Neighboring Distribution Divergence

Official Implementation of Paper "A Novel Metric for Evaluating Semantics Preservation" on ArXiv:
https://arxiv.org/abs/2110.01176 Accepted by ACL Findings 2023

Requirements

numpy==1.20.1
torch==1.9.1
spacy==2.2.0
nltk==3.6.4
transformers==4.12.3

Compress!

You can use the code python main.py to run our demo compressor.

Citation

Cite our code if you find it to be helpful for your projects

@article{DBLP:journals/corr/abs-2110-01176,
  author    = {Letian Peng and
               Zuchao Li and
               Hai Zhao}, 
  title     = {A Novel Metric for Evaluating Semantics Preservation}, 
  journal   = {CoRR}, 
  volume    = {abs/2110.01176}, 
  year      = {2021}, 
  url       = {https://arxiv.org/abs/2110.01176}, 
  eprinttype = {arXiv}, 
  eprint    = {2110.01176}, 
  timestamp = {Fri, 08 Oct 2021 15:47:55 +0200}, 
  biburl    = {https://dblp.org/rec/journals/corr/abs-2110-01176.bib}, 
  bibsource = {dblp computer science bibliography, https://dblp.org}
}