- You need an OpenAI API key.
- If you want to run the open-source models, you will need to install them locally or run them on RunPod.
- Here's a YouTube video showing you how: https://youtu.be/Y5OV58lGTh8
configurations.json
: Add your API key to this file.locations.json
: Update this file with the location of theconfigurations.json
file.
Some important notes about the environment.txt files. Due to the update in the OpenAI API, the Proprietary_vs_OpenSource Rap Battle.ipynb
and speak.py
are running on different requirements.txt
files. I couldn't get the open-source models to work with the latest OpenAI and AutoGen Libs.
So, for the Proprietary_vs_OpenSource Rap Battle.ipynb
, you will need to run this with the requirements.txt
file in the folder Requirements for Notebook
. For speak.py
, you need to run this with the requirements.txt
file in the folder Requirements for speak script
.
The best way to do this is to create two virtual environments, which you can do with Anaconda.
- The main script is the
Proprietary_vs_OpenSource Rap Battle.ipynb
notebook. This generates the battles. You should run this end-to-end first. - The
speak.py
script generates the voices from the lyrics by calling the OpenAI API. Run this to render the voices.