Config fixes for VLLMModel

[anton-l]

Just re-adding some tiny but useful features from the base model back to VLLM

LMK if you notice anything else!

[anton-l]

Important fix: the argument is indeed revision
https://github.com/vllm-project/vllm/blob/04139ade599eedd493ce8effcda7ceabb57f2fb5/vllm/transformers_utils/tokenizer.py#L92

[clefourrier]

Thanks for the update re-revision - not sure about why you're doing the next addition to the path howver?

[anton-l]

I noticed this pattern in other implementations, so it probably makes sense to standardize while we're at it

lighteval/src/lighteval/models/transformers/transformers_model.py

Lines 505 to 507 in fdb12f4

	tokenizer = AutoTokenizer.from_pretrained(
	model_name if tokenizer_name is None else tokenizer_name,
	revision=revision + (f"/{subfolder}" if subfolder is not None else ""),

Pinging @NathanHB for whether it's applicable to vllm

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[kzos]

@anton-l Hope this fixes the below error:
Despite having both vllm and ray installed.

ImportError: You are trying to use an VLLM model, for which you need vllm and ray, which are not
available in your environment. Please install them using pip, pip install vllm ray

[NathanHB]

hi ! did you try pip install lighteval[vllm] ?

[kzos]

@NathanHB in my case it was environment issue, conda/venv. resolved it. thanks