Finetuned DialoGPT on Movie Dialogues

Basic Introduction

• A chatbot you can chat with by sending messages whenever you want.

We finetune the pre-trained microsoft dialoGPT dialogue systems on the movie dialogue corpus. And finally deploy the chatbots using the API provided by google voice.

Datasets

Movie Dialogue Corpus:

https://www.kaggle.com/datasets/Cornell-University/movie-dialog-corpus

The corpus contains a metadata-rich collection of fictional conversations extracted from raw movie scripts:

• 220,579 conversational exchanges between 10,292 pairs of movie characters
• involves 9,035 characters from 617 movies
• in total 304,713 utterances

Pre-trained Models

Seq2seq transformers from huggingface community:

• DialoGPT-small; DialoGPT-medium; DialoGPT-large.

Github Repo Link

https://github.com/kianakaslana648/nlp-final-project