onnx

A browser-based computer vision project that detects cursor hotspots in screen recordings using temporal motion cues, ONNX Runtime Web, and WebGPU. The repo includes a live demo, training scripts, synthetic data generation, and baseline comparisons against YOLO models.

DocsGPT documents a small classifier used for bounded decisions like prompt-injection checks, relevance scoring, and conversation analytics. It explains where the model runs, how it is enabled, what endpoints it uses, and how it behaves on CPU, Mac GPU, or NVIDIA GPU.

A GitHub repository for an engine-agnostic KV cache fragment system for on-device LLM inference, with Android and Go components, adapters for llama.cpp/MLC-LLM/ONNX Runtime, and benchmark and monitoring code.

Browser-only local AI app for running compatible Hugging Face ONNX community models with WebGPU. The page shows a model library, local text generation, and an inference console in the browser.