Welcome to NVTabular’s documentation!

• Introduction
  • Benefits
  • Installation
    • Installing NVTabular Using Conda
    • Installing NVTabular with Docker
  • Examples and Tutorials
  • Feedback and Support
• Core Features
  • TensorFlow and PyTorch Interoperability
  • HugeCTR Interoperability
  • Multi-GPU Support
  • Multi-Node Support
  • Multi-Hot Encoding and Pre-Existing Embeddings
  • Shuffling Datasets
  • Cloud Integration
• Accelerated Training
  • Tensorflow
  • PyTorch
  • HugeCTR
• Examples
  • Getting Started with Movielens
    • Download and Convert
    • ETL with NVTabular
    • Training with TensorFlow
    • Training with PyTorch
  • Advanced Ops with Outbrain
    • Download and Convert
    • ETL with NVTabular
    • Training with TensorFlow
  • Scaling to Large Datasets with Criteo
    • Download and Convert
    • ETL with NVTabular, Training with PyTorch
    • ETL with NVTabular, Training with HugeCTR
  • Winning Solution of the RecSys2020 Competition
    • Overview
    • Getting Started
    • Preparing our dataset
    • Splitting dataset into training and test
    • Feature Engineering
    • Training our model
  • Tabular Data with Rossmann
    • Download and Convert
    • ETL with NVTabular
    • Training with TensorFlow
    • Training with PyTorch
    • Training with FastAI
  • Multi-GPU Example
    • NVTabular + Dask Integration
    • Simple Multi-GPU Toy Example
    • Notes on Shuffling
• Additional Resources
  • API Documentation
    • Workflow
    • Dataset
    • Operators
    • ColumnGroup
    • Torch Dataloader
    • Torch Layers
    • TensorFlow Dataloader
    • TensorFlow Layers
  • Architecture
    • Operations
    • A Higher Level of Abstraction
  • Cloud Integration
    • Amazon Web Services
    • Google Cloud Platform
  • Troubleshooting
    • Checking the Schema of the Parquet File
    • Setting the Row Group Size for the Parquet Files
  • Github Repo

Indices and tables

• Index

• Module Index

• Search Page