and the answer", "794221": "Are these Data sets of Known fake faces or all faces to teach a machine to figure out which ones are fake?", "793166": "I actually used them in a different project. I just wanted to thank you for this!", "742203": "These images are generated by GAN2?", "715997": "You are much appreciated for uploading those fake faces!!", "707939": "after peeking several images, oh boy, I can't tell why the images are fake", "696985": "Hi! Just saw in the competition rules that there should not be external data >= 1GB, I take that this does not include the auxiliary data we use to train our model?", "694503": "How would I download a subset of these? For example, the subdir 1m_faces_00/?", "741346": "", "705406": "", "698860": "", "696751": "", "696047": "", "693987": "", "692767": "Awesome thanks!", "700554": "Thanks for Sharing!!!", "698008": "Nice ,Thanks for sharing", "697908": "Thanks!", "697728": "thanks for your effort Bojan", "697549": "Thanks!", "697383": "Thanks much!", "697321": "Thanks!", "696828": "Thanks much!", "696630": "Thanks", "696524": "Thanks", "696325": "Thanks!", "696177": "Thanks!", "696144": "Thanks, It's helpful for me.", "696137": "Thanks", "695921": "Nice! Thanks.", "695186": "Thanks for sharingπ ", "695059": "Thanks for Sharing", "694529": "Thanks for sharingπ ", "694460": "Thanks for sharing", "693445": "Thanks for sharing!", "693286": "Thanks for sharing", "693158": "Thanks for sharing", "692868": "Thanks for sharing!", "700288": "Thanks for sharing !", "700002": "Thanks for sharing !", "939958": "Thanks! Very useful dataset.", "713553": "Thanks for sharing Bojan ! :)", "707937": "Thanks!", "703503": "Thanks for sharing.", "754073": "Thanks for sharing!\n" } }