🏗️
Modular Design
Easy to add new models, datasets, and evaluation metrics with a clean architecture
Torch-RecHubA Lightweight PyTorch Framework for Recommendation Systems
Easy-to-use, extensible, and high-performance recommendation system framework
Skip to content
⚡
PyTorch-Based
Leverages PyTorch's dynamic graph and GPU acceleration capabilities for high performance
📚
Rich Model Library
Contains various classic and cutting-edge recommendation algorithms including ranking, matching, and multi-task models
🔄
Standardized Pipeline
Provides unified data loading, training, and evaluation workflows for consistent experiments
⚙️
Easy Configuration
Adjust experiment settings via config files or command-line arguments without code changes
🎯
Reproducibility
Designed to ensure reproducible experimental results with fixed random seeds and deterministic operations
🔧
Advanced Features
Supports negative sampling, multi-task learning, and other advanced recommendation techniques
📊
Multiple Datasets
Built-in support for MovieLens, Amazon, Criteo, Avazu, and many other popular datasets