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 TorchXLA dynamo backend config option
- `send_to_device` ignores `skip_keys` for Mappings that have `.to()` method
- Fix send to device skip keys for mappings with `.to` method
- `pad_across_processes` turns `bool` tensors into `int64`, which breaks gather across ranks
- Fix FSDP2 cpu_ram_efficient_loading for models with persistent buffers
- Preserve RNG state during FSDP2 activation checkpointing
- Compile checkpointed repeated blocks across graph breaks
- Preserve shared modules in regional compilation
- Fix PYTHONPATH separator on Windows
- compile_regions doesn't compile gradient-checkpointed transformers layers that contain a graph break
- Docs
- Python not yet supported