The answer I get on "Hello" is total disaster. This model is not usable.
Though it entered GPU of 4 GB in full.
I am using QwQ-LCoT-3B-Instruct.Q4_K_M.gguf which is totally fine.
· Sign up or log in to comment
