Welcome to MMDetection3D’s documentation!

Get Started

• Prerequisites
• Installation
  • Install MMDetection3D
  • Another option: Docker Image
  • A from-scratch setup script
  • Using multiple MMDetection3D versions
• Verification
  • Verify with point cloud demo
  • High-level APIs for testing point clouds
• Demo
  • Introduction
  • Testing
• Model Zoo
  • Common settings
  • Baselines
  • Model Zoo Statistics
• Dataset Preparation
  • Before Preparation
  • Download and Data Preparation

Quick Run

• 1: Inference and train with existing models and standard datasets
  • Inference with existing models
  • Train predefined models on standard datasets
• 2: Train with customized datasets
  • Prepare the customized dataset
  • Prepare a config
  • Train a new model
  • Test and inference

Supported Tasks

• LiDAR-Based 3D Detection
  • Data Preparation
  • Training
  • Quantitative Evaluation
  • Testing and Making a Submission
  • Qualitative Validation
• Vision-Based 3D Detection
  • Data Preparation
  • Training
  • Quantitative Evaluation
  • Testing and Making a Submission
  • Qualitative Validation
• LiDAR-Based 3D Semantic Segmentation
  • Data Preparation
  • Training
  • Quantitative Evaluation
  • Testing and Making a Submission
  • Qualitative Validation

Datasets

• NuScenes Dataset for 3D Object Detection
  • Before Preparation
  • Dataset Preparation
  • Training pipeline
  • Evaluation
  • Metrics
  • Testing and make a submission
  • Notes
• Lyft Dataset for 3D Object Detection
  • Before Preparation
  • Dataset Preparation
  • Training pipeline
  • Evaluation
  • Metrics
  • Testing and make a submission
• Waymo Dataset
  • Prepare dataset
  • Training
  • Evaluation
  • Testing and make a submission
• SUN RGB-D for 3D Object Detection
  • Dataset preparation
  • Train pipeline
  • Metrics
• ScanNet for 3D Object Detection
  • Dataset preparation
  • Training pipeline
  • Metrics
• ScanNet for 3D Semantic Segmentation
  • Dataset preparation
  • Training pipeline
  • Metrics
  • Testing and Making a Submission
• S3DIS for 3D Semantic Segmentation
  • Dataset preparation
  • Training pipeline
  • Metrics

Tutorials

• Tutorial 1: Learn about Configs
  • Config File Structure
  • Config Name Style
  • Deprecated train_cfg/test_cfg
  • An example of VoteNet
  • FAQ
• Tutorial 2: Customize Datasets
  • Support new data format
  • Customize datasets by dataset wrappers
  • Modify Dataset Classes
• Tutorial 3: Customize Data Pipelines
  • Design of Data pipelines
  • Extend and use custom pipelines
• Tutorial 4: Customize Models
  • Develop new components
• Tutorial 5: Customize Runtime Settings
  • Customize optimization settings
  • Customize training schedules
  • Customize workflow
  • Customize hooks

Useful Tools and Scripts

• Log Analysis
• Visualization
  • Results
  • Dataset
• Model Complexity
• Model Conversion
  • RegNet model to MMDetection
  • Detectron ResNet to Pytorch
  • Prepare a model for publishing
• Dataset Conversion
• Miscellaneous
  • Print the entire config

Notes

• Benchmarks
  • Settings
  • Main Results
  • Details of Comparison
• FAQ
  • MMCV/MMDet/MMDet3D Installation
• Compatibility with Previous Versions of MMDetection3D
  • MMDetection3D 0.16.0
  • MMDetection3D 0.15.0
  • MMDetection3D 0.14.0
  • MMDetection3D 0.12.0
  • MMDetection3D 0.6.0

API Reference

• API Documentation
  • mmdet3d.core
    • anchor
    • bbox
    • evaluation
    • visualizer
    • voxel
    • post_processing
  • mmdet3d.datasets
  • mmdet3d.models
    • detectors
    • backbones
    • necks
    • dense_heads
    • roi_heads
    • fusion_layers
    • losses
    • middle_encoders
    • model_utils

Switch Language

• English
• 简体中文

Indices and tables

• Index

• Search Page