Adding corrected generation config for demo

[przepeck]

🛠 Summary

CVS-181828
Adding corrected generation config to demo to stop model generation after correct answer.

🧪 Checklist

• Unit tests added.
• The documentation updated.
• Change follows security best practices.
  ``

[Copilot]

Pull request overview

This PR adds a generation_config.json file for the VLM NPU demo to ensure the model stops generating after producing the correct answer (by setting proper EOS/pad token IDs). It also updates the demo README with instructions to download this config, and separately unpins dependency versions in the agentic AI demo requirements.

Changes:

• Added demos/vlm_npu/generation_config.json with EOS/BOS/pad token IDs for Phi-3.5-vision-instruct
• Updated demos/vlm_npu/README.md with a curl command to download the generation config
• Removed version pins from openai-agents and openai in demos/continuous_batching/agentic_ai/requirements.txt

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
demos/vlm_npu/generation_config.json	New generation config specifying BOS, EOS, and pad token IDs
demos/vlm_npu/README.md	Added curl command to download the generation config after model export
demos/continuous_batching/agentic_ai/requirements.txt	Removed version pins for openai-agents and openai

---

You can also share your feedback on Copilot code review. Take the survey.

[przepeck]

tests: https://ci.iotg.sclab.intel.com/job/ovmsc/job/OVMS_Validation/job/test/job/ovms-c_uat_test/641/