Empathetic-ChatBot

Fine tuning GPT2 on the empathetic dataset to create an open-domain conversation model. REST API and Telegram bot are deployed separately from finely tuned model.

• The transformers library, which contains the latest NLP models (such as BERT, XLNet, GPT-2) will help us in our task. Microsoft’s DialoGPT was added to the Transformers model collection. This model is used.
• Hugginface configuration was used to train the model.
• REST API prototype was built using Python and the Flask web framework.
• python-telegram-bot package helped to create the telegram bot.

USAGE:

1. Clone the Repository
2. pip install - r requirements.txt to install package dependency
3. train.py to train and save the empathy model.
4. interface_empathy.py to test the model in console.
5. app_empathy.py to deploy the model in REST API.
6. tele_empathy.py to crate the telegram bot.

DATA:

Model training on publicly-available empathetic dialogue generation and EMPATHETICDIALOGUES from Allen School of Computer Science & Engineering, University of Washington and Facebook AI Research. Dataset is cleaned and convert into conversation data.

OUTPUT:

REFERENCE:

[1]. https://research.fb.com/publications/towards-empathetic-open-domain-conversation-models-a-new-benchmark-and-dataset/

[2]. https://huggingface.co/