keras
https://github.com/keras-team/keras
Python
Deep Learning for humans
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported271 Subscribers
Add a CodeTriage badge to keras
Help out
- Issues
- RandomShear: boxes are sheared in normalized coordinates while the image is sheared in pixel coordinates
- Resizing with crop_to_aspect_ratio / pad_to_aspect_ratio: boxes use real-valued crop/pad geometry while the image uses integer-rounded geometry
- RandomRotation: boxes are rotated about a different center than the image (half-pixel offset)
- Give each user their own /tmp fallback dir for .keras
- Python scalars, `convert_to_tensor` and 64-bit dtypes give different dtypes on each backend and in eager and symbolic calls
- Float-valued ops do not follow one dtype rule for integer, bool and 16-bit inputs
- Losses and metrics round class ids, counts and probabilities in `bfloat16` and `float16`, and do not cast a non-float `y_pred` to `floatx`
- Integer and `bool` inputs give incorrect or backend-dependent results in contractions, `bool` arithmetic and `clip`
- Layers return a dtype different from their symbolic output, or fail, for integer and bool inputs and half-precision dtypes
- Backend ops do not canonicalize list and tuple arguments or convert Python inputs, so Functional models fail after a `.keras` reload
- Docs
- Python not yet supported