{
  "id": 177563,
  "title": "[2nd Place] Source Code",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/177563",
  "author_name": "",
  "post_date": "2020-08-26T11:12:04.556831500Z",
  "votes": 34,
  "comment_count": 8,
  "views": 0,
  "content": "<p>Hi Kagglers,</p>\n<p>The source code for my solution is available here: <a href=\"https://github.com/i-pan/kaggle-melanoma\" target=\"_blank\">https://github.com/i-pan/kaggle-melanoma</a>. I apologize in advance for its messiness and lack of comments. I need to work on best practices for coding. 😂</p>\n<p>There are instructions to train/run inference on the competition data. Inevitably there will be errors when some of you try to run it, so please open issue and I will try to address it in a timely manner. </p>\n<p>This is also the code I've been using over the past 6 months or so for Kaggle. Hopefully its organization can help people create their own Kaggle pipelines. My focus was on modularity and reproducibility for rapid experimentation. In particular, I started using YAML config files for each experiment I run to keep track of what I was doing. </p>",
  "messages": [
    {
      "id": "986307",
      "postDate": "08/26/2020 11:12:04",
      "content": "<p>Hi Kagglers,</p>\n<p>The source code for my solution is available here: <a href=\"https://github.com/i-pan/kaggle-melanoma\" target=\"_blank\">https://github.com/i-pan/kaggle-melanoma</a>. I apologize in advance for its messiness and lack of comments. I need to work on best practices for coding. 😂</p>\n<p>There are instructions to train/run inference on the competition data. Inevitably there will be errors when some of you try to run it, so please open issue and I will try to address it in a timely manner. </p>\n<p>This is also the code I've been using over the past 6 months or so for Kaggle. Hopefully its organization can help people create their own Kaggle pipelines. My focus was on modularity and reproducibility for rapid experimentation. In particular, I started using YAML config files for each experiment I run to keep track of what I was doing. </p>",
      "rawMarkdown": "Hi Kagglers,\n\nThe source code for my solution is available here: [https://github.com/i-pan/kaggle-melanoma](https://github.com/i-pan/kaggle-melanoma). I apologize in advance for its messiness and lack of comments. I need to work on best practices for coding. 😂\n\nThere are instructions to train/run inference on the competition data. Inevitably there will be errors when some of you try to run it, so please open issue and I will try to address it in a timely manner. \n\nThis is also the code I've been using over the past 6 months or so for Kaggle. Hopefully its organization can help people create their own Kaggle pipelines. My focus was on modularity and reproducibility for rapid experimentation. In particular, I started using YAML config files for each experiment I run to keep track of what I was doing.",
      "votes": null
    },
    {
      "id": "986325",
      "postDate": "08/26/2020 11:32:19",
      "content": "<p>Thank you <a href=\"https://www.kaggle.com/vaillant\" target=\"_blank\">@vaillant</a> for sharing! Was looking for top solutions! Also congrats!</p>",
      "rawMarkdown": "Thank you @vaillant for sharing! Was looking for top solutions! Also congrats!",
      "votes": null
    },
    {
      "id": "987103",
      "postDate": "08/27/2020 02:48:43",
      "content": "<p>Thank you for sharing <a href=\"https://www.kaggle.com/vaillant\" target=\"_blank\">@vaillant</a> , I want to post my code on github as well. How did you learn to structure your code like this? Any resources you would recommend to help with structuring code? Thank you, will greatly appreciate it</p>",
      "rawMarkdown": "Thank you for sharing @vaillant , I want to post my code on github as well. How did you learn to structure your code like this? Any resources you would recommend to help with structuring code? Thank you, will greatly appreciate it",
      "votes": null
    },
    {
      "id": "987116",
      "postDate": "08/27/2020 03:03:38",
      "content": "<p>Thank you for sharing <a href=\"https://www.kaggle.com/vaillant\" target=\"_blank\">@vaillant</a> 👍!</p>",
      "rawMarkdown": "Thank you for sharing @vaillant 👍!",
      "votes": null
    },
    {
      "id": "987188",
      "postDate": "08/27/2020 04:38:55",
      "content": "<p>Thanks for sharing <a href=\"https://www.kaggle.com/vaillant\" target=\"_blank\">@vaillant</a>.</p>\n<p>I clicked your github link, it is go to this:<br>\n<code>kaggle.com/c/siim-isic-melanoma-classification/discussion/url</code></p>\n<p>Anyway, Thanks!</p>",
      "rawMarkdown": "Thanks for sharing @vaillant.\n\nI clicked your github link, it is go to this:\n`kaggle.com/c/siim-isic-melanoma-classification/discussion/url`\n\nAnyway, Thanks!",
      "votes": null
    },
    {
      "id": "987617",
      "postDate": "08/27/2020 11:53:34",
      "content": "<p>Sorry about that, I fixed the link.</p>",
      "rawMarkdown": "Sorry about that, I fixed the link.",
      "votes": null
    },
    {
      "id": "987619",
      "postDate": "08/27/2020 11:54:21",
      "content": "<p>I wanted to modularize my code as much as possible and breaking it into the parts you see in <code>src/factory</code> made the most sense to me.</p>",
      "rawMarkdown": "I wanted to modularize my code as much as possible and breaking it into the parts you see in `src/factory` made the most sense to me.",
      "votes": null
    },
    {
      "id": "1004653",
      "postDate": "09/09/2020 22:34:59",
      "content": "<p>Congratulations on the 2nd place!</p>",
      "rawMarkdown": "Congratulations on the 2nd place!",
      "votes": null
    },
    {
      "id": "1140886",
      "postDate": "01/06/2021 10:47:45",
      "content": "<p>It is good!<br>\nWhat is the key ideas of solution?</p>",
      "rawMarkdown": "It is good!\nWhat is the key ideas of solution?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 986325,
      "author_name": "heyytanay",
      "author_url": "",
      "post_date": "08/26/2020 11:32:19",
      "content": "<p>Thank you <a href=\"https://www.kaggle.com/vaillant\" target=\"_blank\">@vaillant</a> for sharing! Was looking for top solutions! Also congrats!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 987103,
      "author_name": "alanchn31",
      "author_url": "",
      "post_date": "08/27/2020 02:48:43",
      "content": "<p>Thank you for sharing <a href=\"https://www.kaggle.com/vaillant\" target=\"_blank\">@vaillant</a> , I want to post my code on github as well. How did you learn to structure your code like this? Any resources you would recommend to help with structuring code? Thank you, will greatly appreciate it</p>",
      "votes": null,
      "replies": [
        {
          "id": 987619,
          "author_name": "vaillant",
          "author_url": "",
          "post_date": "08/27/2020 11:54:21",
          "content": "<p>I wanted to modularize my code as much as possible and breaking it into the parts you see in <code>src/factory</code> made the most sense to me.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 987116,
      "author_name": "hsinwenchang",
      "author_url": "",
      "post_date": "08/27/2020 03:03:38",
      "content": "<p>Thank you for sharing <a href=\"https://www.kaggle.com/vaillant\" target=\"_blank\">@vaillant</a> 👍!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 987188,
      "author_name": "piantic",
      "author_url": "",
      "post_date": "08/27/2020 04:38:55",
      "content": "<p>Thanks for sharing <a href=\"https://www.kaggle.com/vaillant\" target=\"_blank\">@vaillant</a>.</p>\n<p>I clicked your github link, it is go to this:<br>\n<code>kaggle.com/c/siim-isic-melanoma-classification/discussion/url</code></p>\n<p>Anyway, Thanks!</p>",
      "votes": null,
      "replies": [
        {
          "id": 987617,
          "author_name": "vaillant",
          "author_url": "",
          "post_date": "08/27/2020 11:53:34",
          "content": "<p>Sorry about that, I fixed the link.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1004653,
      "author_name": "alincijov",
      "author_url": "",
      "post_date": "09/09/2020 22:34:59",
      "content": "<p>Congratulations on the 2nd place!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1140886,
      "author_name": "zavodrobotov",
      "author_url": "",
      "post_date": "01/06/2021 10:47:45",
      "content": "<p>It is good!<br>\nWhat is the key ideas of solution?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "986307": "Hi Kagglers,\n\nThe source code for my solution is available here: [https://github.com/i-pan/kaggle-melanoma](https://github.com/i-pan/kaggle-melanoma). I apologize in advance for its messiness and lack of comments. I need to work on best practices for coding. 😂\n\nThere are instructions to train/run inference on the competition data. Inevitably there will be errors when some of you try to run it, so please open issue and I will try to address it in a timely manner. \n\nThis is also the code I've been using over the past 6 months or so for Kaggle. Hopefully its organization can help people create their own Kaggle pipelines. My focus was on modularity and reproducibility for rapid experimentation. In particular, I started using YAML config files for each experiment I run to keep track of what I was doing.",
    "986325": "Thank you @vaillant for sharing! Was looking for top solutions! Also congrats!",
    "987103": "Thank you for sharing @vaillant , I want to post my code on github as well. How did you learn to structure your code like this? Any resources you would recommend to help with structuring code? Thank you, will greatly appreciate it",
    "987116": "Thank you for sharing @vaillant 👍!",
    "987188": "Thanks for sharing @vaillant.\n\nI clicked your github link, it is go to this:\n`kaggle.com/c/siim-isic-melanoma-classification/discussion/url`\n\nAnyway, Thanks!",
    "987617": "Sorry about that, I fixed the link.",
    "987619": "I wanted to modularize my code as much as possible and breaking it into the parts you see in `src/factory` made the most sense to me.",
    "1004653": "Congratulations on the 2nd place!",
    "1140886": "It is good!\nWhat is the key ideas of solution?"
  },
  "source": "meta"
}