MCR: Multi-perspective Coherent Reasoning for Helpfulness Prediction of Multimodal Reviews

Code for the paper "Multi-perspective Coherent Reasoning for Helpfulness Prediction of Multimodal Reviews" (ACL 2021).

If you use this code, please cite the paper using the BibTeX reference below.

@inproceedings{mcr,
    title={Multi-perspective Coherent Reasoning for Helpfulness Prediction of Multimodal Reviews},
    author={Junhao Liu, Zhen Hai, Min Yang, and Lidong Bing},
    booktitle={Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics, {ACL} 2021},
    year={2021},
}

Requirements

• Python >= 3.7
• PyTorch >= 1.6

You can install all required Python packages with

pip install -r requirements.txt

Datasets

To obtain the multimodal datasets of Lazada-MRHP and Amazon-MRHP, please read the details provided here.

Running the code

Train

Use the following commands to train the model based on the specific configuration file.

# single gpu or data parallel, [ckpt] is optional for continual training
sh scripts/train.sh device_ids config_file [ckpt]

# or distributed training
sh scripts/train_dist.sh device_ids n_procs config_file [ckpt]

Sample configuration files are provided in the config folder.

Evaluate

Do evaluation on a specific dataset based on the saved model checkpoint and corresponding configuration file.

sh scripts/eval.sh device_ids config_file ckpt

Contact

If you have any questions, please feel free to email me at [email protected].

Acknowledgments

Our code is built based on the text matching library MatchZoo and the PyTorch version bottom-up-attention.