zhouyonglong / DIN_DIEN_tf2.0

Deep Interest Network for Click-Through Rate Prediction Deep Interest Evolution Network for Click-Through Rate Prediction

DIN_DIEN_tf2

Deep Interest Network for Click-Through Rate Prediction / Deep Interest Evolution Network for Click-Through Rate Prediction。

代码来源：zxxwin/DIEN-tf2

包含了三个模型的底层复现。具体细节可以参考以下两篇博客：

[CTR模型] DIN（Deep Interest Network）模型解读与Deepctr实现

[CTR模型] DIEN（Deep Interest Evolution Network）模型解读与Deepctr实现

Deep Interest Netw

ork for Click-Through Rate Prediction Deep Interest Evolution Network for Click-Through Rate Prediction

I reference zhougr1993 and mouna99 code and converte it to TensorFlow 2.0. This code performs similarly to the paper on ml-20 and amazon datasets. You can modify the model called in main.py and then utilize a model such as Base, DIN, DIEN.

Requirements

• python 3.6
• tensorflow 2.0

Run python main.py