Can't run webgpu-video-background-removal example

[chmjkb]

System Info

• Apple Macbook M1 Pro
• Browser: Arc 131.0.6778.205
• macOS Sequoia 15.1
• transformers.js: 2.17.2

Environment/Platform

• Website/web-app
• Browser extension
• Server-side (e.g., Node.js, Deno, Bun)
• Desktop app (e.g., Electron)
• Other (e.g., VSCode extension)

Description

I tried to run the webgpu video background removal example and I've encountered two issues:

Uncaught SyntaxError: Unexpected token '<', "<!DOCTYPE "... is not valid JSON
    at JSON.parse (<anonymous>)
    at getModelJSON (@xenova_transformers.js?v=1e7f008e:14370:15)
    at async loadConfig (@xenova_transformers.js?v=1e7f008e:18834:14)
    at async PretrainedConfig.from_pretrained (@xenova_transformers.js?v=1e7f008e:18864:26)
    at async AutoModel.from_pretrained (@xenova_transformers.js?v=1e7f008e:21908:14)
    at async main.js:33:13

Which i managed to fix by setting env.allowLocalModels to false. However, then i encountered another error:

Uncaught Error: no available backend found. ERR: [wasm] TypeError: Cannot read properties of undefined (reading 'buffer')
    at resolveBackend (@xenova_transformers.js?v=1e7f008e:166:13)
    at async _InferenceSession.create (@xenova_transformers.js?v=1e7f008e:764:25)
    at async constructSession (@xenova_transformers.js?v=1e7f008e:19513:12)
    at async Promise.all (:5173/index 1)
    at async PreTrainedModel.from_pretrained (@xenova_transformers.js?v=1e7f008e:19900:14)
    at async AutoModel.from_pretrained (@xenova_transformers.js?v=1e7f008e:21924:14)
    at async main.js:33:13

Which is surprising, since i expected it to run on WebGPU instead of WASM.
I tried different versions of transformers.js but couldn't fix it though.
Thank you for creating this awesome library! :D

Reproduction

1. go into webgpu-video-background-removal dir
2. run yarn && yarn dev