Example not working on Chrome/Arc v.124(M1 Mac)

[justrach]

System Info

M1 Mac Sonoma Using the example code.

Environment/Platform

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

Description

Using the repository from : https://github.com/xenova/transformers.js/tree/v3/examples/webgpu-chat and "@xenova/transformers": "github:xenova/transformers.js#v3", it gets stuck on ..loading model..

On the huggingface demo here is the output on the console:

Meanwhile on the local page, here is how the output looks like:

Reproduction

1. Clone repo from https://github.com/xenova/transformers.js/tree/v3/examples/webgpu-chat
2. npm i and do npm run dev;
3. Go to url
4. Load Model
5. Model gets stuck on ...loading...

[xenova]

Duplicate of https://github.com/xenova/transformers.js/issues/748.

[xenova]

It should now work (commits: here and here)!

[justrach]

Thanks! Do you have a page where there's documentation on how to convert ONNX models to ONNX-Web models by any chance

[xenova]

Sure, we use Optimum, and you can find additional information there.

You can also use our helper conversion script for helping with quantization.

[justrach]

Sure, we use Optimum, and you can find additional information there.

You can also use our helper conversion script for helping with quantization.

Thanks Joshua!