Git Product home page Git Product logo

debatebot's Introduction

Debate bot - AIstiny

This project is an attempt at creating a "debate bot" based on the content generated in the twitch "debatesphere," mostly concentrating on the style of the streamer Destiny.

The architecture is based on GPT2 leveraging the DialoGPT codebase, tutorials and models. The repository is structured into 3 parts: tutorial, data and training.

  • tutorial: the ipython notebooks and other relevant information
  • training: the streamlined training scripts
  • data: all the relevant scripts and raw data related to generating the data

Currently, there is only the tutorial notebook based on the Make your own Rick Sanchez (bot) with Transformers and DialoGPT fine-tuning, it has been revised and updated for the changes in the huggingface. You can run this notebook directly in Google Colaboratory by using this link.

Phase 0, current to do list:

  • Check and maintain the Rick Sanchez tutorial notebook.
  • Tutorial notebook should train only for Rick’s responses
  • Create a streamlined training script based on huggingface.
  • Find a new dataset to train that is similar to streamer debates, i.e. lengthy responses allowed.
  • Develop data preparation scripts

Other information:

The initial Phase 0 reddit post.

Current state: Phase 0

This is a project to train an open AI model for the ultimate debate bot/debate bro AIstiny (the name is cringe and preliminary, I am open to suggestions), based on the past debates of Destiny. The architecture uses open source DialoGPT architecture, pretrained models, and the structured subtitles of the youtube videos of Destiny.

This documents is the initial resource document, which has the technical information and most importantly the data gathering efforts, i.e. the video links, and relevant information.

Tech infrastructure:

References DialoGPT:

Make your own Rick Sanchez (bot) with Transformers and DialoGPT fine-tuning

Easy Chatbot with DialoGPT, Machine Learning and HuggingFace Transformers – MachineCurve

Github: microsoft/DialoGPT: Large-scale pretraining for dialogue

DialoGPT — transformers 4.10.1 documentation

Further reading:

🦄 How to build a State-of-the-Art Conversational AI with Transfer-Learning

Practical Applications of Open AI’s GPT-2 Deep Learning Model

Latitude-Archives/AIDungeon: Infinite adventures await!

philip-bl/gpt2_chatbot: Chatbot using GPT-2

Data references:

The most intensive part of the process is to create the clean and structured data for Destiny's debates in dialogue format. As a first step we are creating a list of interesting debates. These debates must have good recording quality, preferably 1v1 and relatively less talking over each other (all for better transcription), and be on interesting topics

List of target debates:

The community contributed list can be found here. If anyone wants to suggest more videos please fill up the form.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.