an edge-real-time anchor-free object detector with decent performance

YOLOv5 in PyTorch > ONNX > RKNN

Run your yolov7 object detection with Rockchip NPU platforms (RK3566, RK3568, RK3588, RK3588S, RV1103, RV1106, RK3562).

Inference YOLOv8 segmentation on ONNX, RKNN, Horizon and TensorRT

Automated script to convert Huggingface and GGUF models to rkllm format for running on Rockchip NPU

export any your YOLOv7 model to TensorFlow, TensorFlowJs, ONNX, OpenVINO, RKNN,...

Inference YOLOv8 detection on ONNX, RKNN, Horizon and TensorRT

Superresolution running on Rockchip NPU (RK3588, etc..)

fatigue detect rknn/onnx model deploy in rk3568 npu（ROCK 3A）

Convert and run scikit-learn MLPs on Rockchip NPU.

Inference YOLOv7 segmentation on ONNX, RKNN, Horizon and TensorRT

Run Yolov8 on Rockchip NPU (from Orange PI 5)

CSD-RkNN: Conic Section Discriminances for Large Scale Reverse k Nearest Neighbors Queries

Inference YOLOv7 tiny detection on ONNX, RKNN, Horizon and TensorRT

Try to provide inference(face, pose, ocr, tts, etc.) for onnx and tensorRT and rknn.

This project focuses on deploying YOLO model and SGBM (Distance Measurement Algorithm) on RKNN platform

ImageAnalyst RKNN is an extension to the ImageAnalyst library, providing additional models and functions using rknn-toolkit-lite2.

kan的rknn2部署