EasyRec is a deep learning framework for recommendation scenarios, offering a variety of models and features for personalized recommendations. It processes large-scale data, captures high-level features, and automates feature engineering and model selection to improve recommendation performance and user satisfaction.
PAI encapsulates EasyRec's core features as drag-and-drop algorithm components in Designer, with tools to build end-to-end recommendation pipelines for your scenarios.
The following examples show how to use EasyRec components in Designer for recommendation tasks such as feature engineering, sample generation, ranking, vector recall, and model evaluation:
The example code comes from the recommendation algorithm customization module of the End-to-end Deep Customization Platform for Recommendation.