dx-all-suite
DEEPX NPU SDK (DXNN) for edge AI: compile, quantize and deploy ONNX models on DEEPX NPUs
- 80
- Python
- 3 d ago
DEEPX NPU SDK (DXNN) for edge AI: compile, quantize and deploy ONNX models on DEEPX NPUs
DEEPX NPU tutorials in Jupyter: install the SDK, compile ONNX to .dxnn, run inference and build Python/C++ edge AI apps. Demos: YOLO26, PP-OCR, CLIP, hand landmarks, 36-channel video.
DEEPX Model Zoo: 490+ pretrained computer vision models for DEEPX NPUs.
DEEPX NPU runtime and Linux driver for edge AI inference: DX-RT, firmware, DX-APP C++/Python examples and DX-Stream GStreamer plugins. Runs .dxnn models on DX-M1 M.2 accelerators.
DEEPX NPU compiler (DX-COM): convert ONNX models into optimized .dxnn binaries with INT8 quantization for DEEPX NPUs such as DX-M1. Part of the DXNN SDK