feat: Add GPU support to experimental JAX inference framework #272
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This commit introduces GPU support for the JAX-based experimental inference framework located in
experimental/jax
.Key changes include:
experimental/jax/requirements.txt
to usejax[cuda-pip]
allowing JAX to utilize NVIDIA GPUs.experimental/jax/inference/parallel/mesh.py
to correctly handle GPU devices during mesh creation, ensuring robust platform detection alongside existing TPU support.experimental/jax/inference/runtime/offline_inference.py
correctly usesjax.devices()
and is compatible with the new GPU handling in the mesh creation logic.experimental/jax/README.md
to include instructions for setting up JAX with GPU support and to reflect that NVIDIA GPUs are now a supported backend.experimental/jax/inference/entrypoint/run_gpu_test.py
and instructions for you to verify GPU functionality with a small number of prompts.These changes allow you, if you have compatible NVIDIA GPUs and CUDA setups, to run the experimental JAX inference framework, expanding its usability beyond TPUs.