TanGentleman / Augmenta

Automate RAG-powered workflows
MIT License
1 stars 0 forks source link

Choosing LLM is non-intuitive #31

Closed TanGentleman closed 6 months ago

TanGentleman commented 6 months ago

Currently, models are set via a field in settings.json. These have to be the name of the function that calls them, like "get_openai_gpt4". Like other projects in this space, having a dictionary for shortcuts proves useful.

I want to be able to type python chat.py -m gpt4 to have a chat with gpt4 as the primary model