{
  "id": 102572,
  "title": "Has anyone managed character detection inside kernels?",
  "url": "/competitions/kuzushiji-recognition/discussion/102572",
  "author_name": "",
  "post_date": "2019-08-02T21:06:58.021373500Z",
  "votes": 3,
  "comment_count": 7,
  "views": 0,
  "content": "<p>Hey friends,\nhas anyone managed to do the character detection inside the kernels?\nGitHub repos or custom packages can't be used because you have to enable the GPU.</p>\n\n<p>I tried to find a package that\n- can be used inside Kaggle kernels\n- can retrain an object detection model\n- can be implemented rather easily\nBut I couldn't find one!</p>\n\n<p>If anyone has any useful information, I would be grateful\nthanks :)</p>",
  "messages": [
    {
      "id": "590930",
      "postDate": "08/02/2019 21:06:58",
      "content": "<p>Hey friends,\nhas anyone managed to do the character detection inside the kernels?\nGitHub repos or custom packages can't be used because you have to enable the GPU.</p>\n\n<p>I tried to find a package that\n- can be used inside Kaggle kernels\n- can retrain an object detection model\n- can be implemented rather easily\nBut I couldn't find one!</p>\n\n<p>If anyone has any useful information, I would be grateful\nthanks :)</p>",
      "rawMarkdown": "Hey friends,\nhas anyone managed to do the character detection inside the kernels?\nGitHub repos or custom packages can't be used because you have to enable the GPU.\n\nI tried to find a package that\n- can be used inside Kaggle kernels\n- can retrain an object detection model\n- can be implemented rather easily\nBut I couldn't find one!\n\nIf anyone has any useful information, I would be grateful\nthanks :)",
      "votes": null
    },
    {
      "id": "591873",
      "postDate": "08/04/2019 11:58:44",
      "content": "<blockquote>\n  <p>GitHub repos or custom packages can't be used because you have to enable the GPU.</p>\n</blockquote>\n\n<p>I don't think so, you can enabled internet in GPU kernels</p>",
      "rawMarkdown": "&gt; GitHub repos or custom packages can't be used because you have to enable the GPU.\n\nI don't think so, you can enabled internet in GPU kernels",
      "votes": null
    },
    {
      "id": "592152",
      "postDate": "08/04/2019 22:04:38",
      "content": "<p>Haven't tried it yet but <a href=\"https://pytorch.org/tutorials/intermediate/torchvision_tutorial.html\">this</a> seems very promising\nMaking a starter kenel soon</p>",
      "rawMarkdown": "Haven't tried it yet but [this](https://pytorch.org/tutorials/intermediate/torchvision_tutorial.html) seems very promising\nMaking a starter kenel soon",
      "votes": null
    },
    {
      "id": "592433",
      "postDate": "08/05/2019 09:22:16",
      "content": "<p>In my kernel, it shows this message. Or is there another way?\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2063391%2Ff40b69dcc926769cd4844ab41e49e7ff%2FGPU-packages.png?generation=1564996779357888&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "In my kernel, it shows this message. Or is there another way?\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2063391%2Ff40b69dcc926769cd4844ab41e49e7ff%2FGPU-packages.png?generation=1564996779357888&amp;alt=media)",
      "votes": null
    },
    {
      "id": "592443",
      "postDate": "08/05/2019 09:50:45",
      "content": "<p>You can install packages with pip or clone them from github or add them as datasets and install from there</p>",
      "rawMarkdown": "You can install packages with pip or clone them from github or add them as datasets and install from there",
      "votes": null
    },
    {
      "id": "592687",
      "postDate": "08/05/2019 17:25:24",
      "content": "<p>thanks a lot</p>",
      "rawMarkdown": "thanks a lot",
      "votes": null
    },
    {
      "id": "592855",
      "postDate": "08/05/2019 22:31:59",
      "content": "<p>Hi Chris,\nI'm currently trying to develop a solution using this Keras/Tensorflow implementation for <a href=\"https://github.com/matterport/Mask_RCNN/blob/master/samples/demo.ipynb\">MaskRCNN</a>. I've been taking inspiration from <a href=\"https://www.kaggle.com/pednoi/training-mask-r-cnn-to-be-a-fashionista-lb-0-07\">this great kernel</a>. Otherwise, I'm hoping to have a kernel ready to share by the end of the week. Hope this helps!</p>",
      "rawMarkdown": "Hi Chris,\nI'm currently trying to develop a solution using this Keras/Tensorflow implementation for [MaskRCNN]( https://github.com/matterport/Mask_RCNN/blob/master/samples/demo.ipynb). I've been taking inspiration from [this great kernel]( https://www.kaggle.com/pednoi/training-mask-r-cnn-to-be-a-fashionista-lb-0-07). Otherwise, I'm hoping to have a kernel ready to share by the end of the week. Hope this helps!",
      "votes": null
    },
    {
      "id": "593071",
      "postDate": "08/06/2019 06:53:14",
      "content": "<p>perfect, thank you</p>",
      "rawMarkdown": "perfect, thank you",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 591873,
      "author_name": "lopuhin",
      "author_url": "",
      "post_date": "08/04/2019 11:58:44",
      "content": "<blockquote>\n  <p>GitHub repos or custom packages can't be used because you have to enable the GPU.</p>\n</blockquote>\n\n<p>I don't think so, you can enabled internet in GPU kernels</p>",
      "votes": null,
      "replies": [
        {
          "id": 592433,
          "author_name": "christianwallenwein",
          "author_url": "",
          "post_date": "08/05/2019 09:22:16",
          "content": "<p>In my kernel, it shows this message. Or is there another way?\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2063391%2Ff40b69dcc926769cd4844ab41e49e7ff%2FGPU-packages.png?generation=1564996779357888&amp;alt=media\" alt=\"\"></p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 592443,
          "author_name": "lopuhin",
          "author_url": "",
          "post_date": "08/05/2019 09:50:45",
          "content": "<p>You can install packages with pip or clone them from github or add them as datasets and install from there</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 592687,
          "author_name": "christianwallenwein",
          "author_url": "",
          "post_date": "08/05/2019 17:25:24",
          "content": "<p>thanks a lot</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 592152,
      "author_name": "sidhanthholalkere",
      "author_url": "",
      "post_date": "08/04/2019 22:04:38",
      "content": "<p>Haven't tried it yet but <a href=\"https://pytorch.org/tutorials/intermediate/torchvision_tutorial.html\">this</a> seems very promising\nMaking a starter kenel soon</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 592855,
      "author_name": "frlemarchand",
      "author_url": "",
      "post_date": "08/05/2019 22:31:59",
      "content": "<p>Hi Chris,\nI'm currently trying to develop a solution using this Keras/Tensorflow implementation for <a href=\"https://github.com/matterport/Mask_RCNN/blob/master/samples/demo.ipynb\">MaskRCNN</a>. I've been taking inspiration from <a href=\"https://www.kaggle.com/pednoi/training-mask-r-cnn-to-be-a-fashionista-lb-0-07\">this great kernel</a>. Otherwise, I'm hoping to have a kernel ready to share by the end of the week. Hope this helps!</p>",
      "votes": null,
      "replies": [
        {
          "id": 593071,
          "author_name": "christianwallenwein",
          "author_url": "",
          "post_date": "08/06/2019 06:53:14",
          "content": "<p>perfect, thank you</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "590930": "Hey friends,\nhas anyone managed to do the character detection inside the kernels?\nGitHub repos or custom packages can't be used because you have to enable the GPU.\n\nI tried to find a package that\n- can be used inside Kaggle kernels\n- can retrain an object detection model\n- can be implemented rather easily\nBut I couldn't find one!\n\nIf anyone has any useful information, I would be grateful\nthanks :)",
    "591873": "&gt; GitHub repos or custom packages can't be used because you have to enable the GPU.\n\nI don't think so, you can enabled internet in GPU kernels",
    "592152": "Haven't tried it yet but [this](https://pytorch.org/tutorials/intermediate/torchvision_tutorial.html) seems very promising\nMaking a starter kenel soon",
    "592433": "In my kernel, it shows this message. Or is there another way?\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2063391%2Ff40b69dcc926769cd4844ab41e49e7ff%2FGPU-packages.png?generation=1564996779357888&amp;alt=media)",
    "592443": "You can install packages with pip or clone them from github or add them as datasets and install from there",
    "592687": "thanks a lot",
    "592855": "Hi Chris,\nI'm currently trying to develop a solution using this Keras/Tensorflow implementation for [MaskRCNN]( https://github.com/matterport/Mask_RCNN/blob/master/samples/demo.ipynb). I've been taking inspiration from [this great kernel]( https://www.kaggle.com/pednoi/training-mask-r-cnn-to-be-a-fashionista-lb-0-07). Otherwise, I'm hoping to have a kernel ready to share by the end of the week. Hope this helps!",
    "593071": "perfect, thank you"
  },
  "source": "meta"
}