Skip to content
No description, website, or topics provided.
OpenEdge ABL Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.texpadtmp
data
loggan100_dim_cnn_rnn_pair_diff_gan_twoside_mi
pycocoevalcap
save
.DS_Store
README.md
converse_gan.py
data_utils.py
data_utils.pyc
demo.p
embedding_metrics.py
embedding_metrics.pyc
eval_metric.py
lstm_decoder.py
model.py
model.pyc
output.txt
rougescore.py
rougescore.pyc
run.sh
utils.py
utils.pyc

README.md

Converse GAN

Implementations of the models in the paper "Generating Informative and Diverse Conversational Responses via Adversarial Information Maximization" by Yizhe Zhang, Michel Galley, Jianfeng Gao, Zhe Gan, Xiujun Li, Chris Brockett, Bill Dolan, NeurIPS 2018

Prerequisite:

  • Tensorflow (version >1.9)
  • CUDA, cudnn
  • pip install nltk gensim

Run

  • See ./run.sh for training and testing commands

Data:

  • Not directly sharable due to copyright issue.
  • Data format and screening criteria will be released soon. Stay tuned.
  • See demo.p for data format.

For any question or suggestions, feel free to contact yizhe.zhang@microsoft.com

Citation

@inproceedings{zhang2018generating,
  title={Generating informative and diverse conversational responses via adversarial information maximization},
  author={Zhang, Yizhe and Galley, Michel and Gao, Jianfeng and Gan, Zhe and Li, Xiujun and Brockett, Chris and Dolan, Bill},
  booktitle={NeurIPS},
  year={2018}
}
You can’t perform that action at this time.