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
- `sparse_categorical_crossentropy(from_logits=True)` returns NaN on the JAX and NumPy backends when a non-target logit is `-inf`
- `CutMix` docstring: `factor` is the range of the Beta concentration, not of the mixing weight, and a higher factor reduces patch-size variability
- UnitNormalization with a list `axis` causes a graph break in `torch.compile`
- ops.image.map_coordinates raises ZeroDivisionError with fill_mode="mirror" on a size-1 axis on the PyTorch backend
- Repeated sparse gradient indices corrupt Adam updates on TensorFlow
- Deduplicate TF IndexedSlices grads before Adam-style updates
- Loaded models with custom `compile()` wrongly report `compiled=True`, bypassing compile assertion and causing misleading errors
- `torch` and `numpy` `evaluate()` omit `_assert_compile_called("evaluate")`, leading to `AttributeError` / misleading `ValueError` on uncompiled models
- Accept list and tuple repeats in the JAX repeat op
- Fix R2Score for targets far from zero
- Docs
- Python not yet supported