{
  "id": 175402,
  "title": "Pytorch(Stratified-Kfold-5fold - 256*256) - No Augmentation - No Triple Stratified - No Multiple resized images -Public LB- 91.4 Private LB Score- 90.41",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/175402",
  "author_name": "",
  "post_date": "2020-08-18T05:10:20.401686Z",
  "votes": 3,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Experimented 256 * 256 image size dataset<br>\nstratified kfold - 5 fold<br>\nPublic Score : 91.41%<br>\nPrivate Score: 90.40%<br>\nAugmentation: only Normalization no other augmentation <br>\nStill this model stable with this dataset💥 learnt a lot from this competition.<br>\nGPU:-  1050 week long training 😩</p>\n<p>Thanks a lot <a href=\"https://www.kaggle.com/cdeotte\" target=\"_blank\">@cdeotte</a> learnt many things from U😎</p>",
  "messages": [
    {
      "id": "974953",
      "postDate": "08/18/2020 05:10:20",
      "content": "<p>Experimented 256 * 256 image size dataset<br>\nstratified kfold - 5 fold<br>\nPublic Score : 91.41%<br>\nPrivate Score: 90.40%<br>\nAugmentation: only Normalization no other augmentation <br>\nStill this model stable with this dataset💥 learnt a lot from this competition.<br>\nGPU:-  1050 week long training 😩</p>\n<p>Thanks a lot <a href=\"https://www.kaggle.com/cdeotte\" target=\"_blank\">@cdeotte</a> learnt many things from U😎</p>",
      "rawMarkdown": "Experimented 256 * 256 image size dataset\nstratified kfold - 5 fold\nPublic Score : 91.41%\nPrivate Score: 90.40%\nAugmentation: only Normalization no other augmentation \nStill this model stable with this dataset💥 learnt a lot from this competition.\nGPU:-  1050 week long training 😩\n\nThanks a lot @cdeotte learnt many things from U😎",
      "votes": null
    },
    {
      "id": "976055",
      "postDate": "08/18/2020 15:56:29",
      "content": "<p>I used only kaggle notebooks GPU/TPU. Training effnet B4  for 5 fold - 12 epochs each ran for at-most 2.5 hours ! Might I suggest use Kaggle resources next time.</p>",
      "rawMarkdown": "I used only kaggle notebooks GPU/TPU. Training effnet B4  for 5 fold - 12 epochs each ran for at-most 2.5 hours ! Might I suggest use Kaggle resources next time.",
      "votes": null
    },
    {
      "id": "976134",
      "postDate": "08/18/2020 16:47:06",
      "content": "<p><a href=\"https://www.kaggle.com/realsid\" target=\"_blank\">@realsid</a> I have faced some issue with kaggle initially, i couldn't able to run it</p>",
      "rawMarkdown": "realsid I have faced some issue with kaggle initially, i couldn't able to run it",
      "votes": null
    },
    {
      "id": "976144",
      "postDate": "08/18/2020 16:55:03",
      "content": "<p>What sort of problem bud ? If you can elaborate there are a lot of people ready to help in the community !</p>",
      "rawMarkdown": "What sort of problem bud ? If you can elaborate there are a lot of people ready to help in the community !",
      "votes": null
    },
    {
      "id": "976588",
      "postDate": "08/19/2020 01:06:15",
      "content": "<p>time out error very often and internet issue also so i decided to use own gpu that point of time</p>",
      "rawMarkdown": "time out error very often and internet issue also so i decided to use own gpu that point of time",
      "votes": null
    },
    {
      "id": "976590",
      "postDate": "08/19/2020 01:07:34",
      "content": "<p>can u tell me what kind of techniques u followed for this competition? u got 691  rank😎</p>",
      "rawMarkdown": "can u tell me what kind of techniques u followed for this competition? u got 691  rank😎",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 976055,
      "author_name": "realsid",
      "author_url": "",
      "post_date": "08/18/2020 15:56:29",
      "content": "<p>I used only kaggle notebooks GPU/TPU. Training effnet B4  for 5 fold - 12 epochs each ran for at-most 2.5 hours ! Might I suggest use Kaggle resources next time.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 976134,
      "author_name": "santhoshkumarv",
      "author_url": "",
      "post_date": "08/18/2020 16:47:06",
      "content": "<p><a href=\"https://www.kaggle.com/realsid\" target=\"_blank\">@realsid</a> I have faced some issue with kaggle initially, i couldn't able to run it</p>",
      "votes": null,
      "replies": [
        {
          "id": 976144,
          "author_name": "realsid",
          "author_url": "",
          "post_date": "08/18/2020 16:55:03",
          "content": "<p>What sort of problem bud ? If you can elaborate there are a lot of people ready to help in the community !</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 976588,
          "author_name": "santhoshkumarv",
          "author_url": "",
          "post_date": "08/19/2020 01:06:15",
          "content": "<p>time out error very often and internet issue also so i decided to use own gpu that point of time</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 976590,
          "author_name": "santhoshkumarv",
          "author_url": "",
          "post_date": "08/19/2020 01:07:34",
          "content": "<p>can u tell me what kind of techniques u followed for this competition? u got 691  rank😎</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "974953": "Experimented 256 * 256 image size dataset\nstratified kfold - 5 fold\nPublic Score : 91.41%\nPrivate Score: 90.40%\nAugmentation: only Normalization no other augmentation \nStill this model stable with this dataset💥 learnt a lot from this competition.\nGPU:-  1050 week long training 😩\n\nThanks a lot @cdeotte learnt many things from U😎",
    "976055": "I used only kaggle notebooks GPU/TPU. Training effnet B4  for 5 fold - 12 epochs each ran for at-most 2.5 hours ! Might I suggest use Kaggle resources next time.",
    "976134": "realsid I have faced some issue with kaggle initially, i couldn't able to run it",
    "976144": "What sort of problem bud ? If you can elaborate there are a lot of people ready to help in the community !",
    "976588": "time out error very often and internet issue also so i decided to use own gpu that point of time",
    "976590": "can u tell me what kind of techniques u followed for this competition? u got 691  rank😎"
  },
  "source": "meta"
}