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 supported17 Subscribers
Add a CodeTriage badge to deepspeed
Help out
- Issues
- [BUG] network crashes with stage=3 & dynamically-changed network
- Add explicit gradient_accumulation_dtype config
- [BUG]AttributeError: 'DeepSpeedEngine' object has no attribute 'quantizer'
- Consolidate norm computation logic
- Is "High-performance INT8 inference kernels" released?
- [BUG] Same train time with DeepSpeed (despite increased batch size)
- Can deepspeed be used to infer fairseq trained gpt models?
- [REQUEST] High CPU memory usage at initialization
- [BUG] Problem with optimizer update step
- [BUG] ZeRO/bf16 grad accumulation in bf16 needs higher precision accumulator
- Docs
- Python not yet supported