4 ms·
onnx on the web has the most models available and can use webgpu which is available everywhere. Huggingface’s transformers.js uses it. And I use that for https
by sroussey 10mo ago
onnx on the web has the most models available and can use webgpu which is available everywhere.
Huggingface’s transformers.js uses it. And I use that for https://workglow.dev https://workglow.dev (also tensorflow mediapipe though that is using wasm).
I don’t think webnn has gone anywhere and is too restrictive.
- ekzhang 10mo agoSince ONNX is just a model data format, you can actually parse and run ONNX files in jax-js as well. Here’s an example of running DETR ResNet-50 from Xenova’s transformers.js checkpoint in jax-js https://jax-js.com/detr-resnet-50 https://jax-js.com/detr-resnet-50 I don’t think I intend to support everything in ONNX right now, especially quant/dequant, but eventually it would be interesting to see if we can help accelerate transformers.js with a jax-js backend + goodies like kernel fusion jax-js is more trying to explore being an ML research library, rather than ONNX which is a runtime for exported models