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
- Make PrefixedDataset satisfy the Mapping contract
- Use non-deprecated API for GPU gather
- Prefer all_gather_single over deprecated all_gather_into_tensor
- Honor scheduler adjustment during gradient accumulation
- docs: fix typo 'ore details' → 'more details' in CONTRIBUTING.md
- Clamp skipped dataloader lengths at zero
- Align `Accelerator.clip_grad_norm_` with Transformers
- compile_regions runs the uncompiled model when the root module has an accelerate hook
- Stop accelerate launch from forcing torch.compile dynamic=False
- Fix unwrap_model(keep_fp32_wrapper=False) breaking hooked models
- Docs
- Python not yet supported