deepspeedai / deepspeedai/DeepSpeed
[BUG] Handle dangling partitioned parameters
Open
@tjruwase is already working on this.
Since Jul 18, 2023.
bug
training
- Dominant language
- Python
- Stars
- 43.1k
- Forks
- 5k
- Avg merge
- 4d 15h
- Merged PRs (30d)
- 112
Description
Detect dangling partitioned parameters of a module, probably at the end of deepspeed.initialize().
# TODO Delay error checking for dangling partitioned parameters to post module init
# raise RuntimeError(f"not all parameters for {module_to_apply_fn_to.__class__.__name__}, "
# f"were zero params, is it possible that the parameters were "
# f"overwritten after they were initialized? "
# f"params: {[p for p in module_to_apply_fn_to.parameters(recurse=False)]} ")
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.