deepspeedai / deepspeedai/DeepSpeedExamples
Why is the PPL so high in the beginning of Step-1 (SFT)?
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 6.8k
- Forks
- 1.1k
- Avg merge
- 2d 16h
- Merged PRs (30d)
- 1
Description
Why is the PPL here 4k when we are starting with a pretrained model?
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 the linked training log at training_log_output/opt-1.3b-globalBatchSize128.log, especially line 4, and inspect the Step-1 supervised fine-tuning setup that produced it. Determine why perplexity begins near 4k despite the pretrained model, then document the confirmed cause and reproduce or validate the explanation with the same training configuration.
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