3 ms·
Mlx-ONNX: Run your MLX models in the browser using WebGPU
- skryl 7mo agoI built mlx-onnx (as part of mlx-ruby), a standalone exporter that converts MLX graphs to ONNX. Web Demo: https://skryl.github.io/mlx-ruby/demo Repo: https://github.com/skryl/mlx-onnx It provides: - MLX callable -> ONNX export - Python API + native C++ API The goal is to make it easier to move MLX models into ONNX tooling (onnxruntime, validation, downstream deployment), while keeping export behavior testable and explicit. Quick example: import mlx.core as mx import mlx_onnx as mxonnx def f(x): return mx.exp(x + 1.0) x = mx.array([[1.0, 2.0]], dtype=mx.float32) mxonnx.export_onnx("model.onnx", f, x, model_name="demo", opset=18) If you try it and hit an issue, I’d love a repro.