Factorization Machine models in PyTorch

This package provides a PyTorch implementation of factorization machine models and common datasets in CTR prediction.

Available Datasets

• MovieLens Dataset
• Criteo Display Advertising Challenge
• Avazu Click-Through Rate Prediction

Models

Model	Reference
Logistic Regression
Factorization Machine	Factorization Machines
Field-aware Factorization Machine	Field-aware Factorization Machines for CTR Prediction
Factorization-Supported Neural Network	Deep Learning over Multi-field Categorical Data - A Case Study on User Response Prediction
Wide&Deep	Wide & Deep Learning for Recommender Systems
Attentional Factorization Machine	Attentional Factorization Machines: Learning the Weight of Feature Interactions via Attention Networks
Neural Factorization Machine	Neural Factorization Machines for Sparse Predictive Analytics
Field-aware Neural Factorization Machine	Field-aware Neural Factorization Machine for Click-Through Rate Prediction
Product Neural Network	Product-based Neural Networks for User Response Prediction
Deep Cross Network	Deep & Cross Network for Ad Click Predictions
DeepFM	DeepFM: A Factorization-Machine based Neural Network for CTR Prediction
xDeepFM	xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems
AutoInt (Automatic Feature Interaction Model)	AutoInt: Automatic Feature Interaction Learning via Self-Attentive Neural Networks

Each model's AUC values are about 0.80 for criteo dataset, and about 0.78 for avazu dataset. (please see example code)

Installation

pip install torchfm

API Documentation

https://rixwew.github.io/pytorch-fm

Licence

MIT