Input Profile
Run Controls
Run the ORT-style inference profile to capture init, first output, throughput, and total latency for the active provider mode.
Execution Trace
No inference run yet.
AI WebGPU Lab Experiment
ORT WebGPU Readiness
`exp-ort-webgpu-baseline` now exposes a provider-readiness harness for ORT-Web style inference. It records provider metadata, fallback state, worker mode, and deterministic transformer-block timing before the real ONNX Runtime Web integration lands.
Use `?mode=fallback` to keep the same input profile while switching the provider metadata and latency profile to a Wasm fallback path.
Metrics
Environment
Activity Log
Schema-Aligned Result Draft
{
"status": "pending"
}
Provider Notes
- Default mode records a WebGPU provider profile.
- `?mode=fallback` records a Wasm fallback profile with the same batch and sequence budget.
- This readiness harness fixes the result contract before real ORT-Web assets and build steps are added.