CMN metric

Code for the paper "Evaluating open-domain dialogues in latent space with next sentence prediction and mutual information" ACL2023

Model can be found at this google drive folder https://drive.google.com/drive/folders/1efxSMLXBWKnxu6jDHbmZDFaBfTV5XJjc

Download checkpoints and move to the corresponding datasets folders (dd, personachat)

Training commands

python Training.py --datasets [dd, personachat]

Evaluation commands

python Evaluation.py --datasets [dd, personachat]

Citation

If you found this repository or paper is helpful to you, please cite our paper.

@article{zhao2023evaluating,
  title={Evaluating Open-Domain Dialogues in Latent Space with Next Sentence Prediction and Mutual Information},
  author={Zhao, Kun and Yang, Bohao and Lin, Chenghua and Rong, Wenge and Villavicencio, Aline and Cui, Xiaohui},
  journal={arXiv preprint arXiv:2305.16967},
year={2023}
}