accelerate
https://github.com/huggingface/accelerate
Python
🚀 A simple way to train and use PyTorch models with multi-GPU, TPU, mixed-precision
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported1 Subscribers
Add a CodeTriage badge to accelerate
Help out
- Issues
- Fix DistributedType documentation
- Fix: Adds missing model_has_params4bit guard to match the existing guard pattern in the same function.
- fix: use key-based matching in fsdp2_load_full_state_dict
- [FSDP2] auto-exclude incompatible Params4bit from fully_shard to prevent silent QLoRA corruption
- fix segmentation fault
- [Bug] FSDP2 mixed-precision upcast to fp32 is a silent no-op since v1.13.0
- Fix TrackioTracker.log() ignoring step parameter
- Prevent double-wrapping models in prepare_model()
- Remove is_torch_xpu_tensor
- loading weight is so slow for fsdp2
- Docs
- Python not yet supported