deepspeedai / deepspeedai/DeepSpeed

[BUG] RepeatingLoader may be invalid in the pipe stages neither the fist nor last

Open
#5,593 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

bug training
Dominant language
Python
Stars
43.1k
Forks
5k
Avg merge
4d 15h
Merged PRs (30d)
112

Description

trainer struked if num_train_steps is grater than num_iters( len(train_dataloader)//gradient_accumulation_steps)

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by tracing RepeatingLoader through the trainer and pipe-stage handling, then inspect how num_train_steps is compared with num_iters, defined as len(train_dataloader)//gradient_accumulation_steps. Reproduce the failure with num_train_steps greater than num_iters and verify that intermediate pipe stages remain valid and training completes without the reported failure.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
distributed-systems, machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.