Existing AI models
LLaMa
Built by Meta (Facebook) - collection of language models ranging from 7B to 65B parameters
LLaMA-13B is 10x more powerful then ChatGPT-3 while being 10x smaller
GPT
ChatGPT-3.5, ChatGPT-4 are examples of LLMs
ChatGPT has 175 billion parameters
llama.cpc
C/C++ model to quiery LLaMA model
Completely open source
Allows to run on the Mac
Only runs on the CPU
Lower accuracy & degradation - leads to lower precision (loss of information)
alpaca.cpp
Same as llama model
It is an instruction-follow model
Fine-tuned version of llama 7B
Last updated