It seems not many people know about it today, but the people I know who've tried it love it.
Maybe we can bring the Keras API to JAX as an extension in the future 
Show this thread
JAX is a really fun low-level library for ML hacking: https://github.com/google/jax It's basically Numpy with gradients. And it can compile to XLA, for strong GPU/TPU acceleration. It's an ideal fit for researchers who want maximum flexibility when implementing new ideas from scratch.
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Pytorch === JAX
-
Its more like: TF + Pytorch + Numpy = Jax
End of conversation
New conversation -
-
-
Differentiable programming? https://julialang.org/blog/2018/12/ml-language-compiler …
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
What's the main difference w.r.t. PyTorch ?
-
JIT compilation runs all operations at once, pytorch runs them in sequence. Reuses the exact same numpy API.
End of conversation
New conversation -
-
-
So why not use Pytorch
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
A Keras API would be great
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Could you elaborate its differences with PyTorch?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.