{
  "id": 206699,
  "title": "Mixed Precision Augmentation",
  "url": "/competitions/ranzcr-clip-catheter-line-classification/discussion/206699",
  "author_name": "Ameya",
  "post_date": "2020-12-26T03:51:03.530000",
  "votes": 1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>How can we use tf.keras.layers.experimental.preprocessing with mixed precision as while doing it, I get the following error:</p>\n<p>TypeError: Value passed to parameter 'images' has DataType bfloat16 not in list of allowed values: uint8, int32, int64, float16, float32, float64</p>",
  "messages": [
    {
      "id": 1126862,
      "postDate": "2020-12-26T03:51:03.530Z",
      "content": "<p>How can we use tf.keras.layers.experimental.preprocessing with mixed precision as while doing it, I get the following error:</p>\n<p>TypeError: Value passed to parameter 'images' has DataType bfloat16 not in list of allowed values: uint8, int32, int64, float16, float32, float64</p>",
      "rawMarkdown": "How can we use tf.keras.layers.experimental.preprocessing with mixed precision as while doing it, I get the following error:\n\nTypeError: Value passed to parameter 'images' has DataType bfloat16 not in list of allowed values: uint8, int32, int64, float16, float32, float64",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1126862": "How can we use tf.keras.layers.experimental.preprocessing with mixed precision as while doing it, I get the following error:\n\nTypeError: Value passed to parameter 'images' has DataType bfloat16 not in list of allowed values: uint8, int32, int64, float16, float32, float64"
  }
}