A-LLMRec : Large Language Models meet Collaborative Filtering: An Efficient All-round LLM-based Recommender System
The source code for A-LLMRec : Large Language Models meet Collaborative Filtering: An Efficient All-round LLM-based Recommender System paper, accepted at KDD 2024.
In this paper, we propose an efficient all-round LLM-based recommender system, called A-LLMRec (All-round LLM-based Recommender system). The main idea is to enable an LLM to directly leverage the collaborative knowledge contained in a pre-trained collaborative filtering recommender system (CF-RecSys) so that the emergent ability of the LLM can be jointly exploited. By doing so, A-LLMRec can outperform under the various scenarios including warm/cold, few-shot, cold user, and cross-domain scenarios.
conda create -n [env name] pip
conda activate [env name]
pip install -r requirements.txt
Download dataset of 2018 Amazon Review dataset for the experiment. Should download metadata and reviews files and place them into data/amazon direcotory.
cd data/amazon
wget https://datarepo.eng.ucsd.edu/mcauley_group/data/amazon_v2/categoryFiles/Movies_and_TV.json.gz # download review dataset
wget https://datarepo.eng.ucsd.edu/mcauley_group/data/amazon_v2/metaFiles2/meta_Movies_and_TV.json.gz # download metadata
gzip -d meta_Movies_and_TV.json.gz
cd pre_train/sasrec
python main.py --device=cuda --dataset Movies_and_TV
- train stage1
cd ../../
python main.py --pretrain_stage1 --rec_pre_trained_data Movies_and_TV
- train stage2
python main.py --pretrain_stage2 --rec_pre_trained_data Movies_and_TV
To run with multi-gpu setting, assign devices with CUDA_VISIBLE_DEVICES command and add '--multi_gpu' argument.
- ex) CUDA_VISIBLE_DEVICES = 0,1 python main.py ... --multi_gpu
Inference stage generates "recommendation_output.txt" file and write the recommendation result generated from the LLMs into the file. To evaluate the result, run the eval.py file.
python main.py --inference --rec_pre_trained_data Movies_and_TV
python eval.py
