onnxruntime/docs/execution-providers/RKNPU-ExecutionProvider.md

title	description	parent	nav_order	redirect_from
RKNPU	Instructions to execute ONNX Runtime on Rockchip NPUs with the RKNPU execution provider	Execution Providers	11	/docs/reference/execution-providers/RKNPU-ExecutionProvider

RKNPU Execution Provider

PREVIEW

RKNPU DDK is an advanced interface to access Rockchip NPU. The RKNPU Execution Provider enables deep learning inference on Rockchip NPU via RKNPU DDK.

Contents

{: .no_toc }

• TOC placeholder {:toc}

Build

For build instructions, please see the BUILD page.

Usage

C/C++

To use RKNPU as an execution provider for inferencing, please register it as below.

Ort::Env env = Ort::Env{ORT_LOGGING_LEVEL_ERROR, "Default"};
Ort::SessionOptions sf;
Ort::ThrowOnError(OrtSessionOptionsAppendExecutionProvider_RKNPU(sf));
Ort::Session session(env, model_path, sf);

The C API details are here.

Support Coverage

Supported Platform

{: .no_toc }

• RK1808 Linux

Note: RK3399Pro platform is not supported.

Supported Operators

{: .no_toc }

The table below shows the ONNX Ops supported using the RKNPU Execution Provider and the mapping between ONNX Ops and RKNPU Ops.

ONNX Ops	RKNPU Ops
Add	ADD
Mul	MULTIPLY
Conv	CONV2D
QLinearConv	CONV2D
Gemm	FULLCONNECT
Softmax	SOFTMAX
AveragePool	POOL
GlobalAveragePool	POOL
MaxPool	POOL
GlobalMaxPool	POOL
LeakyRelu	LEAKY_RELU
Concat	CONCAT
BatchNormalization	BATCH_NORM
Reshape	RESHAPE
Flatten	RESHAPE
Squeeze	RESHAPE
Unsqueeze	RESHAPE
Transpose	PERMUTE
Relu	RELU
Sub	SUBTRACT
Clip(0~6)	RELU6
DequantizeLinear	DATACONVERT
Clip	CLIP

Supported Models

{: .no_toc }

The following models from the ONNX model zoo are supported using the RKNPU Execution Provider

Image Classification

• squeezenet
• mobilenetv2-1.0
• resnet50v1
• resnet50v2
• inception_v2

Object Detection

• ssd
• yolov3