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 supported21 Subscribers
Add a CodeTriage badge to deepspeed
Help out
- Issues
- [MoE] Add AutoEP routing replay
- Forward the per-head Muon tag at every muon_update call site
- Include the expert gradients in the unfused fp16 clip norm
- Add end-to-end BF16 stage 1 tests for zero and invalid gradient norms
- fix: preserve CUDA architecture variant suffixes
- Stop top-2 MoE gating from sampling the second expert in eval
- Clear the gradient ZenFlow just offloaded, not param.grad
- feat(rollout): segment-based kernel injection for hybrid-engine greedy decode
- fix: skip None buffers in AutoTP Loading.load_buffer (checkpoint load with nn.InstanceNorm1d)
- [BUG] Skip None buffers in AutoTP checkpoint loading
- Docs
- Python not yet supported