original LoRa quantized weights in ggml format

repo: https://github.com/nomic-ai/gpt4all

modified for gpt4all alpaca.cpp repo: https://github.com/zanussbaum/gpt4all.cpp