Model-to-NPU pipelines for Qualcomm Snapdragon: QNN/ONNX/Android runtimes for on-device image and video generation.
Run LLMs on Snapdragon NPU — including the 'unsupported' 8 Gen 1 (Hexagon v69). Verified at 31 tok/s on OnePlus 10 Pro.
Native push-to-talk voice-to-text for Windows on ARM. Powered by NVIDIA Parakeet TDT v3 on the Snapdragon X Elite Hexagon NPU. Completely local.
Why the AYN Odin 3's (Qualcomm CQ8725S / SM8750) Hexagon NPU can't be enabled — a 47-phase reverse-engineering teardown proving the NSP subsystem is fused off at the factory (~94% confidence). On-device fuse reads, FeatureEnabler trustlet RE, kernel-source + patent attribution, AVB/dm-verity brick postmortem.
On-device AI exercise coach for Snapdragon — recognises 7 exercises, counts reps, scores form, and coaches you by voice. 100% offline on the Hexagon NPU; no INTERNET permission.
Keyboard-driven .NET terminal UI to find, run, chat with, and serve GGUF/LLM models on the Qualcomm Snapdragon NPU via GenieX — with auto CPU/NPU/GPU compute selection and one-key offload to external/NAS drives.
Add a description, image, and links to the hexagon-npu topic page so that developers can more easily learn about it.
To associate your repository with the hexagon-npu topic, visit your repo's landing page and select "manage topics."