deepspeedai / deepspeedai/DeepSpeedExamples
The min_length setting force the model generate to max length, which produce repeated or nonsense result
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 6.8k
- Forks
- 1.1k
- Avg merge
- 2d 16h
- Merged PRs (30d)
- 1
Description
When i try to reproduce bloom, i meet the same problem:
"The min_length setting force the model generate to max length, which produce repeated or nonsense result."
fix ppo_trainer generate and scores calculation in stage 2
So i try to delete the "min_length setting", but i find the program can't continue to run at https://github.com/microsoft/DeepSpeedExamples/blob/8f8099a813f3b223d5df39e0c15c748de4eb1669/applications/DeepSpeed-Chat/training/step3_rlhf_finetuning/ppo_trainer.py#L105
Contributor guide
No contributing guide indexed for this repository
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.
Research direction
Start with applications/DeepSpeed-Chat/training/step3_rlhf_finetuning/ppo_trainer.py at lines 76 and 105, then reproduce the reported failure after removing the min_length setting. Compare the generation and score-calculation behavior with the referenced PR #347. Done should mean the training stage continues and generation no longer produces repeated or nonsensical results.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100