Add vLLM build script for vLLM v0.25.1

[Rukhaiya2004]

• Have you checked and followed all the points mention in the [CONTRIBUTING.MD]

• Have you validated script on UBI 9 container

• [x ] Did you run the script(s) on fresh container with set -e option enabled and observe success ?

• [NA] Did you have Legal approvals for patch files ?

[Rukhaiya2004]

Hi @cdeepali,

The wheel build and offline inference test completed successfully on my local setup.

From the CI logs, it looks like the LPAR isn't detecting any available CPU cores (core ids=[]), which leads to the OMP_NUM_THREADS error during the test. This seems to be an environment/configuration issue on the CI system rather than an issue with the build script.

Local successful run

Verifying vLLM installation...
0.25.1
vLLM import OK

---------------- Testing installed vLLM ----------------

Generated Outputs:
------------------------------------------------------------
Prompt:    'Hello, my name is'
Output:    'Mike who also lives in the US and was raised in California...'

Prompt:    'The president of the United States is'
Output:    'back in the White House...'

Prompt:    'The capital of France is'
Output:    'home to the largest number of millionaires...'

Prompt:    'The future of AI is'
Output:    'coming, people are already making the connection...'

===============================================================
vLLM v0.25.1 BUILD AND TEST COMPLETED SUCCESSFULLY
===============================================================

[cdeepali]

Merged the PR.