Feat: GPT4All Model Integration

[SaturnCassini]

In case anyone prefers running it locally for free.

Beware it's quite slower than OpenAI!

[winglian]

This should really be refactored into a new module so that people could choose one or the other. Removing openai support completely locks some people out of using this project.

[SaturnCassini]

I agree, if someone feels like doing that fix please do so!

[MLTQ]

This shouldn't contain a hardcoded model, but should reference something in utils.py, and max_token counts should be similarly parameterizable.

[SaturnCassini]

Added the feature to select your own model, will look at max tokens and temperature next @MLTQ thanks for the feedback

[SaturnCassini]

Enabled issues if anyone finds any I'll try to keep it tidy where possible. Also looking forward to someone's refactor to use either gpt4all or openai

[muzian666]

Can add GPU support? Currently is too slow to run on the CPU.

[mikeyang01]

Is there any opensource LLM good for running this agent? Thanks a lot