export LD_LIBRARY_PATH="/home/eaman/llama/bin_vulkan" ;
export LLAMA_CACHE="/home/eaman/lm/models/unsloth/Qwen3.5-4B-GGUF"
/home/eaman/llama/bin_vulkan/llama-server \
	-m /home/eaman/.lmstudio/models/unsloth/Qwen3.5-4B-GGUF/Qwen3.5-4B-Q4_K_S.gguf \
	--host 0.0.0.0  -np 1 -fa on --no-mmap --swa-full \
	-ngl 99 \
	--ctx-size 60000 \
	--jinja \
	--temp 0.3 \
	--top-p 0.8 \
	--top-k 20 \
	--min-p 0.05 \
	--cache-type-k q8_0 \
	--cache-type-v q8_0 \
	--reasoning-budget 1 \