mask utils在glm系列会少算stop token
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 8.5k
- Forks
- 1.3k
- Avg merge
- 5h 36m
- Merged PRs (30d)
- 22
Description
glm系列的stop token是<|observation|>（以及一些其他的）会出现在user轮的开头而不是assistant轮的结尾，导致这个token没算loss
针对glm47还有个问题，似乎</think> token的loss不会算进去
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.
Research direction
No file or test is named. Start by locating the GLM-specific mask utilities and tracing how stop tokens at the beginning of the user turn and <|think>|? are selected for loss; verify the related GLM47 case. Done means these tokens are included in loss accounting for the affected GLM variants, with regression coverage for both cases.
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
- 35/100