{
  "id": 203351,
  "title": "GPU/TPU accelerated Mel-spectrogram transformation",
  "url": "/competitions/rfcx-species-audio-detection/discussion/203351",
  "author_name": "Alex",
  "post_date": "2020-12-14T22:15:42.601000",
  "votes": 3,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I apologize if I've missed it, but I haven't yet stumbled upon a GPU/TPU implementation of the mel-spec transformation. Hence, I'm sharing one <a href=\"https://www.kaggle.com/akensert/rfcx-gpu-tpu-accelerated-mel-spectrogram\" target=\"_blank\">here</a>.</p>\n<p>A complete iteration over all training examples (4727 x 5 sec audio clips) in ~20-25 seconds.</p>",
  "messages": [
    {
      "id": 1112783,
      "postDate": "2020-12-14T22:15:42.600Z",
      "content": "<p>I apologize if I've missed it, but I haven't yet stumbled upon a GPU/TPU implementation of the mel-spec transformation. Hence, I'm sharing one <a href=\"https://www.kaggle.com/akensert/rfcx-gpu-tpu-accelerated-mel-spectrogram\" target=\"_blank\">here</a>.</p>\n<p>A complete iteration over all training examples (4727 x 5 sec audio clips) in ~20-25 seconds.</p>",
      "rawMarkdown": "I apologize if I've missed it, but I haven't yet stumbled upon a GPU/TPU implementation of the mel-spec transformation. Hence, I'm sharing one [here](https://www.kaggle.com/akensert/rfcx-gpu-tpu-accelerated-mel-spectrogram).\n\nA complete iteration over all training examples (4727 x 5 sec audio clips) in ~20-25 seconds.",
      "votes": 3
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1112783": "I apologize if I've missed it, but I haven't yet stumbled upon a GPU/TPU implementation of the mel-spec transformation. Hence, I'm sharing one [here](https://www.kaggle.com/akensert/rfcx-gpu-tpu-accelerated-mel-spectrogram).\n\nA complete iteration over all training examples (4727 x 5 sec audio clips) in ~20-25 seconds."
  }
}