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 supported8 Subscribers
Add a CodeTriage badge to deepspeed
Help out
- Issues
- [REQUEST] implementing interleaved 1F1B in pipeline parallelism
- [REQUEST] How to use int8 quantization inference without training?
- Exception: Current loss scale already at minimum - cannot decrease scale anymore. Exiting run.
- [BUG] misc false: parameter object has no attribute "comm"
- Compression training: when quantizer_kernel is enabled, the check target of "rounding" param should be "nearest" instead of "nearest_neighbor"
- [BUG] In step3, a runtime error will be thrown when inference_tp_size>1
- [BUG]GPU resources cannot be released properly when DeepSpeed is canceled. GPU A100 40G Driver Version: 535.54.03 CUDA Version: 12.2
- [BUG] BFloat16 training accuracy issue with zero 0/1 (the more rank, the less accuracy) in CIFAR10
- [BUG] Handle dangling partitioned parameters
- [BUG]Zero++ quantizer unsupport BFloat16
- Docs
- Python not yet supported