Technical requirements
The code in this chapter requires the use of a GPU. For the section on using HuggingGPT in particular, both a GPU and plenty of space on the hard disk drive are required (several models will be downloaded, including diffusion models. For this, it will be necessary to use Git Large File Storage (LFS), which allows downloading wide files via Git). Anaconda should be installed to obtain the various libraries (the necessary libraries will be set up directly during installation). For readers who do not have these resources, the Using HuggingGPT on the web section shows how you can use HuggingGPT on the web. For local use of HuggingGPT, it is necessary to have an OpenAI token, while for web use, it is also necessary to have a Hugging Face token. The multi-agent system is based on Python libraries (NumPy, scikit-learn, SentenceTransformers, and Transformers).
HuggingGPT should be run on a GPU. The multi-agent system should be run on a GPU, but it could also be run...
