deepspeedai / deepspeedai/DeepSpeed

[BUG] incorrectly loading layer_keys in deepspeed_checkpoint.py

Open Beginner friendly
#3,305 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

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

Description

Hello,

I'm referring to this function in deepspeed_checkpoint.py https://github.com/microsoft/DeepSpeed/blob/master/deepspeed/checkpoint/deepspeed_checkpoint.py#L256-L260

This function is to get layer_keys and it is not correct when number of layers is >= 100 because "layer_100" will be seen as "layer_10".

Best regards,
Trung

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 in deepspeed/checkpoint/deepspeed_checkpoint.py at lines 256-260, where layer_keys are loaded. Reproduce the issue with a checkpoint containing 100 or more layers and inspect how layer_100 is identified. Update the handling so all layer keys are distinguished correctly, then verify loading with such a checkpoint.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Bug
Difficulty
2/5
Estimated time
1-3 hours
Activity status
Active
Clarity
Clearly specified
Newbie friendliness
82/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.