Add Phi-3 Vision ONNX example

[web3nomad]

This PR adds a new example demonstrating the usage of Microsoft's Phi-3 Vision model using ONNX Runtime.

The example implements:

• Multimodal inference combining vision and language processing
• Image preprocessing and text tokenization
• End-to-end example using multiple ONNX models

It showcases:

• Loading and utilizing vision, text embedding, and text generation models
• Combining vision and text embeddings
• Autoregressive text generation

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 53.20%. Comparing base (f915bca) to head (d043107).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #291   +/-   ##
=======================================
  Coverage   53.20%   53.20%           
=======================================
  Files          34       34           
  Lines        2327     2327           
=======================================
  Hits         1238     1238           
  Misses       1089     1089           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[web3nomad]

commits squashed and rebased onto latest main