models

They're data-dependence graphs for a neural-network scheduling problem. Like this but way bigger to start with and then lowered to more detailed representations several times: https://netron.app/?url=https://github.com/onnx/models/raw/m... My home-grown layout engine can handle the 12k nodes for llama2 in its highest-level form in 20s or so, but its not the most featureful, and they only get bigger from there. So I always have an eye out for potential tools.

No, the inference accuracy of the image classifier is dependent on the model used and this is a demo of the code executing the model in a browser with GPU acceleration not the model itself. You can plug and play any model in the onnx format e.g. https://github.com/onnx/models. As a comparison, complaining about the "abysmal quality" of the dummy model on display here is like saying blender is bad 3d modeling software after opening it for the first time because all it models is a blank cube.

People complaining about quality here are missing the point, this is ONNX compatible inference engine written Rust, it just using 5MB SqueezeNet from 2016 for simplicity. Question is, is it worth to invest time and effort into ONNX ? https://en.wikipedia.org/wiki/SqueezeNet https://github.com/onnx/models?tab=readme-ov-file#image-clas... here is the same model using tensorflowjs https://hpssjellis.github.io/beginner-tensorflowjs-examples-... https://t-shaped.nl/posts/running-ai-models-in-the-browser-u...