[bugfix] small fix logic issue

[reidliu41]

Model in vllm serve should only use positional arg.

$ cat /tmp/config.yaml
# config.yaml

model: meta-llama/Meta-Llama-3-8B-Instruct
host: "127.0.0.1"
port: 6379
uvicorn-log-level: "info"

vllm serve  --config /tmp/config.yaml
usage: vllm serve [model_tag] [options]
vllm serve: error: ambiguous option: --model could match --model-impl, --model-loader-extra-config


vllm serve meta-llama/Meta-Llama-3-8B-Instruct --config /tmp/config.yaml
usage: vllm serve [model_tag] [options]
vllm serve: error: ambiguous option: --model could match --model-impl, --model-loader-extra-config

--------------------------------------------
$ cat /tmp/config.yaml
# config.yaml

host: "127.0.0.1"
port: 6379
uvicorn-log-level: "info"

vllm serve meta-llama/Meta-Llama-3-8B-Instruct --config /tmp/config.yaml
INFO:     Started server process [12414]
INFO:     Waiting for application startup.
INFO:     Application startup complete.

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[DarkLight1337]

Hmm, I thought we wanted config.yaml to be able to supply model ID as well?

[reidliu41]

Since the positional arg for model id, seems will not get model id from somewhere...
I guess the change from: vllm serve and vllm serve --model, both worked, and then changed to only work for vllm serve

[hmellor]

Yeah this seems like a bug rather than something we should document

[reidliu41]

@DarkLight1337 sorry, the previous logic is not rigorous:

sys.argv[1:]
['serve', 'meta-llama/Meta-Llama-3-8B-Instruct', '--config', '/tmp/config.yaml']

'serve' not in sys.argv[1:] → False
'--help' not in sys.argv[1:] → True

if False and True → False

pls merge this one.