Working with HuggingGPT
There are two ways you can use HuggingGPT:
- Clone the repository locally
- Use the web service
Here, we will look at the two methods. The main difference is that when we clone the repository locally, we download all the models, and the system execution will be conducted locally. In contrast, the web service method requires that the execution is conducted in a service. In both cases, all models are used in inference; the difference lies in where the models are executed and the resources employed. Additionally, both approaches support the use of a web-based GUI.
Using HuggingGPT locally
To clone HuggingGPT (the corresponding repository is called Jarvis), it is useful to use Git LFS. Git LFS is an open source extension of Git. Git is designed to manage code repositories but not large binary files (such as videos, datasets, or high-resolution images). Git LFS is crucial for repositories that include large assets (e.g., datasets, videos, or...
