Welcome to MinkowskiEngine’s documentation!

The MinkowskiEngine is an auto-differentiation library for sparse tensors.

Table of Contents

Introduction

• Minkowski Engine
• Sparse Tensor Networks
• Quick Start
• Definitions and Terminology

Tutorials

• Sparse Tensor Basics
  • Data Generation
  • Sparse Tensor Initialization
  • Sparse Tensor for Continuous Coordinates
  • Sparse Tensor Arithmetics
  • Feature Concatenation
  • Batch-wise Decomposition

Demos

• Training Pipeline
  • Making a Dataset
  • Quantization
  • Making a DataLoader
  • Training
• ModelNet40 Classification
  • Making a ModelNet40 data loader
  • Data Augmentation
  • Training a ResNet for ModelNet40 Classification
  • Running the Example
• Semantic Segmentation
  • Segmentation of a hotel room
• 3D Sparsity Pattern Reconstruction
  • Making a Sparsity Pattern Reconstruction Network
  • Running the Example
• Working with Pytorch Layers
  • Example: Features for Classification
• Multi-GPU with Pytorch-Lightning
  • Dataset
  • Network
  • Lightning Module
  • Training
• PointNet

API

• SparseTensor and TensorField
  • SparseTensor
  • TensorField
  • SparseTensorOperationMode
  • SparseTensorQuantizationMode
  • set_sparse_tensor_operation_mode
  • sparse_tensor_operation_mode
  • global_coordinate_manager
  • set_global_coordinate_manager
  • clear_global_coordinate_manager
• MinkowskiConvolution
  • MinkowskiConvolution
  • MinkowskiChannelwiseConvolution
  • MinkowskiConvolutionTranspose
  • MinkowskiGenerativeConvolutionTranspose
• MinkowskiPooling
  • MinkowskiMaxPooling
  • MinkowskiAvgPooling
  • MinkowskiSumPooling
  • MinkowskiPoolingTranspose
  • MinkowskiGlobalMaxPooling
  • MinkowskiGlobalAvgPooling
  • MinkowskiGlobalSumPooling
• MinkowskiBroadcast
  • MinkowskiBroadcastAddition
  • MinkowskiBroadcastMultiplication
  • MinkowskiBroadcastConcatenation
  • MinkowskiBroadcast
• MinkowskiNormalization
  • MinkowskiBatchNorm
  • MinkowskiSyncBatchNorm
  • MinkowskiInstanceNorm
• MinkowskiNonlinearities
  • MinkowskiReLU
  • MinkowskiPReLU
  • MinkowskiSELU
  • MinkowskiSELU
  • MinkowskiCELU
  • MinkowskiDropout
  • MinkowskiThreshold
  • MinkowskiSigmoid
  • MinkowskiTanh
  • MinkowskiSoftmax
• MinkowskiPruning
  • MinkowskiPruning
• MinkowskiInterpolation
  • MinkowskiInterpolation
• MinkowskiUnion
  • MinkowskiUnion
• Coordinate Management
  • CoordinateMapKey
  • CoordinateManager
  • GPU Memory Allocator
• Utility Functions and Classes
  • sparse_quantize
  • batched_coordinates
  • sparse_collate
  • batch_sparse_collate
  • cat
  • to_sparse
  • to_sparse_all
  • SparseCollation
  • MinkowskiToSparseTensor
  • MinkowskiToDenseTensor
  • MinkowskiToFeature
  • MinkowskiStackCat
  • MinkowskiStackSum
  • MinkowskiStackMean
  • MinkowskiStackVar
• Miscellaneous Classes
  • Kernel Generator
  • RegionType
• Miscellanea
  • Controlling the number of threads
  • is_cuda_available
  • cuda_version
  • get_gpu_memory_info
  • set_memory_manager_backend
  • References

Miscellanea

• Common Issues
• Guidelines for Faster Networks
• Migration Guide from v0.4.x to v0.5.0
• Benchmark

Indices and tables

• Index

• Search Page