{
  "id": 132986,
  "title": "GridMask data augmentation on GPU/TPU",
  "url": "/competitions/flower-classification-with-tpus/discussion/132986",
  "author_name": "",
  "post_date": "2020-02-29T05:37:45.871305900Z",
  "votes": 18,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hi Guys, I had posted a <a href=\"https://www.kaggle.com/xiejialun/gridmask-data-augmentation-with-tensorflow\">kernel</a> about how to perform GridMask data augmentation on GPU/TPU. </p>\n\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3617078%2F5bd38f98a37535a92970b5b338e519f6%2F2020-02-29%201.32.04.png?generation=1582954461136321&amp;alt=media\" alt=\"\"></p>\n\n<p>And I also posted a kernel several days ago with several data augmentation methods(gaussian blur, random cutout...) on GPU/TPU <a href=\"https://www.kaggle.com/xiejialun/customize-data-augmentation-with-tensorflow\">here</a></p>\n\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3617078%2F54d7001fa13156814b0b7ec12816847b%2F2020-02-29%201.35.38.png?generation=1582954662845009&amp;alt=media\" alt=\"\"></p>",
  "messages": [
    {
      "id": "759498",
      "postDate": "02/29/2020 05:37:45",
      "content": "<p>Hi Guys, I had posted a <a href=\"https://www.kaggle.com/xiejialun/gridmask-data-augmentation-with-tensorflow\">kernel</a> about how to perform GridMask data augmentation on GPU/TPU. </p>\n\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3617078%2F5bd38f98a37535a92970b5b338e519f6%2F2020-02-29%201.32.04.png?generation=1582954461136321&amp;alt=media\" alt=\"\"></p>\n\n<p>And I also posted a kernel several days ago with several data augmentation methods(gaussian blur, random cutout...) on GPU/TPU <a href=\"https://www.kaggle.com/xiejialun/customize-data-augmentation-with-tensorflow\">here</a></p>\n\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3617078%2F54d7001fa13156814b0b7ec12816847b%2F2020-02-29%201.35.38.png?generation=1582954662845009&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "Hi Guys, I had posted a [kernel](https://www.kaggle.com/xiejialun/gridmask-data-augmentation-with-tensorflow) about how to perform GridMask data augmentation on GPU/TPU. \n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3617078%2F5bd38f98a37535a92970b5b338e519f6%2F2020-02-29%201.32.04.png?generation=1582954461136321&amp;alt=media)\n\nAnd I also posted a kernel several days ago with several data augmentation methods(gaussian blur, random cutout...) on GPU/TPU [here](https://www.kaggle.com/xiejialun/customize-data-augmentation-with-tensorflow)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3617078%2F54d7001fa13156814b0b7ec12816847b%2F2020-02-29%201.35.38.png?generation=1582954662845009&amp;alt=media)",
      "votes": null
    },
    {
      "id": "759504",
      "postDate": "02/29/2020 05:52:30",
      "content": "<p>Nice work. Awesome. Thanks.</p>",
      "rawMarkdown": "Nice work. Awesome. Thanks.",
      "votes": null
    },
    {
      "id": "759506",
      "postDate": "02/29/2020 05:57:32",
      "content": "<p>Thanks for your awesome kernels too! I have learned a lot from your great works.</p>",
      "rawMarkdown": "Thanks for your awesome kernels too! I have learned a lot from your great works.",
      "votes": null
    },
    {
      "id": "759507",
      "postDate": "02/29/2020 05:58:22",
      "content": "<p>FYI, both links in your post go to your GridMask notebook.</p>\n\n<p>EDIT: links correct now</p>",
      "rawMarkdown": "FYI, both links in your post go to your GridMask notebook.\n\nEDIT: links correct now",
      "votes": null
    },
    {
      "id": "759510",
      "postDate": "02/29/2020 06:00:57",
      "content": "<p>Oops, thanks for the reminder!</p>",
      "rawMarkdown": "Oops, thanks for the reminder!",
      "votes": null
    },
    {
      "id": "761698",
      "postDate": "03/02/2020 21:13:48",
      "content": "<p>Fantastic work. Thank you.</p>",
      "rawMarkdown": "Fantastic work. Thank you.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 759504,
      "author_name": "cdeotte",
      "author_url": "",
      "post_date": "02/29/2020 05:52:30",
      "content": "<p>Nice work. Awesome. Thanks.</p>",
      "votes": null,
      "replies": [
        {
          "id": 759506,
          "author_name": "xiejialun",
          "author_url": "",
          "post_date": "02/29/2020 05:57:32",
          "content": "<p>Thanks for your awesome kernels too! I have learned a lot from your great works.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 759507,
      "author_name": "cdeotte",
      "author_url": "",
      "post_date": "02/29/2020 05:58:22",
      "content": "<p>FYI, both links in your post go to your GridMask notebook.</p>\n\n<p>EDIT: links correct now</p>",
      "votes": null,
      "replies": [
        {
          "id": 759510,
          "author_name": "xiejialun",
          "author_url": "",
          "post_date": "02/29/2020 06:00:57",
          "content": "<p>Oops, thanks for the reminder!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 761698,
      "author_name": "mgorner",
      "author_url": "",
      "post_date": "03/02/2020 21:13:48",
      "content": "<p>Fantastic work. Thank you.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "759498": "Hi Guys, I had posted a [kernel](https://www.kaggle.com/xiejialun/gridmask-data-augmentation-with-tensorflow) about how to perform GridMask data augmentation on GPU/TPU. \n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3617078%2F5bd38f98a37535a92970b5b338e519f6%2F2020-02-29%201.32.04.png?generation=1582954461136321&amp;alt=media)\n\nAnd I also posted a kernel several days ago with several data augmentation methods(gaussian blur, random cutout...) on GPU/TPU [here](https://www.kaggle.com/xiejialun/customize-data-augmentation-with-tensorflow)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3617078%2F54d7001fa13156814b0b7ec12816847b%2F2020-02-29%201.35.38.png?generation=1582954662845009&amp;alt=media)",
    "759504": "Nice work. Awesome. Thanks.",
    "759506": "Thanks for your awesome kernels too! I have learned a lot from your great works.",
    "759507": "FYI, both links in your post go to your GridMask notebook.\n\nEDIT: links correct now",
    "759510": "Oops, thanks for the reminder!",
    "761698": "Fantastic work. Thank you."
  },
  "source": "meta"
}