loss mask question
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 8.5k
- Forks
- 1.3k
- Avg merge
- 5h 36m
- Merged PRs (30d)
- 22
Description
sample.tokens = prompt_tokens_ids + response_token_ids
sample.response_length = len(response_token_ids)
sample.response = response
sample.loss_mask = loss_masks
这里loss_masks只需要计算在(system+user_query)之后的response就可以了吗？不需要包括的[0]*len(prompt_tokens_ids)吗？
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
Start by reading the code that assigns sample.tokens, sample.response_length, sample.response, and sample.loss_mask. Confirm how the loss mask is consumed during post-training, then document whether prompt positions need explicit zero entries; done means the expected mask shape and coverage are stated clearly.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Documentation
- Difficulty
- 1/5
- Estimated time
- Under an hour
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100