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 supported9 Subscribers
Add a CodeTriage badge to deepspeed
Help out
- Issues
- need an an easier way to activate memory usage debug
- optimizer.step and lr_scheduler.step order
- Training accuracy heavily impacted when using activation checkpointing
- Improved checkpoint load/save and state_dict interfaces
- zero_to_fp32.py drops buffers
- Severe performance degradation with deepspeed > v.0.3.12
- Enhancement: Clean up zero_to_fp32.py
- Some Problems with ZeRO Optimizations for FP16 Training
- subprocess.CalledProcessError: Command '['/home/wushupei1997/anaconda3/envs/torch/bin/python', '-u','main2.py', '--local_rank=0', '--deepspeed_config', 'ds_config.json']' returned non-zero exit status1.
- how to use torch.save() to save my model
- Docs
- Python not yet supported