[Bugfix][Harmless] Fix hardcoded float16 dtype for model_is_embedding

[mgoin]

FIX #7561

On main you would see a concerning but harmless message about casting to float16 for models that are bfloat16:

vllm serve meta-llama/Meta-Llama-3.1-8B-Instruct
...
WARNING 08-15 21:06:16 config.py:1514] Casting torch.bfloat16 to torch.float16.
INFO 08-15 21:06:16 api_server.py:111] Multiprocessing frontend to use ipc:///tmp/5c13e2e7-a88a-4ffa-bb14-492eaa9851f8 for RPC Path.
INFO 08-15 21:06:16 api_server.py:122] Started engine process with PID 265910

This PR removes that by making the ModelConfig made in model_is_embedding just use dtype="auto":

vllm serve meta-llama/Meta-Llama-3.1-8B-Instruct
...
INFO 08-15 21:05:58 api_server.py:111] Multiprocessing frontend to use ipc:///tmp/79226d04-56b8-4839-a94e-8da715e35d1c for RPC Path.
INFO 08-15 21:05:58 api_server.py:122] Started engine process with PID 265174

This is just a cosmetic issue since this config isn't actually used for loading the model, just to check if the model is for embedding

[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 consists a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of default ones by unblocking the steps in your fast-check build on Buildkite UI.

Once the PR is approved and ready to go, please make sure to run full CI as it is required to merge (or just use auto-merge).

To run full CI, you can do one of these:

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

🚀

[DarkLight1337]

Thanks for the QoL improvement!