{
  "id": 212601,
  "title": "EfficientNet Noisy Student Weights for PyTorch [For non timm users]",
  "url": "/competitions/cassava-leaf-disease-classification/discussion/212601",
  "author_name": "Kaushal Shah",
  "post_date": "2021-01-19T13:39:13.684000",
  "votes": 6,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I've converted the EfficientNet Noisy Student Weights using this <a href=\"https://github.com/lukemelas/EfficientNet-PyTorch/blob/master/tf_to_pytorch/convert_tf_to_pt/load_tf_weights.py\" target=\"_blank\">script</a>:  and create a Kaggle Dataset: <a href=\"https://www.kaggle.com/kaushal2896/effnet-noisy-student-pytorch\" target=\"_blank\">https://www.kaggle.com/kaushal2896/effnet-noisy-student-pytorch</a></p>\n<p>I know \"timm\" package already provides this but if you are non timm PyTorch user, this might help you as NS weights work slightly better in this competition. </p>\n<p>Here is the colab notebook: <a href=\"https://colab.research.google.com/drive/19_mkoOdHt35X4o8EotOUORtXZTVRql9P\" target=\"_blank\">https://colab.research.google.com/drive/19_mkoOdHt35X4o8EotOUORtXZTVRql9P</a></p>",
  "messages": [
    {
      "id": 1159814,
      "postDate": "2021-01-19T13:39:13.683Z",
      "content": "<p>I've converted the EfficientNet Noisy Student Weights using this <a href=\"https://github.com/lukemelas/EfficientNet-PyTorch/blob/master/tf_to_pytorch/convert_tf_to_pt/load_tf_weights.py\" target=\"_blank\">script</a>:  and create a Kaggle Dataset: <a href=\"https://www.kaggle.com/kaushal2896/effnet-noisy-student-pytorch\" target=\"_blank\">https://www.kaggle.com/kaushal2896/effnet-noisy-student-pytorch</a></p>\n<p>I know \"timm\" package already provides this but if you are non timm PyTorch user, this might help you as NS weights work slightly better in this competition. </p>\n<p>Here is the colab notebook: <a href=\"https://colab.research.google.com/drive/19_mkoOdHt35X4o8EotOUORtXZTVRql9P\" target=\"_blank\">https://colab.research.google.com/drive/19_mkoOdHt35X4o8EotOUORtXZTVRql9P</a></p>",
      "rawMarkdown": "I've converted the EfficientNet Noisy Student Weights using this [script](https://github.com/lukemelas/EfficientNet-PyTorch/blob/master/tf_to_pytorch/convert_tf_to_pt/load_tf_weights.py):  and create a Kaggle Dataset: https://www.kaggle.com/kaushal2896/effnet-noisy-student-pytorch\n\nI know \"timm\" package already provides this but if you are non timm PyTorch user, this might help you as NS weights work slightly better in this competition. \n\nHere is the colab notebook: https://colab.research.google.com/drive/19_mkoOdHt35X4o8EotOUORtXZTVRql9P",
      "votes": 6
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1159814": "I've converted the EfficientNet Noisy Student Weights using this [script](https://github.com/lukemelas/EfficientNet-PyTorch/blob/master/tf_to_pytorch/convert_tf_to_pt/load_tf_weights.py):  and create a Kaggle Dataset: https://www.kaggle.com/kaushal2896/effnet-noisy-student-pytorch\n\nI know \"timm\" package already provides this but if you are non timm PyTorch user, this might help you as NS weights work slightly better in this competition. \n\nHere is the colab notebook: https://colab.research.google.com/drive/19_mkoOdHt35X4o8EotOUORtXZTVRql9P"
  }
}