# For Flickr30K
cd datasets
python split_flickr_data.py

# For COCO
cd datasets
python process_coco.py

python macd_pretraining.py --cfg cfg/pretrain-flickr-resnet.yml

# For STS-B
python unsupervised_nli.py --cfg cfg/unsupervised/sts-b.yml
# For SNLI
python unsupervised_nli.py --cfg cfg/unsupervised/snli.yml
python snli_unsupervised.py --data_folder /home/a/MACD/unsupervised/flickr-resnet/snli

# For STS-B
python run_finetuning.py --cfg cfg/finetune/sts-b.yml
# For SNLI
python run_finetuning.py --cfg cfg/finetune/snli.yml

@inproceedings{cui2020unsupervised,
  title={Unsupervised Natural Language Inference via Decoupled Multimodal Contrastive Learning},
  author={Cui, Wanyun and Zheng, Guangyu and Wang, Wei},
  booktitle={Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)},
  pages={5511--5520},
  year={2020}
}