[Misc] Avoid calling unnecessary hf_list_repo_files for local model path

[Isotr0py]

• There will be error logs due to hf_list_repo_files calling when model repo is local:

INFO 02-16 13:46:11 __init__.py:190] Automatically detected platform cuda.
ERROR 02-16 13:46:11 config.py:102] Error retrieving file list: Repo id must use alphanumeric chars or '-', '_', '.', '--' and '..' are forbidden, '-' and '.' cannot start or end the name, max length is 96: '../Florence-2-base'., retrying 1 of 2
ERROR 02-16 13:46:13 config.py:100] Error retrieving file list: Repo id must use alphanumeric chars or '-', '_', '.', '--' and '..' are forbidden, '-' and '.' cannot start or end the name, max length is 96: '../Florence-2-base'.
INFO 02-16 13:46:13 config.py:2407] Downcasting torch.float32 to torch.float16.

• Avoid unnecessary hf_list_repo_files calling if the model repo is local so that we can prevent misleading error logs.

[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.

🚀

[ywang96]

@maxdebayser actually had the same PR #13292 but I feel like this PR is a bit cleaner?

[Isotr0py]

Oh, didn't notice that there has been a same PR, my bad! 😅

[maxdebayser]

@Isotr0py , no problem, thanks for also proposing a fix for this problem!

@ywang96 , I agree that this PR also fixes the issue. I think it comes down to personal style preferences, so either way is fine by me. My personal view here is that the code in config.py has to abstract three different locations for the files (remotely, in local cache or local path) and deal with the fact that the information we have might be incomplete (in offline mode we only know that we don't have the file in cache, not that it doesn't exist). I like to do the abstraction part only in the higher-level functions like get_sentence_transformers_tokenizer_config() and leave lower level functions like list_repo_files specialized on a single task, in this case looking up remote hf repos. But again, either way is fine I think.

[maxdebayser]

LGTM