{
  "id": 471424,
  "title": "Fine-tuned models for EEG classification",
  "url": "/competitions/hms-harmful-brain-activity-classification/discussion/471424",
  "author_name": "Andreas Bisiadis",
  "post_date": "2024-01-28T10:13:18.003000",
  "votes": 8,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hello Kagglers,</p>\n<p>After extensive trial and error - making the most out of my gpu quota - I created three datasets, including the weights of:<br>\n<a href=\"https://www.kaggle.com/datasets/andreasbis/resnet34d-weights\" target=\"_blank\">ResNet34d</a><br>\n<a href=\"https://www.kaggle.com/datasets/andreasbis/efficientnetb0-weights\" target=\"_blank\">EfficientNetB0</a><br>\n<a href=\"https://www.kaggle.com/datasets/andreasbis/efficientnetb1-for-eeg-classification-weights\" target=\"_blank\">EfficientNetB1</a></p>\n<p>The three models were trained exclusively on the kaggle-provided eeg spectrogram data for 9 epochs, using a 5-fold CV. The total training time (for all three models) is around 20 hours. </p>\n<p>🙏 Upvote my work if you find it useful!</p>",
  "messages": [
    {
      "id": 2623621,
      "postDate": "2024-01-28T10:13:18.003Z",
      "content": "<p>Hello Kagglers,</p>\n<p>After extensive trial and error - making the most out of my gpu quota - I created three datasets, including the weights of:<br>\n<a href=\"https://www.kaggle.com/datasets/andreasbis/resnet34d-weights\" target=\"_blank\">ResNet34d</a><br>\n<a href=\"https://www.kaggle.com/datasets/andreasbis/efficientnetb0-weights\" target=\"_blank\">EfficientNetB0</a><br>\n<a href=\"https://www.kaggle.com/datasets/andreasbis/efficientnetb1-for-eeg-classification-weights\" target=\"_blank\">EfficientNetB1</a></p>\n<p>The three models were trained exclusively on the kaggle-provided eeg spectrogram data for 9 epochs, using a 5-fold CV. The total training time (for all three models) is around 20 hours. </p>\n<p>🙏 Upvote my work if you find it useful!</p>",
      "rawMarkdown": "Hello Kagglers,\n\nAfter extensive trial and error - making the most out of my gpu quota - I created three datasets, including the weights of:\n[ResNet34d](https://www.kaggle.com/datasets/andreasbis/resnet34d-weights)\n[EfficientNetB0](https://www.kaggle.com/datasets/andreasbis/efficientnetb0-weights)\n[EfficientNetB1](https://www.kaggle.com/datasets/andreasbis/efficientnetb1-for-eeg-classification-weights)\n\nThe three models were trained exclusively on the kaggle-provided eeg spectrogram data for 9 epochs, using a 5-fold CV. The total training time (for all three models) is around 20 hours. \n\n🙏 Upvote my work if you find it useful!",
      "votes": 7
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2623621": "Hello Kagglers,\n\nAfter extensive trial and error - making the most out of my gpu quota - I created three datasets, including the weights of:\n[ResNet34d](https://www.kaggle.com/datasets/andreasbis/resnet34d-weights)\n[EfficientNetB0](https://www.kaggle.com/datasets/andreasbis/efficientnetb0-weights)\n[EfficientNetB1](https://www.kaggle.com/datasets/andreasbis/efficientnetb1-for-eeg-classification-weights)\n\nThe three models were trained exclusively on the kaggle-provided eeg spectrogram data for 9 epochs, using a 5-fold CV. The total training time (for all three models) is around 20 hours. \n\n🙏 Upvote my work if you find it useful!"
  }
}