A simple chatbot implemented using pytorch library.
It uses Seq2Seq model built with multi-layered Gated Recurrent Unit
encoder and decoder.
Model is configured for
Cornell Movie-Dialogs Corpus.
By modifying Loader class and config.yaml it could be used on other datasets as well.
Repository is inspired by a Pytorch Chatbot Tutorial.
Project originally used Python 3.8.
To install required libraries execute:
pip install -r requirments.txt
Training script will download and preprocess data needed for training the model, results will be saved
to a directory specified in a config.yaml file.
To train model simply run:
python train.py
After model is trained chat with a model can be performed by running.
python chat.py