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
- Update Sparse attention to use latest triton/tl.dot
- [BUG] Optimizer calculation seems to run on GPU although I set CPU optimizer options
- [BUG] Assertion error with HF gradient checkpointing option (param already assigned swap buffer)
- Zero Stage-2 Frozen Layers[BUG]
- [BUG] Unexpected caculations at backward pass with ZeRO-Infinity SSD offloading
- Issue with DeepSpeed Inference - Multiple Processes for Model Loading and Memory Allocation
- [PROBLEM] P2p recv waiting for data will cause other threads under the same process to be unable to perform any operations
- [BUG] DeepSpeed distributed training crashes with "TypeError: can't convert complex to float"
- [REQUEST] split zero3 checkpoint files into optim states and master weights
- [REQUEST] [zero_to_fp32.py] Allow user to specify the dtype (e.g. `torch.bfloat16`) in the output file
- Docs
- Python not yet supported