{
  "id": 122639,
  "title": "[Grapheme] LB LB 0.9259 Inference kernel and model kernels",
  "url": "/competitions/bengaliai-cv19/discussion/122639",
  "author_name": "",
  "post_date": "2019-12-21T16:30:51.052306800Z",
  "votes": 12,
  "comment_count": 1,
  "views": 0,
  "content": "<p>This is inference kernel of \n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-9259\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-9259</a></p>\n\n<p>previous kernel\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-8566\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-8566</a></p>\n\n<p>There were no Augmentation // train-valid split or any other techniques, it is a lot of potentials to achieve above +0.95 in the near future.</p>\n\n<p>I am sharing my kernels(to show step by step progress) to help someone study image classification and neural net.\nI hope it will help your study ;-)\n- ResNet-18\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning</a>\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-2\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-2</a>\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-3\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-3</a>\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-4\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-4</a>\n- Resnet-18 mix-up\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet18-mixup-naive-learning-1\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet18-mixup-naive-learning-1</a> (check version 2)</p>\n\n<ul>\n<li>BornoNet\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-1\">https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-1</a>\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-2\">https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-2</a>\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-3\">https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-3</a></li>\n</ul>",
  "messages": [
    {
      "id": "700216",
      "postDate": "12/21/2019 16:30:51",
      "content": "<p>This is inference kernel of \n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-9259\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-9259</a></p>\n\n<p>previous kernel\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-8566\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-8566</a></p>\n\n<p>There were no Augmentation // train-valid split or any other techniques, it is a lot of potentials to achieve above +0.95 in the near future.</p>\n\n<p>I am sharing my kernels(to show step by step progress) to help someone study image classification and neural net.\nI hope it will help your study ;-)\n- ResNet-18\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning</a>\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-2\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-2</a>\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-3\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-3</a>\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-4\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-4</a>\n- Resnet-18 mix-up\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-resnet18-mixup-naive-learning-1\">https://www.kaggle.com/hanjoonchoe/grapheme-resnet18-mixup-naive-learning-1</a> (check version 2)</p>\n\n<ul>\n<li>BornoNet\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-1\">https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-1</a>\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-2\">https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-2</a>\n<a href=\"https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-3\">https://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-3</a></li>\n</ul>",
      "rawMarkdown": "This is inference kernel of \nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-9259\n\nprevious kernel\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-8566\n\nThere were no Augmentation // train-valid split or any other techniques, it is a lot of potentials to achieve above +0.95 in the near future.\n\nI am sharing my kernels(to show step by step progress) to help someone study image classification and neural net.\nI hope it will help your study ;-)\n- ResNet-18\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-2\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-3\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-4\n- Resnet-18 mix-up\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet18-mixup-naive-learning-1 (check version 2)\n\n- BornoNet\nhttps://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-1\nhttps://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-2\nhttps://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-3",
      "votes": null
    },
    {
      "id": "700825",
      "postDate": "12/22/2019 17:00:23",
      "content": "<p>Thanks for sharing!!😄 👍 </p>",
      "rawMarkdown": "Thanks for sharing!!😄 👍",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 700825,
      "author_name": "mashlyn",
      "author_url": "",
      "post_date": "12/22/2019 17:00:23",
      "content": "<p>Thanks for sharing!!😄 👍 </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "700216": "This is inference kernel of \nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-9259\n\nprevious kernel\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-n-l-inference-lb-0-8566\n\nThere were no Augmentation // train-valid split or any other techniques, it is a lot of potentials to achieve above +0.95 in the near future.\n\nI am sharing my kernels(to show step by step progress) to help someone study image classification and neural net.\nI hope it will help your study ;-)\n- ResNet-18\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-2\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-3\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet-18-naive-learning-4\n- Resnet-18 mix-up\nhttps://www.kaggle.com/hanjoonchoe/grapheme-resnet18-mixup-naive-learning-1 (check version 2)\n\n- BornoNet\nhttps://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-1\nhttps://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-2\nhttps://www.kaggle.com/hanjoonchoe/grapheme-modified-bornonet-naive-learning-3",
    "700825": "Thanks for sharing!!😄 👍"
  },
  "source": "meta"
}