What version of llamacpp that u use to run this model?
#1
by
saipangon
- opened
Works with latest!
have tried to run this model with koboldcpp and it makes terrible response from the first message. only 2 letter response, H F H F H F like that continue. for sure something is wrong and you didn't tested it, isn't? by the way, I have tested the 4bit K.L one.
It's working for me with llama.cpp backend:
