Skip to content

Various implementations of the Neural Belief Tracker (Mrksic et al., 2017)

License

Notifications You must be signed in to change notification settings

ivulic/neural-belief-tracker

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Neural Belief Tracker

Contact: Nikola Mrkšić ([email protected])

An implementation of the Fully Data-Driven version of the Neural Belief Tracking (NBT) model.

This version of the model uses a learned belief state update in place of the rule-based mechanism used in the original paper.

Configuring the Tool

The config file in the config directory specifies the model hyperparameters, training details, dataset, ontologies, etc.

Running Experiments

train.sh and test.sh can be used to train and test the model (using the default config file). track.sh uses the trained models to 'simulate' a conversation where the developer can enter sequential user turns and observe the change in belief state.

About

Various implementations of the Neural Belief Tracker (Mrksic et al., 2017)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 99.8%
  • Shell 0.2%