Jiao

So chatted with @amogkam a bit offline, a few changes we need: 1) batch_format should be `BatchMapper` only thing, and base class `Preprocessor` should still have fallback paths that decides...

Failed tests are RLlib docs that irrelevant to this PR. ``` gym.error.NameNotFound: The environment `Pong` has been moved out of Gym to the package `ale-py`. Please install the package via...

Failed test are due to transient error on github (dask and hovorod on github 404) ``` (ray-overview/ray-libraries: line 25) broken https://github.com/dask/dask) - 404 Client Error: Not Found for url: https://github.com/dask/dask)...

Test failures are irrelevant, due to gRPC upgrade on ray client, which this PR did not touch.

Good findings, later you can also compare this with AIR's data loading such as our benchmark: https://sourcegraph.com/github.com/ray-project/ray/-/blob/release/air_tests/air_benchmarks/workloads/pytorch_training_e2e.py?L51 Note that AIR doesn't use pytorch dataloader and use `iter_torch_batches` instead, which will...

Hi @EnjianGong can you edit the title with component and make issue description in English ? This will help a lot to communicate with broader ray community.

keep alive

Thanks for the detailed instruction and fast response @gaogaotiantian :) I think this could work with Ray given Ray also pickles the function / class along with args, and we...

Hi @OrenLeung , i actually just chatted with Alpa team about the exact topic yesterday, and the latest status is I'm pushing to upstream ray collective for JAX in Ray...

Hi @ntlm1686 thanks for your interest and willingness to contribute ! So the most active user of ray collective is currently [Alpa](https://github.com/alpa-projects/alpa) that has both cupy and xla based implementations...