[Bugfix/CI] Fix test_whisper by setting task to "generate"

[tlrmchlsmth]

Test is broken on current main with the error:

ValueError: LLM.generate() is only supported for (conditional) generation models (XForCausalLM, XForConditionalGeneration). Your model supports the 'generate' runner, but is currently initialized for the 'transcription' runner. Please initialize vLLM using `--task generate`.

[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.

🚀

[tlrmchlsmth]

See https://buildkite.com/vllm/ci/builds/13489#019506c7-b139-4ffd-a561-76ee5f89e8fd/208-8973

[mgoin]

@NickLucche maybe I was hasty to have you make a new task.. This is fine for now but we should allow for offline generation while still enforcing online transcription. Or we should make an offline transcript interface

[NickLucche]

Fix here #13274