NeurIPS 2018 presentation.

BPE vocabulary: https://www.dropbox.com/s/n2jbjyq32x6jgr6/parameters.zip?dl=1

Model checkpoint file: https://www.dropbox.com/s/cs6zd9yntn6ixea/last_checkpoint?dl=1

• Alexander Tselousov [email protected]
• Sergey Golovanov [email protected]

Unzip BPE vocabulary files into ./parametes folder and save checkpoint into ./checkpoints folder or use scripts (see below).

The easiest way to prepare environment is to run script prepare_environment.sh. After that docker container with retrieval server must be run in demon mode and image with transformer_chatbot must be built. Run scripts from the root folder of this repository.

Retrieval server can be run with script run_retrieval_servet.sh. Server do not need the internet connection for its work, for connection with transformer_chatbot port 9200 is used (containers must be in the same docker network).

After preparations metrics can be evaluated with corresponding docker_*.sh scripts or *.py scripts can be used during interactive container run.

Usage of docker container for transformer_chatbot is not necessary, but retrieval server must always be run for correct work of transformer_chatbot.

List of used python modules is in requirements.txt. Also pytorch=0.4.1 is used.

f1: 0.178

hits@1: 0.173