Chatbot implementation using a trained in seq2seq model developed in Tensorflow and serving in a Flask server.
Picture from Deep Learning for Chatbots
This project is divided in two parts: - The tensorflow implementation of the seq2seq model: chat.py - Webchat to get input from the user, "respond" them by running the model inference and display the output: main.py
The model construction is based on the following implementation Advanced dynamic seq2seq, with modifications involving new datasets and ways of training.
Front-end is using this simple chat implementation on https://bootsnipp.com/snippets/featured/simple-chat, modified to make requests to the Flask server.