Creator of Keras confirmed that the new version comes out in a few days. Keras becomes multi-backend again with support for PyTorch, TensorFlow and JAX. Personally, I’m excited to be able to try JAX without having to deep dive into documentation and entire ecosystem. What about you?
One draw of keras that would get people to switch over would be how easy it is to model parallelism, but you’d need to get better mfu than fsdp/deepspeed, ideally competitive to megatron while being way easier and more flexible for people to switch