TheBloke

/

galpaca-30B-GPTQ

like 48

First off, thank you so much for doing so much work to convert these llama models for the community! It's been such a bit help in decreasing cost and increasing speed by making the LLMs fit on my GPU in my projects.

Would you consider also generating a GPTQ 4bit version of the medalpaca model (llama finetuned on a bunch of medical datasets + alpaca) - https://huggingface.co/medalpaca/medalpaca-13b - as well? Thank you so much in advance!