doesn't work on llama.cpp-b5135
#1
by
oldshun
- opened
command line is
llama-cli -ngl 99 -c 1024 -b 512 -p "<|system|>You are a helpful assistant.<|user|>Hello<|assistant|>" -m THUDM_GLM-Z1-32B-0414-Q4_K_S.gguf
it generates only "3333333..."
I've tried other models, qwq 32b, Gemma 3 27b, all works well.