Please add "logprobs" param for vLLM itegration
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 39.5k
- Forks
- 4.8k
- PR merge metrics
- No merged PRs in 30d
Description
Please add "logprobs" params for vLLM integration, and also the ability to pass the param from OpenAI compatible API.
Refer to "logprobs" param in https://github.com/vllm-project/vllm/blob/main/vllm/sampling_params.py, vLLM supports "logprobs" params to output logprobs of the most likely possible generations, similar to OpenAI API. This function is necessary for classification tasks usually with a score threshold.
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 by tracing the OpenAI-compatible API request path into the vLLM integration, using vLLM's sampling_params.py as the parameter reference. Confirm how request parameters are forwarded and define done as accepting logprobs and returning the corresponding generation log probabilities for classification use cases.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- ai, api
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100