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
- Support for 5060GPU on Cuda 12.8 and python 3.11.11
- [BUG] ZeRO-3 with HuggingFace Transformers: Gathered Parameters in Forward Pass Not Released Promptly
- [BUG] Rank index out of range during BWD when sp_size < world_size
- compile error / linear_kernels_cuda.cu(127)
- [BUG] Significant Performance Difference between DeepSpeed's zero_stage=1 and zero_stage=2
- [BUG] Grad norm value differs significantly across DeepSpeed version
- [BUG] ZenFlow: loss values becomes NaN after 'update_interval' number of steps.
- Fix bf16 dtype mismatch in ZeRO-3 with zero_quantized_weights
- Enabling fp_quantizer extension unit tests on ROCm
- [BUG] DeepCompile in ZeRO-3 fails to do the forward pass
- Docs
- Python not yet supported