This repo provides the source code & data of our paper "GreaseLM: Graph REASoning Enhanced Language Models".

• Python == 3.8
• PyTorch == 1.8.0
• transformers == 3.4.0
• torch-geometric == 1.7.0

Run the following commands to create a conda environment (assuming CUDA 10.1):

conda create -y -n greaselm python=3.8
conda activate greaselm
pip install numpy==1.18.3 tqdm
pip install torch==1.8.0+cu101 torchvision -f https://download.pytorch.org/whl/torch_stable.html
pip install transformers==3.4.0 nltk spacy
pip install wandb
conda install -y -c conda-forge tensorboardx
conda install -y -c conda-forge tensorboard

# for torch-geometric
pip install torch-scatter==2.0.7 -f https://pytorch-geometric.com/whl/torch-1.8.0+cu101.html
pip install torch-cluster==1.5.9 -f https://pytorch-geometric.com/whl/torch-1.8.0+cu101.html
pip install torch-sparse==0.6.9 -f https://pytorch-geometric.com/whl/torch-1.8.0+cu101.html
pip install torch-spline-conv==1.2.1 -f https://pytorch-geometric.com/whl/torch-1.8.0+cu101.html
pip install torch-geometric==1.7.0 -f https://pytorch-geometric.com/whl/torch-1.8.0+cu101.html

Download all the raw data -- ConceptNet, CommonsenseQA, OpenBookQA -- by

./download_raw_data.sh

You can preprocess the raw data by running

CUDA_VISIBLE_DEVICES=0 python preprocess.py -p <num_processes>

You can specify the GPU you want to use in the beginning of the command CUDA_VISIBLE_DEVICES=.... The script will:

• Setup ConceptNet (e.g., extract English relations from ConceptNet, merge the original 42 relation types into 17 types)
• Convert the QA datasets into .jsonl files (e.g., stored in data/csqa/statement/)
• Identify all mentioned concepts in the questions and answers
• Extract subgraphs for each q-a pair

TL;DR. The preprocessing may take long; for your convenience, you can download all the processed data here into the top-level directory of this repo and run

unzip data_preprocessed.zip

Add MedQA-USMLE. Besides the commonsense QA datasets (CommonsenseQA, OpenBookQA) with the ConceptNet knowledge graph, we added a biomedical QA dataset (MedQA-USMLE) with a biomedical knowledge graph based on Disease Database and DrugBank. You can download all the data for this from [here]. Unzip it and put the medqa_usmle and ddb folders inside the data/ directory.

The resulting file structure should look like this:

.
├── README.md
└── data/
    ├── cpnet/                 (preprocessed ConceptNet)
    └── csqa/
        ├── train_rand_split.jsonl
        ├── dev_rand_split.jsonl
        ├── test_rand_split_no_answers.jsonl
        ├── statement/             (converted statements)
        ├── grounded/              (grounded entities)
        ├── graphs/                (extracted subgraphs)
        ├── ...

To train GreaseLM on CommonsenseQA, run

CUDA_VISIBLE_DEVICES=0 ./run_greaselm.sh csqa --data_dir data/

You can specify up to 2 GPUs you want to use in the beginning of the command CUDA_VISIBLE_DEVICES=....

Similarly, to train GreaseLM on OpenbookQA, run

CUDA_VISIBLE_DEVICES=0 ./run_greaselm.sh obqa --data_dir data/

To train GreaseLM on MedQA-USMLE, run

CUDA_VISIBLE_DEVICES=0 ./run_greaselm__medqa_usmle.sh

You can download a pretrained GreaseLM model on CommonsenseQA here, which achieves an IH-dev acc. of 79.0 and an IH-test acc. of 74.0.

You can also download a pretrained GreaseLM model on OpenbookQA here, which achieves an test acc. of 84.8.

You can also download a pretrained GreaseLM model on MedQA-USMLE here, which achieves an test acc. of 38.5.

To evaluate a pretrained GreaseLM model checkpoint on CommonsenseQA, run

CUDA_VISIBLE_DEVICES=0 ./eval_greaselm.sh csqa --data_dir data/ --load_model_path /path/to/checkpoint

Again you can specify up to 2 GPUs you want to use in the beginning of the command CUDA_VISIBLE_DEVICES=....

SimilarlyTo evaluate a pretrained GreaseLM model checkpoint on OpenbookQA, run

CUDA_VISIBLE_DEVICES=0 ./eval_greaselm.sh obqa --data_dir data/ --load_model_path /path/to/checkpoint

• Convert your dataset to {train,dev,test}.statement.jsonl in .jsonl format (see data/csqa/statement/train.statement.jsonl)
• Create a directory in data/{yourdataset}/ to store the .jsonl files
• Modify preprocess.py and perform subgraph extraction for your data
• Modify utils/parser_utils.py to support your own dataset

This repo is built upon the following work:

QA-GNN: Question Answering using Language Models and Knowledge Graphs
https://github.com/michiyasunaga/qagnn

Many thanks to the authors and developers!