Using local LLM's to run Multi-GPT

[redfort1987]

As the title suggests, wouldn't it be smart to use local LLMs to run the program? I spent about 80 iterations and it cost me about 8 euro's already in API cost.

I use models like ggml-gpt4all-j-v1.3-groovy.bin and ggml-model-q4_0.bin.