{
  "id": 167795,
  "title": "[Info]: PyTorch, GhostNet-Pretrained Weights",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/167795",
  "author_name": "",
  "post_date": "2020-07-17T20:34:29.146860900Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Paper: <a href=\"https://arxiv.org/abs/1911.11907\">https://arxiv.org/abs/1911.11907</a>, CVPR 2020</p>\n\n<hr>\n\n<p><strong>GhostNet</strong>:\n - <a href=\"https://www.kaggle.com/ipythonx/ghostnetpretrained\">https://www.kaggle.com/ipythonx/ghostnetpretrained</a></p>\n\n<p><strong>How to import:</strong></p>\n\n<p>```\nsys.path.insert(0, \"/kaggle/input/ghostnetpretrained\")\nfrom ghost_net import ghost_net</p>\n\n<p>net = ghost_net()\nnet.load_state_dict(torch.load(...)) # weight location\nnet.classifier[4] = nn.Linear(1280, num_classes).cuda()\n```</p>",
  "messages": [
    {
      "id": "933623",
      "postDate": "07/17/2020 20:34:29",
      "content": "<p>Paper: <a href=\"https://arxiv.org/abs/1911.11907\">https://arxiv.org/abs/1911.11907</a>, CVPR 2020</p>\n\n<hr>\n\n<p><strong>GhostNet</strong>:\n - <a href=\"https://www.kaggle.com/ipythonx/ghostnetpretrained\">https://www.kaggle.com/ipythonx/ghostnetpretrained</a></p>\n\n<p><strong>How to import:</strong></p>\n\n<p>```\nsys.path.insert(0, \"/kaggle/input/ghostnetpretrained\")\nfrom ghost_net import ghost_net</p>\n\n<p>net = ghost_net()\nnet.load_state_dict(torch.load(...)) # weight location\nnet.classifier[4] = nn.Linear(1280, num_classes).cuda()\n```</p>",
      "rawMarkdown": "Paper: https://arxiv.org/abs/1911.11907, CVPR 2020\n\n---\n**GhostNet**:\n - https://www.kaggle.com/ipythonx/ghostnetpretrained\n\n**How to import:**\n\n```\nsys.path.insert(0, \"/kaggle/input/ghostnetpretrained\")\nfrom ghost_net import ghost_net\n\nnet = ghost_net()\nnet.load_state_dict(torch.load(...)) # weight location\nnet.classifier[4] = nn.Linear(1280, num_classes).cuda()\n```",
      "votes": null
    },
    {
      "id": "934424",
      "postDate": "07/18/2020 12:30:31",
      "content": "<p>Have you used it?</p>",
      "rawMarkdown": "Have you used it?",
      "votes": null
    },
    {
      "id": "934602",
      "postDate": "07/18/2020 15:21:29",
      "content": "<p><a href=\"/doanquanvietnamca\">@doanquanvietnamca</a> \nin time: <a href=\"https://www.kaggle.com/ipythonx/training-cv-melanoma-starter-ghostnet?scriptVersionId=39019060\">[Training CV] Melanoma Starter. GhostNet</a>\nbased on <a href=\"https://www.kaggle.com/shonenkov\">Alex</a> kernel. </p>",
      "rawMarkdown": "doanquanvietnamca \nin time: [[Training CV] Melanoma Starter. GhostNet](https://www.kaggle.com/ipythonx/training-cv-melanoma-starter-ghostnet?scriptVersionId=39019060)\nbased on [Alex](https://www.kaggle.com/shonenkov) kernel.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 934424,
      "author_name": "doanquanvietnamca",
      "author_url": "",
      "post_date": "07/18/2020 12:30:31",
      "content": "<p>Have you used it?</p>",
      "votes": null,
      "replies": [
        {
          "id": 934602,
          "author_name": "ipythonx",
          "author_url": "",
          "post_date": "07/18/2020 15:21:29",
          "content": "<p><a href=\"/doanquanvietnamca\">@doanquanvietnamca</a> \nin time: <a href=\"https://www.kaggle.com/ipythonx/training-cv-melanoma-starter-ghostnet?scriptVersionId=39019060\">[Training CV] Melanoma Starter. GhostNet</a>\nbased on <a href=\"https://www.kaggle.com/shonenkov\">Alex</a> kernel. </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "933623": "Paper: https://arxiv.org/abs/1911.11907, CVPR 2020\n\n---\n**GhostNet**:\n - https://www.kaggle.com/ipythonx/ghostnetpretrained\n\n**How to import:**\n\n```\nsys.path.insert(0, \"/kaggle/input/ghostnetpretrained\")\nfrom ghost_net import ghost_net\n\nnet = ghost_net()\nnet.load_state_dict(torch.load(...)) # weight location\nnet.classifier[4] = nn.Linear(1280, num_classes).cuda()\n```",
    "934424": "Have you used it?",
    "934602": "doanquanvietnamca \nin time: [[Training CV] Melanoma Starter. GhostNet](https://www.kaggle.com/ipythonx/training-cv-melanoma-starter-ghostnet?scriptVersionId=39019060)\nbased on [Alex](https://www.kaggle.com/shonenkov) kernel."
  },
  "source": "meta"
}