TorchCTR

TorchCTR copied to clipboard

Install | Getting Started | Documentation |

TorchCTR

TorchCTR is a scalable and easy-to-use ML package for CTR (Click Through Rate) prediction and ranking in recommender system with Facebook PyTorch.

Classic Models

• [x] Logistic Regression
• [x] Factorization Machine by Osaka Univ.
• [x] Field-aware Factorization Machine by Criteo, CMU & NTU

Deep Learning Models

• [x] Wide and Deep by Google
• [x] DeepFM by Huawei & HIT
• [x] Neural Factorization Machine by NUS
• [x] Field-aware Neural Factorization Machine
• [ ] Factorization-supported Neural Network
• [ ] Product-based Neural Network
• [ ] Attentional Factorization Machine
• [ ] Deep and Cross Network
• [ ] Deep Interest Network
• [ ] eXtreme Deep Factorization Machine

Install

pip install torchctr --user

Getting Started

please check exmaples/

Documentation

Official Documentation