ea98b39996
This is something tf.keras will do. I'm not totally sure about dtype, we could only save it if it diverges from the global policy. But when it is explicitly set on the layer, it is probably important to persist. |
||
|---|---|---|
| benchmarks | ||
| examples | ||
| integration_tests | ||
| keras_core | ||
| shell | ||
| .gitignore | ||
| demo_functional.py | ||
| jax_integration_test.py | ||
| jax_training_scratchpad.py | ||
| LICENSE | ||
| pip_build.py | ||
| pyproject.toml | ||
| README.md | ||
| requirements.txt | ||
| setup.cfg | ||
| setup.py | ||
| tf_integration_test.py | ||
Keras Core: a new multi-backend Keras
Keras Core is a new multi-backend implementation of the Keras API, with support for TensorFlow, JAX, and PyTorch.
Backwards compatibility
Keras Core is intend to work as a drop-in replacement for tf.keras (when using the TensorFlow backend).
In addition, Keras models can consume datasets in any format, regardless of the backend you're using: you can train your models with your existing tf.data.Dataset pipelines or Torch DataLoaders.
Why use Keras Core?
- Write custom components (e.g. layers, models, metrics) that you can move across framework boundaries.
- Make your code future-proof by avoiding framework lock-in.
- As a PyTorch user: get access to power of Keras, at last!
- As a JAX user: get access to a fully-featured, battle-tested modeling and training library.
Credits
TODO