Version 0.2.0

This is another complete upgrade of the project, using as a basis version 0.4.0 of AutoGPT. Also, the default model now is the better performing llama-3.1-70b-versatile from the Groq API. This version works better than the previous, but there are still some issues.

Version 0.1.0

This is a complete overhaul of the project. I couldn't get it to act right, so I rebuilt it starting from version 0.3.0 of the AutoGPT project. I have returned all the memory options and the speak options. I am using the simplest of token counting methods for now. I am using the spacy en_core_web_md model for embedding. I have removed any use of any OpenAI product, so as to make it completely free and open source.

Version 0.0.2

To avoid using embeddings, at least until I can find an open source embeddings API, I switched the local memory system to sqlite. Not very efficient, but this thing is still so buggy it'll have to work for now.

Version 0.0.1 - Initial Release

Initial release of the AutoLlama program. This is a bare bones version of an old AutoGPT program, but it uses the free Llama3 model from Groq API.