Deploying RKNN Models: Evaluation and Inference Testing

Differences Between Loading Native and RKNN-Converted Models Models developed in frameworks like PyTorch, TensorFlow, or ONNX must be converted into the proprietary RKNN format to leverage Rockchip’s NPU aceleration. The RKNN format is optimized for Rockchip’s neural processing units, enabling efficient execution on embedded platforms such as t ...

Posted on Mon, 15 Jun 2026 18:27:05 +0000 by dwest

SSD Model Inference Performance Comparison on Rockchip NPU Platforms (RK3568, RK3588, RK1808)

This benchmark evaluates neural processing unit (NPU) performance across Rockchip chipsets using the SSD Inceptoin V2 object detection model with 300×300 RGB input tensors. Performance Results Platform Inference Time Throughput (FPS) RK1808 25 ms/frame 40 RK3588 35–50 ms/frame 20 RK3568 150 ms/frame 6 RK3588 Benchmark Output $ ./r ...

Posted on Sun, 17 May 2026 00:35:14 +0000 by stressedsue