Skip to content

Latest commit

 

History

History
31 lines (20 loc) · 1.17 KB

README.md

File metadata and controls

31 lines (20 loc) · 1.17 KB
Raw

RecoBERT

This repository contains implementation of the paper RecoBERT: A Catalog Language Model for Text-Based Recommendations [1]. This is NOT the original implenetation of the paper.

The paper can be found here: https://arxiv.org/pdf/2009.13292.pdf

Requirements

python==3.9.15
pandas==1.5.2
pytorch==1.13.0
transformers==4.25.1

Datasets

Preprocessed wines dataset used for RecoBERT training and evaluation along with annotated test subset can be found on this link.

Usage

To train and/or evaluate the RecoBERT implementation:

  1. Download preprocessed wines dataset from this link.
  2. Save the dataset to the ./data folder within this repository.
  3. Run training by using command: python ./scripts/training.py wines.
  4. After training, run inference by using command: python ./scripts/inference.py wines.

References

[1] I. Malkiel, O. Barkan, A. Caciularu, N. Razin, O. Katz and N. Koenigstein, "RecoBERT: A Catalog Language Model for Text-Based Recommendations," 2020.