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
- Divide SequenceTiledCompute's incoming gradient out of place
- [BUG] AutoTP multi-query GPTBigCode training fails at TP > 1
- Fix AutoTP multi-query GPTBigCode training at TP > 1
- [ZeRO] Keep the offload overflow flag on the device
- [AffineIR] Calibrate affine transfer links per path (Plan analysis & calibration Part-2)
- [AffineIR] Analyze affine transfer batches before execution (Plan analysis & calibration Part-1)
- Skip redundant Stage-0 gradient reduction at DP=1
- Skip blank and comment lines in .deepspeed_env
- ZeRO: apply the clip coefficient at full precision to bf16/fp16 gradients
- ZeRO-3 partial offload: fix LR schedules and checkpoints of the device sub-groups
- Docs
- Python not yet supported