Add webgpu recipes for qwen3.5 models

[apsonawane]

Add webgpu recipes for Qwen3.5 models. Export the model using this command:
python optimize.py --config-dir webgpu --device webgpu

[Copilot]

Pull request overview

This PR adds WebGPU export/optimization “recipes” for Qwen3.5 multimodal models (0.8B/2B/4B/9B) so they can be exported via the existing builtin/optimize.py flow using Olive and run with the WebGpuExecutionProvider.

Changes:

• Add builtin/webgpu/{text,vision,embedding}.json Olive pipelines for Qwen3.5-0.8B/2B/4B/9B targeting WebGpuExecutionProvider.
• Extend each model’s builtin/optimize.py to accept --device webgpu and emit WebGPU provider options into generated runtime config.

Reviewed changes

Copilot reviewed 16 out of 16 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
Qwen-Qwen3.5-9B/builtin/webgpu/vision.json	New WebGPU vision encoder Olive pipeline targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-9B/builtin/webgpu/text.json	New WebGPU text decoder ModelBuilder pipeline (INT4) targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-9B/builtin/webgpu/embedding.json	New WebGPU embedding/export pipeline (FP16) targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-9B/builtin/optimize.py	Add webgpu device option and provider-options patching path.
Qwen-Qwen3.5-4B/builtin/webgpu/vision.json	New WebGPU vision encoder Olive pipeline targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-4B/builtin/webgpu/text.json	New WebGPU text decoder ModelBuilder pipeline (INT4) targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-4B/builtin/webgpu/embedding.json	New WebGPU embedding/export pipeline (FP16) targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-4B/builtin/optimize.py	Add webgpu device option and provider-options patching path.
Qwen-Qwen3.5-2B/builtin/webgpu/vision.json	New WebGPU vision encoder Olive pipeline targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-2B/builtin/webgpu/text.json	New WebGPU text decoder ModelBuilder pipeline (INT4) targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-2B/builtin/webgpu/embedding.json	New WebGPU embedding/export pipeline (FP16) targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-2B/builtin/optimize.py	Add webgpu device option and provider-options patching path.
Qwen-Qwen3.5-0.8B/builtin/webgpu/vision.json	New WebGPU vision encoder Olive pipeline targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-0.8B/builtin/webgpu/text.json	New WebGPU text decoder ModelBuilder pipeline (INT4) targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-0.8B/builtin/webgpu/embedding.json	New WebGPU embedding/export pipeline (FP16) targeting WebGpuExecutionProvider.
Qwen-Qwen3.5-0.8B/builtin/optimize.py	Add webgpu device option and provider-options patching path.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.