recbole

This is a repository of public data sources for Recommender Systems (RS).

This is a library built upon RecBole for cross-domain recommendation algorithms

Jupyter тетрадка с решением Kaggle соревнования Alfa Campus

PyTorch implementation of our paper "Review-aware Graph Convolution Network for explainable recommendation"

A collection of social datasets for RecBole-GNN.

[WWW '23] The official implmentation of our paper "Fine-tuning Partition-aware Item Similarities for Efficient and Scalable Recommendation"

Two-stage e-commerce recommender system (ALS retrieval + SASRec re-ranking) with Optuna tuning, evaluated on RetailRocket and Amazon Video Games datasets

A simplified and improved version based on the original RecSysDatasets/conversion_tools

recsys course notebooks

The system tracks the emissions of a given recommendation algorithm on a given dataset.

Tecniche di early stopping sostenibili per modelli di raccomandazione, mirate a ridurre le emissioni di CO2 durante l’addestramento senza compromettere in maniera significativa le performance. Basato su RecBole e CodeCarbon.

Multi-objective recommendation pipeline balancing item relevance with environmental impact for sustainable recommendations.

A verified SASRec reproduction and a semantic-ID generative recommender on one backbone, one eval harness, one set of frozen negatives. Full-catalog vs. sampled ranking, a measured seed-noise floor, and every README table CI-asserted against the run that produced it.