Welcome to MMDeploy’s documentation!

You can switch between Chinese and English documents in the lower-left corner of the layout.

Get Started

• Get Started
  • Introduction
  • Prerequisites
  • Installation
  • Convert Model
  • Inference Model
  • Evaluate Model

Build

• Build from Source
• Use Docker Image
• Build from Script
• CMake Build Option Spec

Run & Test

• How to convert model
• How to write config
• How to evaluate model
• Quantize model
• Useful Tools

SDK Usage

• SDK Documentation

Benchmark

• Supported models
• Benchmark
• Test on embedded device
• Test on TVM
• Quantization test result

OpenMMLab Codebase Support

• MMPretrain Deployment
• MMDetection Deployment
• MMSegmentation Deployment
• MMagic Deployment
• MMOCR Deployment
• MMPose Deployment
• MMDetection3d Deployment
• MMRotate Deployment
• MMAction2 Deployment

Backend Support

• Supported ncnn feature
• onnxruntime Support
• OpenVINO Support
• PPLNN Support
• SNPE feature support
• TensorRT Support
• TorchScript support
• Supported RKNN feature
• TVM feature support
• Core ML feature support

Custom Ops

• ONNX Runtime Ops
• TensorRT Ops
• ncnn Ops

Developer Guide

• mmdeploy Architecture
• How to support new models
• How to support new backends
• How to add test units for backend ops
• How to test rewritten models
• How to get partitioned ONNX models
• How to do regression test

Experimental feature

• ONNX export Optimizer

Appendix

• Cross compile snpe inference server on Ubuntu 18

FAQ

• Frequently Asked Questions

API Reference

• apis
• apis/tensorrt
• apis/onnxruntime
• apis/ncnn
• apis/pplnn

Indices and tables

• Index

• Search Page