Announcement
Keras plans pluggable backends, with MLX and PaddlePaddle backends upcoming
A post summarizing the Keras community call says new backends will install as add-on libraries.
TLDR
A post summarizing the Keras community call outlines a pluggable backend plan, with MLX and PaddlePaddle backends upcoming. It says Keras is reducing the operations needed to ship a backend, adding backend-agnostic implementations of secondary operations and streamlining testing with a single harness. The post also notes new KerasHub models and a move from tf-text preprocessing to PyGrain.
Combined views
7.7K
1 Source, first seen ago