{
  "id": 87074,
  "title": "Welcome to iMet Collection 2019 - FGVC6!",
  "url": "/competitions/imet-2019-fgvc6/discussion/87074",
  "author_name": "Chenyang Zhang",
  "post_date": "2019-03-28T15:25:43.376000",
  "votes": 16,
  "comment_count": 47,
  "views": 0,
  "content": "<p>In this competition, we are recognizing multiple attributes from (and beautiful) artwork images from The Metropolitan Museum of Art.</p>\n\n<p>We want to note that this is a <strong>kernels-only competition</strong>, i.e., all submissions to this competition must be made through <strong>kernels</strong>. For fair competition, no external data should be used, but pre-trained model from public released academic datasets are allowed (this includes ImageNet, iNaturalist, etc; full citations needed.) Please specify all the external data used in your submission in the specified discussion post.</p>\n\n<p>GLHF!</p>",
  "messages": [
    {
      "id": 502446,
      "postDate": "2019-03-28T15:25:43.377Z",
      "content": "<p>In this competition, we are recognizing multiple attributes from (and beautiful) artwork images from The Metropolitan Museum of Art.</p>\n\n<p>We want to note that this is a <strong>kernels-only competition</strong>, i.e., all submissions to this competition must be made through <strong>kernels</strong>. For fair competition, no external data should be used, but pre-trained model from public released academic datasets are allowed (this includes ImageNet, iNaturalist, etc; full citations needed.) Please specify all the external data used in your submission in the specified discussion post.</p>\n\n<p>GLHF!</p>",
      "rawMarkdown": "In this competition, we are recognizing multiple attributes from (and beautiful) artwork images from The Metropolitan Museum of Art.\n\nWe want to note that this is a **kernels-only competition**, i.e., all submissions to this competition must be made through **kernels**. For fair competition, no external data should be used, but pre-trained model from public released academic datasets are allowed (this includes ImageNet, iNaturalist, etc; full citations needed.) Please specify all the external data used in your submission in the specified discussion post.\n\nGLHF!",
      "votes": 16
    },
    {
      "id": 502486,
      "postDate": "2019-03-28T16:18:39.343Z",
      "content": "<p>&gt;few-shot learning\n&gt;109k train\n&gt;40k private test\n&gt;kernels only</p>\n\n<p>First competition without ensembling, I guess.</p>",
      "rawMarkdown": "&gt;few-shot learning\n&gt;109k train\n&gt;40k private test\n&gt;kernels only\n\nFirst competition without ensembling, I guess.",
      "votes": 7
    },
    {
      "id": 504053,
      "postDate": "2019-03-30T22:34:00.817Z",
      "content": "<p>Thanks for hosting such a great competition.\nIt would be great if you could increase available shared memory, to be able to use PyTorch DataLoader with multiple workers: <a href=\"https://www.kaggle.com/product-feedback/72606\">https://www.kaggle.com/product-feedback/72606</a></p>",
      "rawMarkdown": "Thanks for hosting such a great competition.\nIt would be great if you could increase available shared memory, to be able to use PyTorch DataLoader with multiple workers: https://www.kaggle.com/product-feedback/72606",
      "votes": 6
    },
    {
      "id": 502674,
      "postDate": "2019-03-28T22:55:21.177Z",
      "content": "<p>&gt; but pre-trained model from public released academic datasets are allowed (this includes ImageNet, iNaturalist, etc; full citations needed.)</p>\n\n<p>How i can get pretraied model into pytorch, when internet connection is blocked? I'm import torchvision, use eg resnet34, specify 'pretrained=True' and ... get connection error :/</p>",
      "rawMarkdown": "&gt; but pre-trained model from public released academic datasets are allowed (this includes ImageNet, iNaturalist, etc; full citations needed.)\n\nHow i can get pretraied model into pytorch, when internet connection is blocked? I'm import torchvision, use eg resnet34, specify 'pretrained=True' and ... get connection error :/",
      "votes": 6
    },
    {
      "id": 502931,
      "postDate": "2019-03-29T09:19:08.190Z",
      "content": "<p>Must we do training and inference in one kernel? (9 hours)</p>",
      "rawMarkdown": "Must we do training and inference in one kernel? (9 hours)",
      "votes": 3
    },
    {
      "id": 542883,
      "postDate": "2019-06-04T06:57:04.220Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a>\n<a href=\"https://github.com/osmr/imgclsmob\">https://github.com/osmr/imgclsmob</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/osmr/imgclsmob",
      "votes": 1
    },
    {
      "id": 542236,
      "postDate": "2019-06-03T15:40:48.607Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch",
      "votes": 1
    },
    {
      "id": 540957,
      "postDate": "2019-06-01T12:29:31.563Z",
      "content": "<p>We may use: <a href=\"https://github.com/lukemelas/EfficientNet-PyTorch\">https://github.com/lukemelas/EfficientNet-PyTorch</a>\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "We may use: https://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/Cadene/pretrained-models.pytorch",
      "votes": 1,
      "replies": [
        {
          "id": 542052,
          "postDate": "2019-06-03T11:45:38.753Z",
          "content": "<p>waiting for the pretrained weights of efficientnet-pytorch</p>",
          "rawMarkdown": "waiting for the pretrained weights of efficientnet-pytorch",
          "votes": 1
        },
        {
          "id": 542165,
          "postDate": "2019-06-03T14:26:20.180Z",
          "content": "<p>EfficientNet b0-b3 pretrained weights have already been released</p>",
          "rawMarkdown": "EfficientNet b0-b3 pretrained weights have already been released"
        }
      ]
    },
    {
      "id": 538732,
      "postDate": "2019-05-29T03:24:34.500Z",
      "content": "<p>I'm using keras and pytorch imagenet pretrained models.\n<a href=\"https://keras.io/applications/\">https://keras.io/applications/</a>\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "I'm using keras and pytorch imagenet pretrained models.\nhttps://keras.io/applications/\nhttps://github.com/Cadene/pretrained-models.pytorch",
      "votes": 1
    },
    {
      "id": 510826,
      "postDate": "2019-04-09T13:59:40.757Z",
      "content": "<p>This is fixed by now (Sorry late to the party)\n<a href=\"https://www.kaggle.com/philmod/limitation-shared-memory-on-multiprocessing?scriptVersionId=12535758\">https://www.kaggle.com/philmod/limitation-shared-memory-on-multiprocessing?scriptVersionId=12535758</a></p>",
      "rawMarkdown": "This is fixed by now (Sorry late to the party)\nhttps://www.kaggle.com/philmod/limitation-shared-memory-on-multiprocessing?scriptVersionId=12535758",
      "votes": 1
    },
    {
      "id": 502481,
      "postDate": "2019-03-28T16:14:07.357Z",
      "content": "<p>It will be hard with kernels only. GLHF anyway</p>",
      "rawMarkdown": "It will be hard with kernels only. GLHF anyway",
      "votes": 1
    },
    {
      "id": 524207,
      "postDate": "2019-04-28T07:57:30.797Z",
      "content": "<p><a href=\"/codingcliff\">@codingcliff</a> Hi, in <code>Kernel Re-runs on Private Test Set</code> stage, will there be a new <code>sample_submission.csv</code> file? Since most of us will use this file for generating <code>submission.csv</code>, it would be kind if you can provide use with this information.</p>",
      "rawMarkdown": "@codingcliff Hi, in `Kernel Re-runs on Private Test Set` stage, will there be a new `sample_submission.csv` file? Since most of us will use this file for generating `submission.csv`, it would be kind if you can provide use with this information.",
      "votes": 2
    },
    {
      "id": 542435,
      "postDate": "2019-06-03T23:22:25.583Z",
      "content": "<p>I'm using pretrained models from:\n<a href=\"https://github.com/lukemelas/EfficientNet-PyTorch\">https://github.com/lukemelas/EfficientNet-PyTorch</a>\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a>\n<a href=\"https://github.com/osmr/imgclsmob/tree/master/pytorch\">https://github.com/osmr/imgclsmob/tree/master/pytorch</a></p>",
      "rawMarkdown": "I'm using pretrained models from:\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/osmr/imgclsmob/tree/master/pytorch",
      "votes": -2
    },
    {
      "id": 545903,
      "postDate": "2019-06-06T02:55:28.353Z",
      "content": "<p><a href=\"https://data.lip6.fr/cadene/pretrainedmodels/\">https://data.lip6.fr/cadene/pretrainedmodels/</a></p>",
      "rawMarkdown": "https://data.lip6.fr/cadene/pretrainedmodels/"
    },
    {
      "id": 545865,
      "postDate": "2019-06-06T01:27:32.910Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 545856,
      "postDate": "2019-06-06T00:52:19.313Z",
      "content": "<p><a href=\"https://github.com/lukemelas/EfficientNet-PyTorch\">https://github.com/lukemelas/EfficientNet-PyTorch</a>\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 545854,
      "postDate": "2019-06-06T00:46:20.997Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 545853,
      "postDate": "2019-06-06T00:42:36.417Z",
      "content": "<p>I’m using pretrained models from:\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "I’m using pretrained models from:\nhttps://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 544022,
      "postDate": "2019-06-05T03:33:46.620Z",
      "content": "<p>I'm using pretrained models from:\n<a href=\"https://github.com/lukemelas/EfficientNet-PyTorch\">https://github.com/lukemelas/EfficientNet-PyTorch</a>\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "I'm using pretrained models from:\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 543985,
      "postDate": "2019-06-05T02:25:26.353Z",
      "content": "<p>I’m using pretrained models from:\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "I’m using pretrained models from:\nhttps://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 543877,
      "postDate": "2019-06-04T22:52:18.073Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 543875,
      "postDate": "2019-06-04T22:49:56.593Z",
      "content": "<p><a href=\"https://www.kaggle.com/igorkrashenyi/pytorch-model-zoo\">https://www.kaggle.com/igorkrashenyi/pytorch-model-zoo</a></p>",
      "rawMarkdown": "https://www.kaggle.com/igorkrashenyi/pytorch-model-zoo"
    },
    {
      "id": 543869,
      "postDate": "2019-06-04T22:37:23.700Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 543853,
      "postDate": "2019-06-04T22:23:30.070Z",
      "content": "<p>Why submit a score for half an hour without responding?</p>",
      "rawMarkdown": "Why submit a score for half an hour without responding?"
    },
    {
      "id": 543641,
      "postDate": "2019-06-04T16:49:30.343Z",
      "content": "<p>My team and I used imagenet pretrained models from :\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "My team and I used imagenet pretrained models from :\nhttps://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 543504,
      "postDate": "2019-06-04T15:09:59.247Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 543482,
      "postDate": "2019-06-04T14:52:56.127Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 543470,
      "postDate": "2019-06-04T14:45:04.487Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 543421,
      "postDate": "2019-06-04T14:21:47.963Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 543409,
      "postDate": "2019-06-04T14:14:51.980Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch",
      "replies": [
        {
          "id": 549642,
          "postDate": "2019-06-10T22:34:14.940Z",
          "content": "<p>Hey, are you or your team will be in CVPR this year?</p>",
          "rawMarkdown": "Hey, are you or your team will be in CVPR this year?",
          "votes": 1
        }
      ]
    },
    {
      "id": 543388,
      "postDate": "2019-06-04T14:03:10.963Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 543285,
      "postDate": "2019-06-04T12:51:03.057Z",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 542857,
      "postDate": "2019-06-04T06:37:54.533Z",
      "content": "<p>I'm using pytorch imagenet pretrained models.\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "I'm using pytorch imagenet pretrained models.\nhttps://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 542679,
      "postDate": "2019-06-04T03:26:42.743Z",
      "content": "<p>I'm using pre-trained models here:  <a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "I'm using pre-trained models here:  https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 542425,
      "postDate": "2019-06-03T22:27:11.347Z",
      "content": "<p>I'm using pre-trained models here: <a href=\"https://www.kaggle.com/ttahara/chainercv-seresnext\">https://www.kaggle.com/ttahara/chainercv-seresnext</a>\nThese are converted for Chainer. Orignal weights are distributed at GitHub: <a href=\"https://github.com/hujie-frank/SENet\">https://github.com/hujie-frank/SENet</a></p>",
      "rawMarkdown": "I'm using pre-trained models here: https://www.kaggle.com/ttahara/chainercv-seresnext\nThese are converted for Chainer. Orignal weights are distributed at GitHub: https://github.com/hujie-frank/SENet"
    },
    {
      "id": 541912,
      "postDate": "2019-06-03T07:29:42.533Z",
      "content": "<p>We are using this: <a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "We are using this: https://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 541910,
      "postDate": "2019-06-03T07:24:28.403Z",
      "content": "<p>I'm using pretrained models from\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "I'm using pretrained models from\nhttps://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 541658,
      "postDate": "2019-06-02T21:01:07.530Z",
      "content": "<p>I'm using pretrained models from\n<a href=\"https://github.com/osmr/imgclsmob/tree/master/pytorch\">https://github.com/osmr/imgclsmob/tree/master/pytorch</a>\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "rawMarkdown": "I'm using pretrained models from\nhttps://github.com/osmr/imgclsmob/tree/master/pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch"
    },
    {
      "id": 518329,
      "postDate": "2019-04-17T05:07:56.087Z",
      "content": "<p>nice</p>",
      "rawMarkdown": "nice"
    },
    {
      "id": 516344,
      "postDate": "2019-04-14T03:20:22.843Z",
      "content": "<p><a href=\"/codingcliff\">@codingcliff</a> \n&gt; Please specify all the external data used in your submission in the specified discussion post.</p>\n\n<p>Which discussion topic that we should declare our use on pretrained models?\nI use standard Keras’ imagenet weights (e.g. ResNet50, InceptionV3, DenseNet, Xception) ... Just in case I have to declare here.</p>",
      "rawMarkdown": "@codingcliff \n&gt; Please specify all the external data used in your submission in the specified discussion post.\n\nWhich discussion topic that we should declare our use on pretrained models?\nI use standard Keras’ imagenet weights (e.g. ResNet50, InceptionV3, DenseNet, Xception) ... Just in case I have to declare here.",
      "replies": [
        {
          "id": 518616,
          "postDate": "2019-04-17T14:20:01.387Z",
          "content": "<p>Hi NE, Thanks for posting your info here. In your kernel submission, can you also put your declaration there? Thanks</p>",
          "rawMarkdown": "Hi NE, Thanks for posting your info here. In your kernel submission, can you also put your declaration there? Thanks",
          "votes": 2
        }
      ]
    },
    {
      "id": 516108,
      "postDate": "2019-04-13T16:13:36.327Z",
      "content": "<p>Is it true that the player must use the network disconnect mode?</p>",
      "rawMarkdown": "Is it true that the player must use the network disconnect mode?\n"
    },
    {
      "id": 502869,
      "postDate": "2019-03-29T07:05:06.200Z",
      "content": "<p>Is it OK to use external deep learning library which is not included in Kaggle's docker?. <br>\nFor example: <a href=\"https://github.com/belskikh/kekas\">kekas</a></p>",
      "rawMarkdown": "Is it OK to use external deep learning library which is not included in Kaggle's docker?.  \nFor example: [kekas](https://github.com/belskikh/kekas)",
      "replies": [
        {
          "id": 502870,
          "postDate": "2019-03-29T07:09:28.193Z",
          "content": "<p><code>\n- 9 hour runtime limit (including GPU Kernels)\n- No internet access enabled\n- Only whitelisted data is allowed\n- No custom packages\n- Submission file must be named \"submission.csv\"\n</code>\nThe custom package is not allowed. I think I got the answer.</p>",
          "rawMarkdown": "```\n- 9 hour runtime limit (including GPU Kernels)\n- No internet access enabled\n- Only whitelisted data is allowed\n- No custom packages\n- Submission file must be named \"submission.csv\"\n```\nThe custom package is not allowed. I think I got the answer.",
          "votes": 1
        }
      ]
    },
    {
      "id": 503550,
      "postDate": "2019-03-30T07:26:50.413Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 502486,
      "author_name": "Dmytro Panchenko",
      "author_url": "",
      "post_date": "2019-03-28T16:18:39.343000",
      "content": "<p>&gt;few-shot learning\n&gt;109k train\n&gt;40k private test\n&gt;kernels only</p>\n\n<p>First competition without ensembling, I guess.</p>",
      "votes": 7,
      "replies": []
    },
    {
      "id": 504053,
      "author_name": "Konstantin Lopukhin",
      "author_url": "",
      "post_date": "2019-03-30T22:34:00.817000",
      "content": "<p>Thanks for hosting such a great competition.\nIt would be great if you could increase available shared memory, to be able to use PyTorch DataLoader with multiple workers: <a href=\"https://www.kaggle.com/product-feedback/72606\">https://www.kaggle.com/product-feedback/72606</a></p>",
      "votes": 6,
      "replies": []
    },
    {
      "id": 502674,
      "author_name": "Igor Kotenkov",
      "author_url": "",
      "post_date": "2019-03-28T22:55:21.177000",
      "content": "<p>&gt; but pre-trained model from public released academic datasets are allowed (this includes ImageNet, iNaturalist, etc; full citations needed.)</p>\n\n<p>How i can get pretraied model into pytorch, when internet connection is blocked? I'm import torchvision, use eg resnet34, specify 'pretrained=True' and ... get connection error :/</p>",
      "votes": 6,
      "replies": []
    },
    {
      "id": 502931,
      "author_name": "seefun",
      "author_url": "",
      "post_date": "2019-03-29T09:19:08.190000",
      "content": "<p>Must we do training and inference in one kernel? (9 hours)</p>",
      "votes": 3,
      "replies": []
    },
    {
      "id": 542883,
      "author_name": "seefun",
      "author_url": "",
      "post_date": "2019-06-04T06:57:04.220000",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a>\n<a href=\"https://github.com/osmr/imgclsmob\">https://github.com/osmr/imgclsmob</a></p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 542236,
      "author_name": "Yiheng Wang",
      "author_url": "",
      "post_date": "2019-06-03T15:40:48.607000",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 540957,
      "author_name": "Yuanhao",
      "author_url": "",
      "post_date": "2019-06-01T12:29:31.563000",
      "content": "<p>We may use: <a href=\"https://github.com/lukemelas/EfficientNet-PyTorch\">https://github.com/lukemelas/EfficientNet-PyTorch</a>\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 1,
      "replies": [
        {
          "id": 542052,
          "author_name": "seefun",
          "author_url": "",
          "post_date": "2019-06-03T11:45:38.753000",
          "content": "<p>waiting for the pretrained weights of efficientnet-pytorch</p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 542165,
          "author_name": "Yuanhao",
          "author_url": "",
          "post_date": "2019-06-03T14:26:20.180000",
          "content": "<p>EfficientNet b0-b3 pretrained weights have already been released</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 538732,
      "author_name": "Guanshuo Xu",
      "author_url": "",
      "post_date": "2019-05-29T03:24:34.500000",
      "content": "<p>I'm using keras and pytorch imagenet pretrained models.\n<a href=\"https://keras.io/applications/\">https://keras.io/applications/</a>\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 510826,
      "author_name": "Aditya Soni",
      "author_url": "",
      "post_date": "2019-04-09T13:59:40.757000",
      "content": "<p>This is fixed by now (Sorry late to the party)\n<a href=\"https://www.kaggle.com/philmod/limitation-shared-memory-on-multiprocessing?scriptVersionId=12535758\">https://www.kaggle.com/philmod/limitation-shared-memory-on-multiprocessing?scriptVersionId=12535758</a></p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 502481,
      "author_name": "Dmitriy Ershov",
      "author_url": "",
      "post_date": "2019-03-28T16:14:07.357000",
      "content": "<p>It will be hard with kernels only. GLHF anyway</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 524207,
      "author_name": "good good study",
      "author_url": "",
      "post_date": "2019-04-28T07:57:30.797000",
      "content": "<p><a href=\"/codingcliff\">@codingcliff</a> Hi, in <code>Kernel Re-runs on Private Test Set</code> stage, will there be a new <code>sample_submission.csv</code> file? Since most of us will use this file for generating <code>submission.csv</code>, it would be kind if you can provide use with this information.</p>",
      "votes": 2,
      "replies": []
    },
    {
      "id": 542435,
      "author_name": "n01z3",
      "author_url": "",
      "post_date": "2019-06-03T23:22:25.583000",
      "content": "<p>I'm using pretrained models from:\n<a href=\"https://github.com/lukemelas/EfficientNet-PyTorch\">https://github.com/lukemelas/EfficientNet-PyTorch</a>\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a>\n<a href=\"https://github.com/osmr/imgclsmob/tree/master/pytorch\">https://github.com/osmr/imgclsmob/tree/master/pytorch</a></p>",
      "votes": -2,
      "replies": []
    },
    {
      "id": 545903,
      "author_name": "Jar",
      "author_url": "",
      "post_date": "2019-06-06T02:55:28.353000",
      "content": "<p><a href=\"https://data.lip6.fr/cadene/pretrainedmodels/\">https://data.lip6.fr/cadene/pretrainedmodels/</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 545865,
      "author_name": "JiaqiFan",
      "author_url": "",
      "post_date": "2019-06-06T01:27:32.910000",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 545856,
      "author_name": "Strideradu",
      "author_url": "",
      "post_date": "2019-06-06T00:52:19.313000",
      "content": "<p><a href=\"https://github.com/lukemelas/EfficientNet-PyTorch\">https://github.com/lukemelas/EfficientNet-PyTorch</a>\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 545854,
      "author_name": "FuLin",
      "author_url": "",
      "post_date": "2019-06-06T00:46:20.997000",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 545853,
      "author_name": "Dmitry Abulkhanov",
      "author_url": "",
      "post_date": "2019-06-06T00:42:36.417000",
      "content": "<p>I’m using pretrained models from:\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 544022,
      "author_name": "David",
      "author_url": "",
      "post_date": "2019-06-05T03:33:46.620000",
      "content": "<p>I'm using pretrained models from:\n<a href=\"https://github.com/lukemelas/EfficientNet-PyTorch\">https://github.com/lukemelas/EfficientNet-PyTorch</a>\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543985,
      "author_name": "kumapo",
      "author_url": "",
      "post_date": "2019-06-05T02:25:26.353000",
      "content": "<p>I’m using pretrained models from:\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543877,
      "author_name": "earhian",
      "author_url": "",
      "post_date": "2019-06-04T22:52:18.073000",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543875,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-06-04T22:49:56.593000",
      "content": "<p><a href=\"https://www.kaggle.com/igorkrashenyi/pytorch-model-zoo\">https://www.kaggle.com/igorkrashenyi/pytorch-model-zoo</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543869,
      "author_name": "nekoder",
      "author_url": "",
      "post_date": "2019-06-04T22:37:23.700000",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543853,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-06-04T22:23:30.070000",
      "content": "<p>Why submit a score for half an hour without responding?</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543641,
      "author_name": "Firas Baba",
      "author_url": "",
      "post_date": "2019-06-04T16:49:30.343000",
      "content": "<p>My team and I used imagenet pretrained models from :\n<a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543504,
      "author_name": "Filemon",
      "author_url": "",
      "post_date": "2019-06-04T15:09:59.247000",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543482,
      "author_name": "rskmoi",
      "author_url": "",
      "post_date": "2019-06-04T14:52:56.127000",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543470,
      "author_name": "e-mon",
      "author_url": "",
      "post_date": "2019-06-04T14:45:04.487000",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543421,
      "author_name": "Y.Nakama",
      "author_url": "",
      "post_date": "2019-06-04T14:21:47.963000",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543409,
      "author_name": "owruby",
      "author_url": "",
      "post_date": "2019-06-04T14:14:51.980000",
      "content": "<p><a href=\"https://github.com/Cadene/pretrained-models.pytorch\">https://github.com/Cadene/pretrained-models.pytorch</a></p>",
      "votes": 0,
      "replies": [
        {
          "id": 549642,
          "author_name": "Chenyang Zhang",
          "author_url": "",
          "post_date": "2019-06-10T22:34:14.940000",
          "content": "<p>Hey, are you or your team will be in CVPR this year?</p>",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 543388,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-06-04T14:03:10.963000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 543285,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-06-04T12:51:03.057000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 542857,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-06-04T06:37:54.533000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 542679,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-06-04T03:26:42.743000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 542425,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-06-03T22:27:11.347000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 541912,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-06-03T07:29:42.533000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 541910,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-06-03T07:24:28.403000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 541658,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-06-02T21:01:07.530000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 518329,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-04-17T05:07:56.087000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 516344,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-04-14T03:20:22.843000",
      "content": "",
      "votes": 0,
      "replies": [
        {
          "id": 518616,
          "author_name": "",
          "author_url": "",
          "post_date": "2019-04-17T14:20:01.387000",
          "content": "",
          "votes": 2,
          "replies": []
        }
      ]
    },
    {
      "id": 516108,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-04-13T16:13:36.327000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 502869,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-03-29T07:05:06.200000",
      "content": "",
      "votes": 0,
      "replies": [
        {
          "id": 502870,
          "author_name": "",
          "author_url": "",
          "post_date": "2019-03-29T07:09:28.193000",
          "content": "",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 503550,
      "author_name": "",
      "author_url": "",
      "post_date": "2019-03-30T07:26:50.413000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "502446": "In this competition, we are recognizing multiple attributes from (and beautiful) artwork images from The Metropolitan Museum of Art.\n\nWe want to note that this is a **kernels-only competition**, i.e., all submissions to this competition must be made through **kernels**. For fair competition, no external data should be used, but pre-trained model from public released academic datasets are allowed (this includes ImageNet, iNaturalist, etc; full citations needed.) Please specify all the external data used in your submission in the specified discussion post.\n\nGLHF!",
    "502486": "&gt;few-shot learning\n&gt;109k train\n&gt;40k private test\n&gt;kernels only\n\nFirst competition without ensembling, I guess.",
    "504053": "Thanks for hosting such a great competition.\nIt would be great if you could increase available shared memory, to be able to use PyTorch DataLoader with multiple workers: https://www.kaggle.com/product-feedback/72606",
    "502674": "&gt; but pre-trained model from public released academic datasets are allowed (this includes ImageNet, iNaturalist, etc; full citations needed.)\n\nHow i can get pretraied model into pytorch, when internet connection is blocked? I'm import torchvision, use eg resnet34, specify 'pretrained=True' and ... get connection error :/",
    "502931": "Must we do training and inference in one kernel? (9 hours)",
    "542883": "https://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/osmr/imgclsmob",
    "542236": "https://github.com/Cadene/pretrained-models.pytorch",
    "540957": "We may use: https://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/Cadene/pretrained-models.pytorch",
    "538732": "I'm using keras and pytorch imagenet pretrained models.\nhttps://keras.io/applications/\nhttps://github.com/Cadene/pretrained-models.pytorch",
    "510826": "This is fixed by now (Sorry late to the party)\nhttps://www.kaggle.com/philmod/limitation-shared-memory-on-multiprocessing?scriptVersionId=12535758",
    "502481": "It will be hard with kernels only. GLHF anyway",
    "524207": "@codingcliff Hi, in `Kernel Re-runs on Private Test Set` stage, will there be a new `sample_submission.csv` file? Since most of us will use this file for generating `submission.csv`, it would be kind if you can provide use with this information.",
    "542435": "I'm using pretrained models from:\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/osmr/imgclsmob/tree/master/pytorch",
    "545903": "https://data.lip6.fr/cadene/pretrainedmodels/",
    "545865": "https://github.com/Cadene/pretrained-models.pytorch",
    "545856": "https://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/Cadene/pretrained-models.pytorch",
    "545854": "https://github.com/Cadene/pretrained-models.pytorch",
    "545853": "I’m using pretrained models from:\nhttps://github.com/Cadene/pretrained-models.pytorch",
    "544022": "I'm using pretrained models from:\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/Cadene/pretrained-models.pytorch",
    "543985": "I’m using pretrained models from:\nhttps://github.com/Cadene/pretrained-models.pytorch",
    "543877": "https://github.com/Cadene/pretrained-models.pytorch",
    "543875": "https://www.kaggle.com/igorkrashenyi/pytorch-model-zoo",
    "543869": "https://github.com/Cadene/pretrained-models.pytorch",
    "543853": "Why submit a score for half an hour without responding?",
    "543641": "My team and I used imagenet pretrained models from :\nhttps://github.com/Cadene/pretrained-models.pytorch",
    "543504": "https://github.com/Cadene/pretrained-models.pytorch",
    "543482": "https://github.com/Cadene/pretrained-models.pytorch",
    "543470": "https://github.com/Cadene/pretrained-models.pytorch",
    "543421": "https://github.com/Cadene/pretrained-models.pytorch",
    "543409": "https://github.com/Cadene/pretrained-models.pytorch",
    "543388": "https://github.com/Cadene/pretrained-models.pytorch",
    "543285": "https://github.com/Cadene/pretrained-models.pytorch",
    "542857": "I'm using pytorch imagenet pretrained models.\nhttps://github.com/Cadene/pretrained-models.pytorch",
    "542679": "I'm using pre-trained models here:  https://github.com/Cadene/pretrained-models.pytorch",
    "542425": "I'm using pre-trained models here: https://www.kaggle.com/ttahara/chainercv-seresnext\nThese are converted for Chainer. Orignal weights are distributed at GitHub: https://github.com/hujie-frank/SENet",
    "541912": "We are using this: https://github.com/Cadene/pretrained-models.pytorch",
    "541910": "I'm using pretrained models from\nhttps://github.com/Cadene/pretrained-models.pytorch",
    "541658": "I'm using pretrained models from\nhttps://github.com/osmr/imgclsmob/tree/master/pytorch\nhttps://github.com/Cadene/pretrained-models.pytorch",
    "518329": "nice",
    "516344": "@codingcliff \n&gt; Please specify all the external data used in your submission in the specified discussion post.\n\nWhich discussion topic that we should declare our use on pretrained models?\nI use standard Keras’ imagenet weights (e.g. ResNet50, InceptionV3, DenseNet, Xception) ... Just in case I have to declare here.",
    "516108": "Is it true that the player must use the network disconnect mode?\n",
    "502869": "Is it OK to use external deep learning library which is not included in Kaggle's docker?.  \nFor example: [kekas](https://github.com/belskikh/kekas)",
    "503550": ""
  }
}