I implemented Laguna (XS.2) as a model in Llama.cpp
Posted by linuxid10t@reddit | LocalLLaMA | View on Reddit | 3 comments
Posted by linuxid10t@reddit | LocalLLaMA | View on Reddit | 3 comments
Weary-Step-8818@reddit
nice. the real adoption test will be whether people can reproduce the GGUF path without living inside your fork forever. model support is cool, maintenance surface is the tax.
linuxid10t@reddit (OP)
There was no model support in Llama.cpp at all. It was a lot more involved than just converting Laguna to GGUF.
linuxid10t@reddit (OP)
You will need a GGUF built with my fork of Llama.cpp. I have some linked below on Huggingface.
https://huggingface.co/linuxid10t/Laguna-XS.2-GGUF/tree/main