Please use this thread to list the pre-trained models you are planning to use, or other external data. Once a model or data source has been posted once, it does not need to be posted again. You must post here one week prior to competition close.
", "messages": [ { "id": 434341, "postDate": "2018-12-06T08:18:26.187Z", "content": "If I go out on a whale watching boat and take pictures of whale tails, would I need to post images I use here? Seriously, I live close enough and have been on whale watching tours here before.
", "rawMarkdown": "If I go out on a whale watching boat and take pictures of whale tails, would I need to post images I use here? Seriously, I live close enough and have been on whale watching tours here before. ", "votes": 11, "replies": [ { "id": 434564, "postDate": "2018-12-06T15:43:42.003Z", "content": "Hi Brian. That is what this is all ultimately for -- those photos of whales are valuable data. And to make it more fun and meaningful, we will notify you through Happywhale.com (which is where you'd submit images) of what we find -- if we can identify an individual, where the whale has traveled, and we'll notify you when we see your whale again :)
", "rawMarkdown": "Hi Brian. That is what this is all ultimately for -- those photos of whales are valuable data. And to make it more fun and meaningful, we will notify you through Happywhale.com (which is where you'd submit images) of what we find -- if we can identify an individual, where the whale has traveled, and we'll notify you when we see your whale again :)" }, { "id": 434846, "postDate": "2018-12-07T03:15:59.367Z", "content": "Sounds like I should sign up for another trip out into the bay :)
", "rawMarkdown": "Sounds like I should sign up for another trip out into the bay :)", "votes": 2 }, { "id": 444489, "postDate": "2018-12-24T05:39:05.880Z", "content": "Follow up question, what is a good charter in Monterrey Bay to see the whales? Last time I went out was years ago in November on a Chardonnay pizza and beer charter. They were great but I can't really do pizza anymore
", "rawMarkdown": "Follow up question, what is a good charter in Monterrey Bay to see the whales? Last time I went out was years ago in November on a Chardonnay pizza and beer charter. They were great but I can't really do pizza anymore" }, { "id": 444604, "postDate": "2018-12-24T11:06:04.537Z", "content": "Chardonnay is a nice time but not the way to see whales. The best operations are, in my opinion, in Moss Landing, Blue Ocean Whale Watch and Fast Raft Ocean Safaris, and out of Fisherman's Wharf in Monterey, Monterey Bay Whale Watch and Discovery Whale Watch
", "rawMarkdown": "Chardonnay is a nice time but not the way to see whales. The best operations are, in my opinion, in Moss Landing, Blue Ocean Whale Watch and Fast Raft Ocean Safaris, and out of Fisherman's Wharf in Monterey, Monterey Bay Whale Watch and Discovery Whale Watch", "votes": 3 }, { "id": 453386, "postDate": "2019-01-10T05:26:57.620Z", "content": "I also was going to watch whales (if I am lucky). I live in Ireland, close to the sea. But I think it's not going to be the same region as in challenge...
", "rawMarkdown": "I also was going to watch whales (if I am lucky). I live in Ireland, close to the sea. But I think it's not going to be the same region as in challenge...", "votes": 1 } ] }, { "id": 430675, "postDate": "2018-11-30T20:38:41.190Z", "content": "Please use this thread to list the pre-trained models you are planning to use, or other external data. Once a model or data source has been posted once, it does not need to be posted again. You must post here one week prior to competition close.
", "rawMarkdown": "Please use this thread to list the pre-trained models you are planning to use, or other external data. Once a model or data source has been posted once, it does not need to be posted again. You must post here one week prior to competition close.", "votes": 8 }, { "id": 480070, "postDate": "2019-02-27T18:46:17.153Z", "content": "LeNet pretrained on MNIST
", "rawMarkdown": "LeNet pretrained on MNIST", "votes": 6, "replies": [ { "id": 480077, "postDate": "2019-02-27T18:52:52.613Z", "content": "Pulled out the big guns I see...
", "rawMarkdown": "Pulled out the big guns I see...", "votes": 3 } ] }, { "id": 437932, "postDate": "2018-12-12T19:22:49.260Z", "content": "Keras Resnet(18,34,50,101,152) and ResNext(50,101) ImageNet pretrained models https://github.com/qubvel/classification_models
", "rawMarkdown": "Keras Resnet(18,34,50,101,152) and ResNext(50,101) ImageNet pretrained models https://github.com/qubvel/classification_models ", "votes": 3 }, { "id": 431039, "postDate": "2018-12-01T13:56:06.517Z", "content": "All pretrained models for pytorch from torchvision package.
", "rawMarkdown": "All pretrained models for pytorch from torchvision package.", "votes": 3 }, { "id": 430879, "postDate": "2018-12-01T05:12:13.023Z", "content": "I plan to use Fastai and retrained model supported by fastai
", "rawMarkdown": "I plan to use Fastai and retrained model supported by fastai", "votes": 4 }, { "id": 476237, "postDate": "2019-02-21T20:52:45.750Z", "content": "pretrainedmodels for pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html
\n\nI may use the model here.（update,can not download the pretrained files, will not use it)\nhttps://github.com/wielandbrendel/bag-of-local-features-models
", "rawMarkdown": "pretrainedmodels for pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\n\nI may use the model here.（update,can not download the pretrained files, will not use it)\nhttps://github.com/wielandbrendel/bag-of-local-features-models", "votes": 1 }, { "id": 475961, "postDate": "2019-02-21T12:43:21.183Z", "content": "pretrainedmodels for pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html
\n\nand pretrainedmodels for keras\nhttps://keras.io/applications/\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/78078
", "rawMarkdown": "pretrainedmodels for pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\n\nand pretrainedmodels for keras\nhttps://keras.io/applications/\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/78078", "votes": 1 }, { "id": 451098, "postDate": "2019-01-06T12:22:15.637Z", "content": "ResNet-50 pretrained from ImageNet
", "rawMarkdown": "ResNet-50 pretrained from ImageNet", "votes": 1 }, { "id": 431028, "postDate": "2018-12-01T13:25:12.860Z", "content": "I use pytorch pretrained models. https://www.kaggle.com/iafoss/pytorch-pretrained-models
", "rawMarkdown": "I use pytorch pretrained models. https://www.kaggle.com/iafoss/pytorch-pretrained-models", "votes": 2 }, { "id": 432198, "postDate": "2018-12-03T14:48:02.860Z", "content": "Why? Why shall we reveal our models here?\nis this a serious post or its just a fun one?
", "rawMarkdown": "Why? Why shall we reveal our models here?\nis this a serious post or its just a fun one?", "votes": -1, "replies": [ { "id": 432249, "postDate": "2018-12-03T15:59:27.373Z", "content": "It is part of the rules of the competition. If you use a pretrained model you have to disclose it.
\n\nEXTERNAL DATA \nThe following provision supersedes General Rules Section 7.C. below: “You may use data, other than the Competition Data, as allowed on the Competition Website to develop and test your models and Submissions; provided, you have the right and authority to use such external data for the purposes of the Competition, and to share such data with Sponsor and Kaggle as may be required.\"\nPublicly, freely available external data is permitted. The source of any external data must be posted to the official competition forum prior to the Entry Deadline. Entrants may re-annotate images in the training set, but may not hand-label predictions, including having human observers rate and evaluate the test data set.
", "rawMarkdown": "It is part of the rules of the competition. If you use a pretrained model you have to disclose it.\n\nEXTERNAL DATA \nThe following provision supersedes General Rules Section 7.C. below: “You may use data, other than the Competition Data, as allowed on the Competition Website to develop and test your models and Submissions; provided, you have the right and authority to use such external data for the purposes of the Competition, and to share such data with Sponsor and Kaggle as may be required.\"\nPublicly, freely available external data is permitted. The source of any external data must be posted to the official competition forum prior to the Entry Deadline. Entrants may re-annotate images in the training set, but may not hand-label predictions, including having human observers rate and evaluate the test data set." }, { "id": 432262, "postDate": "2018-12-03T16:21:00.607Z", "content": "it is about external data!\nnot model!
", "rawMarkdown": "it is about external data!\nnot model! " }, { "id": 432279, "postDate": "2018-12-03T16:54:21.607Z", "content": "Yes, it seems the rules do not state that you need to reveal your model, just the sources of publicly available data (that is the only type of data we can train on). I plan to gather additional data from search engines for training, validation, or both. I may also use part of the Cascadia Research Collective dataset.
", "rawMarkdown": "Yes, it seems the rules do not state that you need to reveal your model, just the sources of publicly available data (that is the only type of data we can train on). I plan to gather additional data from search engines for training, validation, or both. I may also use part of the Cascadia Research Collective dataset.", "votes": 1 }, { "id": 432328, "postDate": "2018-12-03T18:17:49.693Z", "content": "If somebody else already posted the same pre-trained model to the forum for this contest, am I supposed to post it again, saying I am using it also?
", "rawMarkdown": "If somebody else already posted the same pre-trained model to the forum for this contest, am I supposed to post it again, saying I am using it also?" }, { "id": 432334, "postDate": "2018-12-03T18:25:50.867Z", "content": "Once disclosed, you do not need to disclose again. The spirit of revealing any external data or pre-trained models is to keep the competition a Machine Learning competition, not a data collection competition :)
", "rawMarkdown": "Once disclosed, you do not need to disclose again. The spirit of revealing any external data or pre-trained models is to keep the competition a Machine Learning competition, not a data collection competition :)", "votes": 5 }, { "id": 434566, "postDate": "2018-12-06T15:45:48.247Z", "content": "@Simeon Trieu, you are unlikely to find much additional data of substantial value. We've sought to include as large a dataset as we could gather without adding abundant duplicates, from a dataset with broad collaboration around the world :)
", "rawMarkdown": "@Simeon Trieu, you are unlikely to find much additional data of substantial value. We've sought to include as large a dataset as we could gather without adding abundant duplicates, from a dataset with broad collaboration around the world :)", "votes": 2 }, { "id": 434633, "postDate": "2018-12-06T17:48:44.803Z", "content": "Thanks, I haven't had much luck contacting the Cascadia researchers anyways. I'll make do with what I have, and thanks for preparing the dataset.
", "rawMarkdown": "Thanks, I haven't had much luck contacting the Cascadia researchers anyways. I'll make do with what I have, and thanks for preparing the dataset. " } ] }, { "id": 480943, "postDate": "2019-02-28T22:43:06.567Z", "content": "Keras pretrained models: \nhttps://keras.io/applications/
\nhttps://github.com/qubvel/classification_models
Pre-trained models for pytorch and keras.
", "rawMarkdown": "Pre-trained models for pytorch and keras. " }, { "id": 477601, "postDate": "2019-02-25T00:06:05.607Z", "content": "I'm also using all pre-trained models for pytorch from torchvision package.
", "rawMarkdown": "I'm also using all pre-trained models for pytorch from torchvision package." }, { "id": 477591, "postDate": "2019-02-24T23:21:05.480Z", "content": "From Martin's work.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output
", "rawMarkdown": "From Martin's work.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output" }, { "id": 476519, "postDate": "2019-02-22T09:28:48.160Z", "content": "Pre-trained models for pytorch\nResNet-34, ResNet-50
", "rawMarkdown": "Pre-trained models for pytorch\nResNet-34, ResNet-50 " }, { "id": 476085, "postDate": "2019-02-21T15:40:08.273Z", "content": "keras pretrained models
", "rawMarkdown": "keras pretrained models" }, { "id": 475862, "postDate": "2019-02-21T09:48:59.980Z", "content": "https://github.com/Cadene/pretrained-models.pytorch
", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch" }, { "id": 475859, "postDate": "2019-02-21T09:41:12.403Z", "content": "pretrain model
\n\nhttps://github.com/Cadene/pretrained-models.pytorch
\n\nhttps://github.com/fwang91/residual-attention-network
", "rawMarkdown": "pretrain model\n\n\nhttps://github.com/Cadene/pretrained-models.pytorch\n\nhttps://github.com/fwang91/residual-attention-network\n" }, { "id": 475836, "postDate": "2019-02-21T08:59:35.530Z", "content": "Keras pre trained Available models
\n\nModels for image classification with weights trained on ImageNet:\nXception\nVGG16\nVGG19\nResNet, ResNetV2, ResNeXt\nInceptionV3\nInceptionResNetV2\nMobileNet\nMobileNetV2\nDenseNet\nNASNet
", "rawMarkdown": "Keras pre trained Available models\n\nModels for image classification with weights trained on ImageNet:\nXception\nVGG16\nVGG19\nResNet, ResNetV2, ResNeXt\nInceptionV3\nInceptionResNetV2\nMobileNet\nMobileNetV2\nDenseNet\nNASNet" }, { "id": 475524, "postDate": "2019-02-20T21:06:54.113Z", "content": "https://github.com/osmr/imgclsmob/tree/master/pytorch
", "rawMarkdown": "https://github.com/osmr/imgclsmob/tree/master/pytorch" }, { "id": 474030, "postDate": "2019-02-18T20:44:42.673Z", "content": "https://www.syntheticwhales.nl/s/ZXgQcZKCCks50wH
", "rawMarkdown": "https://www.syntheticwhales.nl/s/ZXgQcZKCCks50wH" }, { "id": 473985, "postDate": "2019-02-18T19:43:52.870Z", "content": "keras pretrained weights\nhttps://keras.io/applications/\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/78078
", "rawMarkdown": "keras pretrained weights\nhttps://keras.io/applications/\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/78078" }, { "id": 472597, "postDate": "2019-02-16T08:52:21.333Z", "content": "VGG 16 , weights = imageNet
", "rawMarkdown": "VGG 16 , weights = imageNet" }, { "id": 472110, "postDate": "2019-02-15T11:19:48.517Z", "content": "Resnet152, Densenet161
", "rawMarkdown": "Resnet152, Densenet161" }, { "id": 472002, "postDate": "2019-02-15T07:47:59.717Z", "content": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html
\n\nPretrained Keras-RetinaNet models:\nhttps://github.com/ZFTurbo/Keras-RetinaNet-for-Open-Images-Challenge-2018
\n\npretrained models from https://github.com/tensorflow/models and tensorflow slim
\n\nhttps://keras.io/applications/#mobilenet
\n\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output
\n\n", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\n\nPretrained Keras-RetinaNet models:\nhttps://github.com/ZFTurbo/Keras-RetinaNet-for-Open-Images-Challenge-2018\n\npretrained models from https://github.com/tensorflow/models and tensorflow slim\n\nhttps://keras.io/applications/#mobilenet\n\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output\n\nhttp://www.image-net.org" }, { "id": 471256, "postDate": "2019-02-14T07:55:05.213Z", "content": "Pretrained Keras-RetinaNet models:\nhttps://github.com/ZFTurbo/Keras-RetinaNet-for-Open-Images-Challenge-2018
", "rawMarkdown": "Pretrained Keras-RetinaNet models:\nhttps://github.com/ZFTurbo/Keras-RetinaNet-for-Open-Images-Challenge-2018" }, { "id": 469880, "postDate": "2019-02-12T01:07:51.467Z", "content": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html
", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html", "replies": [ { "id": 475418, "postDate": "2019-02-20T18:21:58.837Z", "content": "pretrainedmodels for pytorch
\n\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html
", "rawMarkdown": "pretrainedmodels for pytorch\n\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html" } ] }, { "id": 469210, "postDate": "2019-02-10T18:00:42.933Z", "content": "Can we use images and labels from happywhale.com?\nIs it legal to manually label the playground testset and propagate the labels to this competitions testset?
", "rawMarkdown": "Can we use images and labels from happywhale.com?\nIs it legal to manually label the playground testset and propagate the labels to this competitions testset?", "replies": [ { "id": 469588, "postDate": "2019-02-11T13:50:07.957Z", "content": "no, that will violate your entry. Good luck!
", "rawMarkdown": "no, that will violate your entry. Good luck!" }, { "id": 469701, "postDate": "2019-02-11T17:33:15.337Z", "content": "Hi Ted,\nTo clarify, you mean we cannot use images download from happywhale.com?
", "rawMarkdown": "Hi Ted,\nTo clarify, you mean we cannot use images download from happywhale.com?" }, { "id": 469706, "postDate": "2019-02-11T17:37:39.337Z", "content": "correct. manual labeling from other data sources invalidates entry. And none of the test images are on happywhale.com. Success is earned by algorithm performance! And I wish you all success :)
", "rawMarkdown": "correct. manual labeling from other data sources invalidates entry. And none of the test images are on happywhale.com. Success is earned by algorithm performance! And I wish you all success :)", "votes": 4 }, { "id": 469710, "postDate": "2019-02-11T17:54:36.720Z", "content": "Thanks
", "rawMarkdown": "Thanks" }, { "id": 470408, "postDate": "2019-02-12T22:35:20.007Z", "rawMarkdown": "", "isDeleted": true } ] }, { "id": 468912, "postDate": "2019-02-10T02:42:55.227Z", "content": "ResNext101 on http://www.image-net.org
", "rawMarkdown": "ResNext101 on http://www.image-net.org" }, { "id": 467732, "postDate": "2019-02-07T16:22:52.893Z", "content": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html
", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html" }, { "id": 467490, "postDate": "2019-02-07T08:09:54.273Z", "content": "Image recognition in R using convolutional neural networks with the MXNet package. https://firsttimeprogrammer.blogspot.com/2016/07/image-recognition-in-r-using.html
", "rawMarkdown": "Image recognition in R using convolutional neural networks with the MXNet package. https://firsttimeprogrammer.blogspot.com/2016/07/image-recognition-in-r-using.html" }, { "id": 467366, "postDate": "2019-02-07T00:29:24.873Z", "content": "My extern data and pretrained model are from https://www.kaggle.com/seesee/siamese-pretrained-0-822.
", "rawMarkdown": "My extern data and pretrained model are from https://www.kaggle.com/seesee/siamese-pretrained-0-822." }, { "id": 466176, "postDate": "2019-02-04T19:59:46.850Z", "content": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output
", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output" }, { "id": 464513, "postDate": "2019-02-01T02:40:54.317Z", "content": "Darknet53 ImageNet pretrained model: http://pjreddie.com/media/files/darknet53.conv.74
", "rawMarkdown": "Darknet53 ImageNet pretrained model: http://pjreddie.com/media/files/darknet53.conv.74" }, { "id": 464132, "postDate": "2019-01-31T08:37:13.357Z", "content": "I intend to use Keras Resnet(18,34,50,101,152) (trained on ImageNet) pretrained models\nhttps://github.com/qubvel/classification_models
", "rawMarkdown": "I intend to use Keras Resnet(18,34,50,101,152) (trained on ImageNet) pretrained models\nhttps://github.com/qubvel/classification_models" }, { "id": 463971, "postDate": "2019-01-31T01:25:51.380Z", "content": "https://github.com/Cadene/pretrained-models.pytorch
", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch" }, { "id": 462354, "postDate": "2019-01-28T07:11:38.297Z", "content": "", "rawMarkdown": "http://www.image-net.org" }, { "id": 461104, "postDate": "2019-01-25T08:19:19.717Z", "content": "From Martin's work.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output
", "rawMarkdown": "From Martin's work.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output" }, { "id": 461017, "postDate": "2019-01-25T02:17:39.840Z", "content": "All pretrained models from torchvision package.
", "rawMarkdown": "All pretrained models from torchvision package." }, { "id": 459743, "postDate": "2019-01-22T09:27:10.860Z", "content": "I would like to use Martin's bounding-box model.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output
", "rawMarkdown": "I would like to use Martin's bounding-box model.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output" }, { "id": 458190, "postDate": "2019-01-19T02:54:46.100Z", "content": "PyTorch pre-trained ResNet-50, so far. Although currently I haven't submit my predictions.
", "rawMarkdown": "PyTorch pre-trained ResNet-50, so far. Although currently I haven't submit my predictions." }, { "id": 457893, "postDate": "2019-01-18T09:34:05.410Z", "content": "I plan to finetune pretrained models from keras applications.
", "rawMarkdown": "I plan to finetune pretrained models from [keras applications](https://keras.io/applications/)." }, { "id": 451446, "postDate": "2019-01-07T05:49:06.803Z", "content": "using https://github.com/Cadene/pretrained-models.pytorch
", "rawMarkdown": "using https://github.com/Cadene/pretrained-models.pytorch" }, { "id": 451349, "postDate": "2019-01-06T23:07:18.237Z", "content": "Resnet-50, trained on imagenet
", "rawMarkdown": "Resnet-50, trained on imagenet" }, { "id": 447712, "postDate": "2018-12-30T11:42:55.353Z", "content": "https://github.com/DagnyT/hardnet\nhttps://github.com/ducha-aiki/affnet\nhttp://cmp.felk.cvut.cz/cnnimageretrieval/
", "rawMarkdown": "https://github.com/DagnyT/hardnet\nhttps://github.com/ducha-aiki/affnet\nhttp://cmp.felk.cvut.cz/cnnimageretrieval/" }, { "id": 446978, "postDate": "2018-12-29T00:01:54.913Z", "content": "Xception V1 model pretrained on imagenet dataset
", "rawMarkdown": "Xception V1 model pretrained on imagenet dataset" }, { "id": 441358, "postDate": "2018-12-18T15:27:37.550Z", "content": "https://github.com/qubvel/classification_models\nAll models with imagenet pretrained weights
", "rawMarkdown": "https://github.com/qubvel/classification_models\nAll models with imagenet pretrained weights" }, { "id": 441311, "postDate": "2018-12-18T14:33:50.023Z", "content": "All pre trained networks from Keras\nhttps://keras.io/applications/
\n\nData from previous competition
", "rawMarkdown": "All pre trained networks from Keras\nhttps://keras.io/applications/\n\nData from previous competition" }, { "id": 439827, "postDate": "2018-12-16T13:32:29.750Z", "content": "for now, MobilenetV2, 1.0/1.4 pretrained on ImageNet, ckpts from https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet
", "rawMarkdown": "for now, MobilenetV2, 1.0/1.4 pretrained on ImageNet, ckpts from https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet" }, { "id": 439408, "postDate": "2018-12-15T11:33:41.657Z", "content": "https://github.com/Cadene/pretrained-models.pytorch
", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch" }, { "id": 437654, "postDate": "2018-12-12T09:28:04.500Z", "content": "I am going to use resnet50 as my initial submission. Not sure if other architectures perform better.
", "rawMarkdown": "I am going to use resnet50 as my initial submission. Not sure if other architectures perform better. " }, { "id": 436679, "postDate": "2018-12-10T18:22:41.987Z", "content": "I have used ResNet50, ResNext50 and ResNet101 for my solutions
", "rawMarkdown": "I have used ResNet50, ResNext50 and ResNet101 for my solutions" }, { "id": 435671, "postDate": "2018-12-08T14:07:04.687Z", "content": "I would like to use Martin's bounding-box model.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output
", "rawMarkdown": "I would like to use Martin's bounding-box model.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output" }, { "id": 434916, "postDate": "2018-12-07T06:39:49.417Z", "content": "Does the dataset from previous whale identification challenge count as external data as well?
", "rawMarkdown": "Does the dataset from previous whale identification challenge count as external data as well?", "replies": [ { "id": 435194, "postDate": "2018-12-07T16:42:29.473Z", "content": "Same question - I am using data from the previous challenge as well.
", "rawMarkdown": "Same question - I am using data from the previous challenge as well." }, { "id": 435920, "postDate": "2018-12-09T04:49:13.850Z", "content": "Both of train and test datasets are same as playground competition.\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/73257
", "rawMarkdown": "Both of train and test datasets are same as playground competition.\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/73257", "votes": 2 }, { "id": 436745, "postDate": "2018-12-10T21:34:57.867Z", "content": "Sorry, I misunderstood. There is another dataset in playground competition, and we can download from side-bar. I would like to try them too.\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/73257
", "rawMarkdown": "Sorry, I misunderstood. There is another dataset in playground competition, and we can download from side-bar. I would like to try them too.\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/73257", "votes": 1 } ] }, { "id": 434772, "postDate": "2018-12-06T23:08:18.127Z", "content": "keras pretrained models
", "rawMarkdown": "keras pretrained models" }, { "id": 433611, "postDate": "2018-12-05T08:17:12.853Z", "content": "ResNet-50 pretrained from ImageNet
", "rawMarkdown": "ResNet-50 pretrained from ImageNet" }, { "id": 432460, "postDate": "2018-12-03T22:27:11.240Z", "content": "Pretrained models from https://github.com/Cadene/pretrained-models.pytorch,\npretrained models from https://github.com/tensorflow/models and tensorflow slim
", "rawMarkdown": "Pretrained models from https://github.com/Cadene/pretrained-models.pytorch,\npretrained models from https://github.com/tensorflow/models and tensorflow slim" }, { "id": 432391, "postDate": "2018-12-03T19:47:01.483Z", "content": "I will be using MobileNet with weights from imagenet if possible: https://keras.io/applications/#mobilenet
", "rawMarkdown": "I will be using MobileNet with weights from imagenet if possible: https://keras.io/applications/#mobilenet" }, { "id": 477255, "postDate": "2019-02-24T08:17:52.153Z", "rawMarkdown": "", "isDeleted": true }, { "id": 475388, "postDate": "2019-02-20T17:38:47.673Z", "rawMarkdown": "", "isDeleted": true }, { "id": 468868, "postDate": "2019-02-09T22:32:46.563Z", "rawMarkdown": "", "isDeleted": true }, { "id": 460731, "postDate": "2019-01-24T10:09:56.927Z", "rawMarkdown": "", "isDeleted": true }, { "id": 453580, "postDate": "2019-01-10T12:24:35.460Z", "rawMarkdown": "", "isDeleted": true }, { "id": 440397, "postDate": "2018-12-17T13:55:47.193Z", "rawMarkdown": "", "votes": 2, "isDeleted": true }, { "id": 2915628, "postDate": "2024-07-10T15:35:33.073Z", "content": "yes! thank you!
", "rawMarkdown": "yes! thank you!" } ], "comments": [ { "id": 434341, "author_name": "Brian", "author_url": "", "post_date": "2018-12-06T08:18:26.187000", "content": "If I go out on a whale watching boat and take pictures of whale tails, would I need to post images I use here? Seriously, I live close enough and have been on whale watching tours here before.
", "votes": 11, "replies": [ { "id": 434564, "author_name": "Ted Cheeseman", "author_url": "", "post_date": "2018-12-06T15:43:42.003000", "content": "Hi Brian. That is what this is all ultimately for -- those photos of whales are valuable data. And to make it more fun and meaningful, we will notify you through Happywhale.com (which is where you'd submit images) of what we find -- if we can identify an individual, where the whale has traveled, and we'll notify you when we see your whale again :)
", "votes": 0, "replies": [] }, { "id": 434846, "author_name": "Brian", "author_url": "", "post_date": "2018-12-07T03:15:59.367000", "content": "Sounds like I should sign up for another trip out into the bay :)
", "votes": 2, "replies": [] }, { "id": 444489, "author_name": "Brian", "author_url": "", "post_date": "2018-12-24T05:39:05.880000", "content": "Follow up question, what is a good charter in Monterrey Bay to see the whales? Last time I went out was years ago in November on a Chardonnay pizza and beer charter. They were great but I can't really do pizza anymore
", "votes": 0, "replies": [] }, { "id": 444604, "author_name": "Ted Cheeseman", "author_url": "", "post_date": "2018-12-24T11:06:04.537000", "content": "Chardonnay is a nice time but not the way to see whales. The best operations are, in my opinion, in Moss Landing, Blue Ocean Whale Watch and Fast Raft Ocean Safaris, and out of Fisherman's Wharf in Monterey, Monterey Bay Whale Watch and Discovery Whale Watch
", "votes": 3, "replies": [] }, { "id": 453386, "author_name": "Blonde", "author_url": "", "post_date": "2019-01-10T05:26:57.620000", "content": "I also was going to watch whales (if I am lucky). I live in Ireland, close to the sea. But I think it's not going to be the same region as in challenge...
", "votes": 1, "replies": [] } ] }, { "id": 480070, "author_name": "Artem.Sanakoev", "author_url": "", "post_date": "2019-02-27T18:46:17.153000", "content": "LeNet pretrained on MNIST
", "votes": 6, "replies": [ { "id": 480077, "author_name": "interneuron", "author_url": "", "post_date": "2019-02-27T18:52:52.613000", "content": "Pulled out the big guns I see...
", "votes": 3, "replies": [] } ] }, { "id": 437932, "author_name": "Pavel Iakubovskii (qubvel)", "author_url": "", "post_date": "2018-12-12T19:22:49.260000", "content": "Keras Resnet(18,34,50,101,152) and ResNext(50,101) ImageNet pretrained models https://github.com/qubvel/classification_models
", "votes": 3, "replies": [] }, { "id": 431039, "author_name": "Andrey Lukyanenko", "author_url": "", "post_date": "2018-12-01T13:56:06.517000", "content": "All pretrained models for pytorch from torchvision package.
", "votes": 3, "replies": [] }, { "id": 430879, "author_name": "Amil Gentili", "author_url": "", "post_date": "2018-12-01T05:12:13.023000", "content": "I plan to use Fastai and retrained model supported by fastai
", "votes": 4, "replies": [] }, { "id": 476237, "author_name": "bestfitting", "author_url": "", "post_date": "2019-02-21T20:52:45.750000", "content": "pretrainedmodels for pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html
\n\nI may use the model here.（update,can not download the pretrained files, will not use it)\nhttps://github.com/wielandbrendel/bag-of-local-features-models
", "votes": 1, "replies": [] }, { "id": 475961, "author_name": "Niclas Stoltenberg", "author_url": "", "post_date": "2019-02-21T12:43:21.183000", "content": "pretrainedmodels for pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html
\n\nand pretrainedmodels for keras\nhttps://keras.io/applications/\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/78078
", "votes": 1, "replies": [] }, { "id": 451098, "author_name": "Qidian213", "author_url": "", "post_date": "2019-01-06T12:22:15.637000", "content": "ResNet-50 pretrained from ImageNet
", "votes": 1, "replies": [] }, { "id": 431028, "author_name": "ManishYathnalli", "author_url": "", "post_date": "2018-12-01T13:25:12.860000", "content": "I use pytorch pretrained models. https://www.kaggle.com/iafoss/pytorch-pretrained-models
", "votes": 2, "replies": [] }, { "id": 432198, "author_name": "Omid Safarzadeh", "author_url": "", "post_date": "2018-12-03T14:48:02.860000", "content": "Why? Why shall we reveal our models here?\nis this a serious post or its just a fun one?
", "votes": -1, "replies": [ { "id": 432249, "author_name": "Amil Gentili", "author_url": "", "post_date": "2018-12-03T15:59:27.373000", "content": "It is part of the rules of the competition. If you use a pretrained model you have to disclose it.
\n\nEXTERNAL DATA \nThe following provision supersedes General Rules Section 7.C. below: “You may use data, other than the Competition Data, as allowed on the Competition Website to develop and test your models and Submissions; provided, you have the right and authority to use such external data for the purposes of the Competition, and to share such data with Sponsor and Kaggle as may be required.\"\nPublicly, freely available external data is permitted. The source of any external data must be posted to the official competition forum prior to the Entry Deadline. Entrants may re-annotate images in the training set, but may not hand-label predictions, including having human observers rate and evaluate the test data set.
", "votes": 0, "replies": [] }, { "id": 432262, "author_name": "Omid Safarzadeh", "author_url": "", "post_date": "2018-12-03T16:21:00.607000", "content": "it is about external data!\nnot model!
", "votes": 0, "replies": [] }, { "id": 432279, "author_name": "Simeon Trieu", "author_url": "", "post_date": "2018-12-03T16:54:21.607000", "content": "Yes, it seems the rules do not state that you need to reveal your model, just the sources of publicly available data (that is the only type of data we can train on). I plan to gather additional data from search engines for training, validation, or both. I may also use part of the Cascadia Research Collective dataset.
", "votes": 1, "replies": [] }, { "id": 432328, "author_name": "impulsecorp", "author_url": "", "post_date": "2018-12-03T18:17:49.693000", "content": "If somebody else already posted the same pre-trained model to the forum for this contest, am I supposed to post it again, saying I am using it also?
", "votes": 0, "replies": [] }, { "id": 432334, "author_name": "Addison Howard", "author_url": "", "post_date": "2018-12-03T18:25:50.867000", "content": "Once disclosed, you do not need to disclose again. The spirit of revealing any external data or pre-trained models is to keep the competition a Machine Learning competition, not a data collection competition :)
", "votes": 5, "replies": [] }, { "id": 434566, "author_name": "Ted Cheeseman", "author_url": "", "post_date": "2018-12-06T15:45:48.247000", "content": "@Simeon Trieu, you are unlikely to find much additional data of substantial value. We've sought to include as large a dataset as we could gather without adding abundant duplicates, from a dataset with broad collaboration around the world :)
", "votes": 2, "replies": [] }, { "id": 434633, "author_name": "Simeon Trieu", "author_url": "", "post_date": "2018-12-06T17:48:44.803000", "content": "Thanks, I haven't had much luck contacting the Cascadia researchers anyways. I'll make do with what I have, and thanks for preparing the dataset.
", "votes": 0, "replies": [] } ] }, { "id": 480943, "author_name": "bluetrain", "author_url": "", "post_date": "2019-02-28T22:43:06.567000", "content": "Keras pretrained models: \nhttps://keras.io/applications/
\nhttps://github.com/qubvel/classification_models
Pre-trained models for pytorch and keras.
", "votes": 0, "replies": [] }, { "id": 477601, "author_name": "daisukelab", "author_url": "", "post_date": "2019-02-25T00:06:05.607000", "content": "I'm also using all pre-trained models for pytorch from torchvision package.
", "votes": 0, "replies": [] }, { "id": 477591, "author_name": "Zhenye Na", "author_url": "", "post_date": "2019-02-24T23:21:05.480000", "content": "From Martin's work.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output
", "votes": 0, "replies": [] }, { "id": 476519, "author_name": "Amit Singh", "author_url": "", "post_date": "2019-02-22T09:28:48.160000", "content": "Pre-trained models for pytorch\nResNet-34, ResNet-50
", "votes": 0, "replies": [] }, { "id": 476085, "author_name": "Igor Praznik", "author_url": "", "post_date": "2019-02-21T15:40:08.273000", "content": "keras pretrained models
", "votes": 0, "replies": [] }, { "id": 475862, "author_name": "earhian", "author_url": "", "post_date": "2019-02-21T09:48:59.980000", "content": "https://github.com/Cadene/pretrained-models.pytorch
", "votes": 0, "replies": [] }, { "id": 475859, "author_name": "hengck23", "author_url": "", "post_date": "2019-02-21T09:41:12.403000", "content": "pretrain model
\n\nhttps://github.com/Cadene/pretrained-models.pytorch
\n\nhttps://github.com/fwang91/residual-attention-network
", "votes": 0, "replies": [] }, { "id": 475836, "author_name": "Lucky Singh", "author_url": "", "post_date": "2019-02-21T08:59:35.530000", "content": "Keras pre trained Available models
\n\nModels for image classification with weights trained on ImageNet:\nXception\nVGG16\nVGG19\nResNet, ResNetV2, ResNeXt\nInceptionV3\nInceptionResNetV2\nMobileNet\nMobileNetV2\nDenseNet\nNASNet
", "votes": 0, "replies": [] }, { "id": 475524, "author_name": "Artyom Palvelev", "author_url": "", "post_date": "2019-02-20T21:06:54.113000", "content": "https://github.com/osmr/imgclsmob/tree/master/pytorch
", "votes": 0, "replies": [] }, { "id": 474030, "author_name": "dromosys", "author_url": "", "post_date": "2019-02-18T20:44:42.673000", "content": "https://www.syntheticwhales.nl/s/ZXgQcZKCCks50wH
", "votes": 0, "replies": [] }, { "id": 473985, "author_name": "Guanshuo Xu", "author_url": "", "post_date": "2019-02-18T19:43:52.870000", "content": "keras pretrained weights\nhttps://keras.io/applications/\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/78078
", "votes": 0, "replies": [] }, { "id": 472597, "author_name": "Prince Gupta", "author_url": "", "post_date": "2019-02-16T08:52:21.333000", "content": "VGG 16 , weights = imageNet
", "votes": 0, "replies": [] }, { "id": 472110, "author_name": "Artificially Intelligent", "author_url": "", "post_date": "2019-02-15T11:19:48.517000", "content": "Resnet152, Densenet161
", "votes": 0, "replies": [] }, { "id": 472002, "author_name": "YaGana Sheriff-Hussaini", "author_url": "", "post_date": "2019-02-15T07:47:59.717000", "content": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html
\n\nPretrained Keras-RetinaNet models:\nhttps://github.com/ZFTurbo/Keras-RetinaNet-for-Open-Images-Challenge-2018
\n\npretrained models from https://github.com/tensorflow/models and tensorflow slim
\n\nhttps://keras.io/applications/#mobilenet
\n\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output
\n\n", "votes": 0, "replies": [] }, { "id": 471256, "author_name": "ZFTurbo", "author_url": "", "post_date": "2019-02-14T07:55:05.213000", "content": "Pretrained Keras-RetinaNet models:\nhttps://github.com/ZFTurbo/Keras-RetinaNet-for-Open-Images-Challenge-2018
", "votes": 0, "replies": [] }, { "id": 469880, "author_name": "Ashish Lal", "author_url": "", "post_date": "2019-02-12T01:07:51.467000", "content": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html
", "votes": 0, "replies": [ { "id": 475418, "author_name": "interneuron", "author_url": "", "post_date": "2019-02-20T18:21:58.837000", "content": "pretrainedmodels for pytorch
\n\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html
", "votes": 0, "replies": [] } ] }, { "id": 469210, "author_name": "Guanshuo Xu", "author_url": "", "post_date": "2019-02-10T18:00:42.933000", "content": "Can we use images and labels from happywhale.com?\nIs it legal to manually label the playground testset and propagate the labels to this competitions testset?
", "votes": 0, "replies": [ { "id": 469588, "author_name": "Ted Cheeseman", "author_url": "", "post_date": "2019-02-11T13:50:07.957000", "content": "no, that will violate your entry. Good luck!
", "votes": 0, "replies": [] }, { "id": 469701, "author_name": "Guanshuo Xu", "author_url": "", "post_date": "2019-02-11T17:33:15.337000", "content": "Hi Ted,\nTo clarify, you mean we cannot use images download from happywhale.com?
", "votes": 0, "replies": [] }, { "id": 469706, "author_name": "Ted Cheeseman", "author_url": "", "post_date": "2019-02-11T17:37:39.337000", "content": "correct. manual labeling from other data sources invalidates entry. And none of the test images are on happywhale.com. Success is earned by algorithm performance! And I wish you all success :)
", "votes": 4, "replies": [] }, { "id": 469710, "author_name": "Guanshuo Xu", "author_url": "", "post_date": "2019-02-11T17:54:36.720000", "content": "Thanks
", "votes": 0, "replies": [] }, { "id": 470408, "author_name": "", "author_url": "", "post_date": "2019-02-12T22:35:20.007000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 468912, "author_name": "MadCoder", "author_url": "", "post_date": "2019-02-10T02:42:55.227000", "content": "ResNext101 on http://www.image-net.org
", "votes": 0, "replies": [] }, { "id": 467732, "author_name": "", "author_url": "", "post_date": "2019-02-07T16:22:52.893000", "content": "", "votes": 0, "replies": [] }, { "id": 467490, "author_name": "", "author_url": "", "post_date": "2019-02-07T08:09:54.273000", "content": "", "votes": 0, "replies": [] }, { "id": 467366, "author_name": "", "author_url": "", "post_date": "2019-02-07T00:29:24.873000", "content": "", "votes": 0, "replies": [] }, { "id": 466176, "author_name": "", "author_url": "", "post_date": "2019-02-04T19:59:46.850000", "content": "", "votes": 0, "replies": [] }, { "id": 464513, "author_name": "", "author_url": "", "post_date": "2019-02-01T02:40:54.317000", "content": "", "votes": 0, "replies": [] }, { "id": 464132, "author_name": "", "author_url": "", "post_date": "2019-01-31T08:37:13.357000", "content": "", "votes": 0, "replies": [] }, { "id": 463971, "author_name": "", "author_url": "", "post_date": "2019-01-31T01:25:51.380000", "content": "", "votes": 0, "replies": [] }, { "id": 462354, "author_name": "", "author_url": "", "post_date": "2019-01-28T07:11:38.297000", "content": "", "votes": 0, "replies": [] }, { "id": 461104, "author_name": "", "author_url": "", "post_date": "2019-01-25T08:19:19.717000", "content": "", "votes": 0, "replies": [] }, { "id": 461017, "author_name": "", "author_url": "", "post_date": "2019-01-25T02:17:39.840000", "content": "", "votes": 0, "replies": [] }, { "id": 459743, "author_name": "", "author_url": "", "post_date": "2019-01-22T09:27:10.860000", "content": "", "votes": 0, "replies": [] }, { "id": 458190, "author_name": "", "author_url": "", "post_date": "2019-01-19T02:54:46.100000", "content": "", "votes": 0, "replies": [] }, { "id": 457893, "author_name": "", "author_url": "", "post_date": "2019-01-18T09:34:05.410000", "content": "", "votes": 0, "replies": [] }, { "id": 451446, "author_name": "", "author_url": "", "post_date": "2019-01-07T05:49:06.803000", "content": "", "votes": 0, "replies": [] }, { "id": 451349, "author_name": "", "author_url": "", "post_date": "2019-01-06T23:07:18.237000", "content": "", "votes": 0, "replies": [] }, { "id": 447712, "author_name": "", "author_url": "", "post_date": "2018-12-30T11:42:55.353000", "content": "", "votes": 0, "replies": [] }, { "id": 446978, "author_name": "", "author_url": "", "post_date": "2018-12-29T00:01:54.913000", "content": "", "votes": 0, "replies": [] }, { "id": 441358, "author_name": "", "author_url": "", "post_date": "2018-12-18T15:27:37.550000", "content": "", "votes": 0, "replies": [] }, { "id": 441311, "author_name": "", "author_url": "", "post_date": "2018-12-18T14:33:50.023000", "content": "", "votes": 0, "replies": [] }, { "id": 439827, "author_name": "", "author_url": "", "post_date": "2018-12-16T13:32:29.750000", "content": "", "votes": 0, "replies": [] }, { "id": 439408, "author_name": "", "author_url": "", "post_date": "2018-12-15T11:33:41.657000", "content": "", "votes": 0, "replies": [] }, { "id": 437654, "author_name": "", "author_url": "", "post_date": "2018-12-12T09:28:04.500000", "content": "", "votes": 0, "replies": [] }, { "id": 436679, "author_name": "", "author_url": "", "post_date": "2018-12-10T18:22:41.987000", "content": "", "votes": 0, "replies": [] }, { "id": 435671, "author_name": "", "author_url": "", "post_date": "2018-12-08T14:07:04.687000", "content": "", "votes": 0, "replies": [] }, { "id": 434916, "author_name": "", "author_url": "", "post_date": "2018-12-07T06:39:49.417000", "content": "", "votes": 0, "replies": [ { "id": 435194, "author_name": "", "author_url": "", "post_date": "2018-12-07T16:42:29.473000", "content": "", "votes": 0, "replies": [] }, { "id": 435920, "author_name": "", "author_url": "", "post_date": "2018-12-09T04:49:13.850000", "content": "", "votes": 2, "replies": [] }, { "id": 436745, "author_name": "", "author_url": "", "post_date": "2018-12-10T21:34:57.867000", "content": "", "votes": 1, "replies": [] } ] }, { "id": 434772, "author_name": "", "author_url": "", "post_date": "2018-12-06T23:08:18.127000", "content": "", "votes": 0, "replies": [] }, { "id": 433611, "author_name": "", "author_url": "", "post_date": "2018-12-05T08:17:12.853000", "content": "", "votes": 0, "replies": [] }, { "id": 432460, "author_name": "", "author_url": "", "post_date": "2018-12-03T22:27:11.240000", "content": "", "votes": 0, "replies": [] }, { "id": 432391, "author_name": "", "author_url": "", "post_date": "2018-12-03T19:47:01.483000", "content": "", "votes": 0, "replies": [] }, { "id": 477255, "author_name": "", "author_url": "", "post_date": "2019-02-24T08:17:52.153000", "content": "", "votes": 0, "replies": [] }, { "id": 475388, "author_name": "", "author_url": "", "post_date": "2019-02-20T17:38:47.673000", "content": "", "votes": 0, "replies": [] }, { "id": 468868, "author_name": "", "author_url": "", "post_date": "2019-02-09T22:32:46.563000", "content": "", "votes": 0, "replies": [] }, { "id": 460731, "author_name": "", "author_url": "", "post_date": "2019-01-24T10:09:56.927000", "content": "", "votes": 0, "replies": [] }, { "id": 453580, "author_name": "", "author_url": "", "post_date": "2019-01-10T12:24:35.460000", "content": "", "votes": 0, "replies": [] }, { "id": 440397, "author_name": "", "author_url": "", "post_date": "2018-12-17T13:55:47.193000", "content": "", "votes": 2, "replies": [] }, { "id": 2915628, "author_name": "", "author_url": "", "post_date": "2024-07-10T15:35:33.073000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "434341": "If I go out on a whale watching boat and take pictures of whale tails, would I need to post images I use here? Seriously, I live close enough and have been on whale watching tours here before. ", "430675": "Please use this thread to list the pre-trained models you are planning to use, or other external data. Once a model or data source has been posted once, it does not need to be posted again. You must post here one week prior to competition close.", "480070": "LeNet pretrained on MNIST", "437932": "Keras Resnet(18,34,50,101,152) and ResNext(50,101) ImageNet pretrained models https://github.com/qubvel/classification_models ", "431039": "All pretrained models for pytorch from torchvision package.", "430879": "I plan to use Fastai and retrained model supported by fastai", "476237": "pretrainedmodels for pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\n\nI may use the model here.（update,can not download the pretrained files, will not use it)\nhttps://github.com/wielandbrendel/bag-of-local-features-models", "475961": "pretrainedmodels for pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\n\nand pretrainedmodels for keras\nhttps://keras.io/applications/\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/78078", "451098": "ResNet-50 pretrained from ImageNet", "431028": "I use pytorch pretrained models. https://www.kaggle.com/iafoss/pytorch-pretrained-models", "432198": "Why? Why shall we reveal our models here?\nis this a serious post or its just a fun one?", "480943": "Keras pretrained models: \nhttps://keras.io/applications/ \nhttps://github.com/qubvel/classification_models ", "479613": "Pre-trained models for pytorch and keras. ", "477601": "I'm also using all pre-trained models for pytorch from torchvision package.", "477591": "From Martin's work.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output", "476519": "Pre-trained models for pytorch\nResNet-34, ResNet-50 ", "476085": "keras pretrained models", "475862": "https://github.com/Cadene/pretrained-models.pytorch", "475859": "pretrain model\n\n\nhttps://github.com/Cadene/pretrained-models.pytorch\n\nhttps://github.com/fwang91/residual-attention-network\n", "475836": "Keras pre trained Available models\n\nModels for image classification with weights trained on ImageNet:\nXception\nVGG16\nVGG19\nResNet, ResNetV2, ResNeXt\nInceptionV3\nInceptionResNetV2\nMobileNet\nMobileNetV2\nDenseNet\nNASNet", "475524": "https://github.com/osmr/imgclsmob/tree/master/pytorch", "474030": "https://www.syntheticwhales.nl/s/ZXgQcZKCCks50wH", "473985": "keras pretrained weights\nhttps://keras.io/applications/\nhttps://www.kaggle.com/c/humpback-whale-identification/discussion/78078", "472597": "VGG 16 , weights = imageNet", "472110": "Resnet152, Densenet161", "472002": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html\n\nPretrained Keras-RetinaNet models:\nhttps://github.com/ZFTurbo/Keras-RetinaNet-for-Open-Images-Challenge-2018\n\npretrained models from https://github.com/tensorflow/models and tensorflow slim\n\nhttps://keras.io/applications/#mobilenet\n\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output\n\nhttp://www.image-net.org", "471256": "Pretrained Keras-RetinaNet models:\nhttps://github.com/ZFTurbo/Keras-RetinaNet-for-Open-Images-Challenge-2018", "469880": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html", "469210": "Can we use images and labels from happywhale.com?\nIs it legal to manually label the playground testset and propagate the labels to this competitions testset?", "468912": "ResNext101 on http://www.image-net.org", "467732": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://pytorch.org/docs/stable/torchvision/models.html", "467490": "Image recognition in R using convolutional neural networks with the MXNet package. https://firsttimeprogrammer.blogspot.com/2016/07/image-recognition-in-r-using.html", "467366": "My extern data and pretrained model are from https://www.kaggle.com/seesee/siamese-pretrained-0-822.", "466176": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output", "464513": "Darknet53 ImageNet pretrained model: http://pjreddie.com/media/files/darknet53.conv.74", "464132": "I intend to use Keras Resnet(18,34,50,101,152) (trained on ImageNet) pretrained models\nhttps://github.com/qubvel/classification_models", "463971": "https://github.com/Cadene/pretrained-models.pytorch", "462354": "http://www.image-net.org", "461104": "From Martin's work.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output", "461017": "All pretrained models from torchvision package.", "459743": "I would like to use Martin's bounding-box model.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output", "458190": "PyTorch pre-trained ResNet-50, so far. Although currently I haven't submit my predictions.", "457893": "I plan to finetune pretrained models from [keras applications](https://keras.io/applications/).", "451446": "using https://github.com/Cadene/pretrained-models.pytorch", "451349": "Resnet-50, trained on imagenet", "447712": "https://github.com/DagnyT/hardnet\nhttps://github.com/ducha-aiki/affnet\nhttp://cmp.felk.cvut.cz/cnnimageretrieval/", "446978": "Xception V1 model pretrained on imagenet dataset", "441358": "https://github.com/qubvel/classification_models\nAll models with imagenet pretrained weights", "441311": "All pre trained networks from Keras\nhttps://keras.io/applications/\n\nData from previous competition", "439827": "for now, MobilenetV2, 1.0/1.4 pretrained on ImageNet, ckpts from https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet", "439408": "https://github.com/Cadene/pretrained-models.pytorch", "437654": "I am going to use resnet50 as my initial submission. Not sure if other architectures perform better. ", "436679": "I have used ResNet50, ResNext50 and ResNet101 for my solutions", "435671": "I would like to use Martin's bounding-box model.\nhttps://www.kaggle.com/martinpiotte/bounding-box-model/output", "434916": "Does the dataset from previous whale identification challenge count as external data as well?", "434772": "keras pretrained models", "433611": "ResNet-50 pretrained from ImageNet", "432460": "Pretrained models from https://github.com/Cadene/pretrained-models.pytorch,\npretrained models from https://github.com/tensorflow/models and tensorflow slim", "432391": "I will be using MobileNet with weights from imagenet if possible: https://keras.io/applications/#mobilenet", "477255": "", "475388": "", "468868": "", "460731": "", "453580": "", "440397": "", "2915628": "yes! thank you!" } }