Add hf_transfer to testing image

[mgoin]

Installs the hf_transfer package into the test image in the dockerfile so all the CUDA buildkite tests can have faster model downloads

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.
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 do one of these:

• Add ready label to the PR
• Enable auto-merge.

🚀

[simon-mo]

i believe you also need to turn it on via env var? just set HF_HUB_ENABLE_HF_TRANSFER=1 globally

[mgoin]

I thought the automatic enablement of the env var would be taken care of by this:

vllm/vllm/model_executor/model_loader/weight_utils.py

Lines 39 to 51 in a62bc01

	def enable_hf_transfer():
	"""automatically activates hf_transfer
	"""
	if "HF_HUB_ENABLE_HF_TRANSFER" not in os.environ:
	try:
	# enable hf hub transfer if available
	import hf_transfer # type: ignore # noqa
	huggingface_hub.constants.HF_HUB_ENABLE_HF_TRANSFER = True
	except ImportError:
	pass
	enable_hf_transfer()

But the fastcheck times are the same - I will try manually enabling.

[mergify]

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @mgoin.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

[DarkLight1337]

The models tests are significantly faster on average, so I think this is working now.