Skip to content

Latest commit

 

History

History
30 lines (22 loc) · 857 Bytes

README.md

File metadata and controls

30 lines (22 loc) · 857 Bytes
Raw

Text-Based Sequential Recommendation

Reproduction of text-based sequential recommendation results. e.g., UniSRec.

Note

The original code has been refactored to be more concise and clean. As a result, statistics of the processed dataset, as well as the recommendation model results could be slightly different from the numbers in our paper.

Quick Start

Process the dataset

cd seq_rec_results/dataset/
python process_amazon_2023.py \
    --domain All_Beauty \
    --device cuda:0 \
    --plm hyp1231/blair-roberta-base

For all available domains/categories, please refer to Hugging Face Hub.

Train and evaluate the models

cd seq_rec_results/
python run.py \
    -m UniSRec \
    -d All_Beauty \
    --gpu_id=0