Had a request
Searched around and couldn't find anyway to contact you directly, but I had a request. I'm still learning about LLMs, but MosaicML just released https://www.mosaicml.com/blog/mpt-7b and I was trying to use Llama.cpp to convert to ggml for use in LangChain. I seem to be missing something though as I can't get it to convert properly. You seem really good at doing these conversions, so I was wondering if I could bug you about this. My Discord username is Oh-now-uh#6551 if you wanted to reach out there. Thank you for all of the work that you do!
Best way to reach me is on Discord, I'm TheBloke#9741 and you can find me on these Discords:
Alpaca Lora: https://discord.gg/eSdptpkm
GPT4All: https://discord.gg/sfWUbDKH
LmSys (Vicuna): https://discord.gg/CPz84krv
MPT 7B looks amazing but I'm afraid right now you can't use it in llama.cpp or in ooba text-gen-webui or anywhere else like that, as it's a brand new model format. It'll take a little time - a few days or a week maybe - until it's usable in the popular software.
The efforts are already beginning. Check out #announcements in the GPT4All Discord I linked above - the GPT4All team are already looking to start work on making GGML support for it.