Add `--max-model-len` param and description to `serve.vllm_worker`?
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 39.5k
- Forks
- 4.8k
- PR merge metrics
- No merged PRs in 30d
Description
Currently the `--max-model-len` can be passed to vllm through `kwarg`, but could it be added to the default param like `gpu-utilization-limit`? It is often needed when using models that can accept long contexts.
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 at the `serve.vllm_worker` entry point and locate the existing `gpu-utilization-limit` parameter definition. Check how keyword arguments are forwarded to vLLM, then add the `--max-model-len` option and its description alongside the existing default parameters. Confirm the option appears in the worker help output and can be passed through to vLLM.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- cli
- Issue type
- Feature
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100