Upgrade to vllm 0.6.0

[stikkireddy]

https://github.com/vllm-project/vllm/releases

• incorporate --num-scheduler-steps 8
• test openai tool spec
• test multi multi modal input support
• test tensor parallel for ViT models in multi modal llms
• supported tensor-parallel ViT models. [(also test the new xformers ]([VLM][Model] TP support for ViTs vllm-project/vllm#7186)
  • Siglip
  • Clip
  • Blip
  • Intern ViT

[stikkireddy]

upgraded to 0.6.0 new feature support remaining

[stikkireddy]

• will not support --num-scheduler-steps 8, too many limitations
• test openai tool spec (misunderstanding of release)
• test tensor parallel for ViT models in multi modal llms (tensor parallel issues in azure databricks)