deepspeed
https://github.com/microsoft/deepspeed
Python
DeepSpeed is a deep learning optimization library that makes distributed training easy, efficient, and effective.
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 supported16 Subscribers
Add a CodeTriage badge to deepspeed
Help out
- Issues
- Fix zero 1 and 2 CPU-offloaded gradient norm
- [REQUEST] Add Trackio as a New Backend for Experiment Monitoring
- Fix overlap-comm buffer lifetimes
- [BUG] It seems that SuperOfflload is unable to overlap the CPU (param update) and GPU (backwawrd).
- Support NexusQuant KV cache compression for memory reduction
- [Blog] Muon Optimizer Support in DeepSpeed
- [BUG][ZeRO-3 / BF16] No public API to reassemble optimizer moments (exp_avg, exp_avg_sq) from sharded checkpoint files post-hoc
- Fix DeepCompile+Z3 on PyTorch v2.9/2.10
- Add Gram Newton-Schulz orthogonalization for Muon optimizer
- Refactor/torch autocast encapsulate global state
- Docs
- Python not yet supported