Logo

Getting Started

  • Technical Overview
  • File Formats
  • Supported Platforms
  • Evaluate & Install

Programming Guide

  • CAD Fundamentals
    • Overview
    • What is CAD?
    • Types of CAD Representations
      • Boundary Representation (B-rep)
      • Mesh Representation
    • B-rep Topology Hierarchy
      • The B-rep Hierarchy
      • Vertices
      • Edges
      • Faces
      • Shells
      • Bodies (Solids)
      • Model (Assembly)
    • B-rep Topology vs. Geometry
      • Topology
      • Geometry
    • UV Grids: Sampling Face Geometry
      • What are UV Coordinates?
      • UV Grids in Machine Learning
    • Face Adjacency Graphs
      • What is a Face Adjacency Graph?
      • Building Face Adjacency Graphs
      • Graph Properties for ML
    • From CAD to Graph Neural Networks
      • Mapping B-rep to GNN
    • Advanced B-rep Concepts
      • Dihedral Angles
      • Surface Curvature
      • Machining Features
    • CAD File Formats
    • Coordinate Systems and Units
      • Right-Handed Coordinate System
      • Units
    • Best Practices for CAD ML
      • 1. Understand Your Data
      • 2. Normalize Features
      • 3. Handle Variable Graph Sizes
      • 4. Augmentation for CAD Data
      • 5. Feature Selection
    • Next Steps
  • Machine Learning Fundamentals
    • Overview
    • What is Machine Learning?
      • Types of Machine Learning Tasks
    • Neural Networks Basics
      • What is a Neural Network?
      • Key Components
      • Training Process
    • Graph Neural Networks (GNNs)
      • Why GNNs for CAD Data?
      • Graph Representation of CAD Models
      • How GNNs Work
      • Common GNN Architectures
      • Point Cloud Networks
      • Convolutional Neural Networks (CNNs)
      • Hybrid Architectures for CAD
    • PyTorch and PyTorch Lightning
      • PyTorch Basics
      • PyTorch Lightning
    • DGL: Deep Graph Library
      • DGL Graph Structure
    • PyTorch Geometric: Alternative Graph Library
      • When You Might See PyTorch Geometric
      • PyG Graph Structure
    • Loss Functions
      • Cross-Entropy Loss (Classification)
      • Binary Cross-Entropy (Node Classification)
      • Mean Squared Error (Regression)
    • Advanced Concepts
      • Transfer Learning and Domain Adaptation
      • Regularization Techniques
      • Attention Mechanisms
      • Self-Supervised Learning
    • Best Practices for CAD Machine Learning
      • Data Quality
      • Feature Engineering
      • Model Training
      • Debugging ML Models
    • Resources and Further Reading
    • Next Steps
  • Data Flow Management
    • CAD Data Access
    • CAD Data Encoding
    • Datasets - ML-Ready Inputs
    • Data Storage
      • Data Merging in HOOPS AI
    • Data Flow Customisation
      • Flow module - Quick Reference
  • Machine Learning Model
    • Dataset Exploration and Mining
    • Parts Classification Model
    • CAD Feature Recognition Model
    • Develop Your own ML Model
  • Data Visualization Experience

Python API Reference

  • hoops_ai module
    • hoops_ai.cadaccess
    • hoops_ai.cadencoder
    • hoops_ai.dataset
      • hoops_ai.dataset.dataset_explorer
      • hoops_ai.dataset.dataset_loader
      • hoops_ai.dataset.graph_dataset
      • hoops_ai.dataset.torch_adapter
    • hoops_ai.flowmanager
      • hoops_ai.flowmanager.base_task
      • hoops_ai.flowmanager.flow_builder
      • hoops_ai.flowmanager.parallel_task
      • hoops_ai.flowmanager.sequential_task
    • hoops_ai.insights
      • hoops_ai.insights.dataset_viewer
      • hoops_ai.insights.utils
      • hoops_ai.insights.viewer
    • hoops_ai.ml
      • hoops_ai.ml.EXPERIMENTAL
      • hoops_ai.ml.metric_explorer
    • hoops_ai.storage
      • hoops_ai.storage.cadfile_retriever
      • hoops_ai.storage.datastorage
      • hoops_ai.storage.datasetstorage
      • hoops_ai.storage.helpers
      • hoops_ai.storage.label_storage
      • hoops_ai.storage.loggers
      • hoops_ai.storage.metric_storage

Tutorials

  • 1. Accessing a CAD File
    • HOOPS AI: CAD Access Module
    • Holes extraction for certain Files.
    • Interactive 3D Visualization with HOOPS AI Insights
  • 2. Encoding a CAD File
    • View your cad model in the notebook
    • HOOPS AI: Encoder Module
    • BRep data as numerical features for ML
    • Encoding the Geometry
    • Encoding the attributes
  • 3. HOOPS AI - Minimal ETL Demo
    • HOOPS AI - Minimal ETL Demo
    • DATA SERVING : Use the DatasetExplorer to navigate your data
  • 4. Fabwave - Part Classification using HOOPS AI
    • Fabwave - Part Classification using HOOPS AI
    • Data Transformation : Encoded data to be used as ml input
    • Machine Learning Training
  • 5. Data Mining a 5K CAD Dataset
    • Use the Dataset Explorer to navigate the dataset
    • Gather files that fulfilled a given condition. Filter
    • Query data for single file
  • 6. Data Mining a 162K CAD Dataset
    • HOOPS AI: Use the Dataset Explorer to navigate the dataset
    • Gather files that fulfilled a given condition. Filter
    • Query data for single file
    • Create subsets (train, validation, test) based on the label distribution
  • 7. Train a Machine Learning Model For Parts Classification
    • HOOPS AI: EXPERIMENTAL - Flow Trainer
    • Make inference, test your current trained model
  • 8. Infer features using CAD as Input
    • HOOPS AI: Using a Pre-trained Model
    • ML prediction
    • Visualize Predictions on CAD Model
    • A second case

Additional Resources

  • Release Notes
    • HOOPS AI 1.0-preview
    • Fixed Bugs List
  • Public Roadmap
  • Acknowledgments
  • Distribute Your Application
  • Downloads
  • Glossary
  • Archives

Support

  • Forum
  • Knowledge Base
  • Support
  • Contact Us
HOOPS AI
  • Search


© Copyright 2025, Tech Soft 3D

Welcome Conversation saved
HOOPSY BETA

Hello! I'm HOOPSY

Your Docs generative AI assistant. Ask me anything about HOOPS AI or choose a sample question below to start a conversation:

What is HOOPS AI?
What are the latest release notes for HOOPS AI?
How do I start evaluating HOOPS AI?
How do I write my first sample application?
How can I get assistance?

HOOPSY may store conversations according to OpenAI's policy. Responses may not always be accurate.

HOOPSY is thinking...
Return to top