LLM Debates
LLM Debates is a sophisticated project that leverages Large Language Models to generate and analyze debates on various topics. This tool is designed to create structured, multi-round debates with arguments for and against a given topic, complete with audio generation capabilities.
Example debate (topic: Veganism)
Table of Contents
Installation
To set up the project, follow these steps:
-
Clone the repository:
git clone https://github.com/yourusername/llm_debates.git cd llm_debates -
Install the required dependencies:
pip install -r requirements.txt -
Install Playwright:
Usage
The project includes several scripts for generating debates and processing audio. Here are the main components:
- Debate Generation (not shown in the provided code snippets)
- Audio Processing
To run the audio processing script:
./make_full_speech.sh <number_of_rounds> [output_filename]
This script will concatenate the debate audio files with added silence between segments.
Features
- Multi-round debate generation
- Argument creation for and against a given topic
- Web search integration for sourcing relevant information
- Audio generation for debate speeches
- Structured data management for debate rounds and arguments
Audio Processing
The make_full_speech.sh script (lines 1-53) provides a comprehensive solution for concatenating debate audio files with added silence between segments.
You can convert the resulting mp3 file into a video for social media by using an online tool or simply something like this:
ffmpeg -i veganism.mp3 -i image.png -filter_complex "[0:a]aformat=channel_layouts=mono,showwaves=s=1280x720:mode=line:rate=25:colors=0xFFFFFFFF|0x00000000,format=yuva420p[wave];[1][wave]overlay=W/2-w/2:H-h-50:format=auto" -c:v libx264 -c:a aac -b:a 192k -shortest output_video_veganism.mp4License
This project is licensed under the Apache License 2.0. For more details, see the LICENSE file in the repository.
