Torch Points 3D
stable

Developer guide

  • Getting Started
    • Installation
      • Install Python 3.6 or higher
      • Install dependencies using poetry
      • Minkowski engine support
      • Installation within a virtual environment
    • Train!
    • Visualise your results
    • Project structure
  • Tutorials
    • Create a new dataset
      • Create a dataset that the framework recognises
      • Create a new configuration file
    • Create a new model
      • Create the basic modules
      • Assemble all the basic blocks
      • Create a new configuration
      • Another example with RSConcv
    • Launch an experiment
    • Train and Test on tasks already implemented
      • Registration Task
  • Advanced
    • Configuration
      • Overview
      • Configuration architecture
      • Understanding config.yaml
      • Training arguments
      • Eval arguments
    • Data formats for point cloud
      • Dense
      • Sparse formats
    • Backbone Architectures
      • UnetBasedModel
      • UnwrappedUnetBasedModel
    • Datasets
      • Segmentation
      • Classification
      • Registration
    • Model checkpoint
      • Model Saving
      • Model Loading
      • Adding a new metric
    • Visualization
    • Custom logging

API

  • Models
  • Datasets
    • ShapeNet
      • Raw dataset
      • Wrapped dataset
    • S3DIS
      • Raw dataset
      • Wrapped dataset
    • Scannet
      • Raw dataset
      • Wrapped dataset
  • Transforms
  • Filters
Torch Points 3D
  • Docs »
  • Search
  • Edit on GitHub

© Copyright 2020, Thomas Chaton and Nicolas Chaulet Revision b2a69d01.

Built with Sphinx using a theme provided by Read the Docs.