RVM × LiteRT.js

Browser matting,
on the GPU.

Model not loaded

Accelerator Input size

Load the model, then try the sample video or your camera.

—FPS

—inference ms

—requested backend

512 × 288model input

Runtime log