set -e

SRC_DIR="/models/"

./llama.cpp/build/bin/llama-server -m ${SRC_DIR}Qwen3-30B-A3B.vocab.gguf \
-ca ${SRC_DIR}Qwen3-30B-A3B-dynamic-W8A8-W4AF16-full_layers_merged_2_chips_4096_fa_calbin_2026-04-21 --port 8031 --temp 0 --no-warmup --top-p 0.9 --seed 42 -np 1 2>&1 | tee server_run.log

# ./llama.cpp/build/bin/llama-server -m ${SRC_DIR}DeepSeek-R1-Distill-Qwen-7B.vocab.gguf \
# -ca ${SRC_DIR}DeepSeek-R1-Distill-Qwen-7B-dynamic-W8A8-W4AF16_2_chips_4096_fa --port 8031 --temp 0 --no-warmup --top-p 0.9 --seed 42 -np 1 2>&1 | tee server_run.log

# ./llama.cpp/build/bin/llama-server -m ${SRC_DIR}Qwen2.5-0.5B-Instruct.vocab.gguf \
# -ca ${SRC_DIR}Qwen2.5-0.5B-Instruct-dynamic-W8A8-W4AF16_2_chips_128_fa_calbin_2026-04-21 --port 8031 --temp 0 --no-warmup --top-p 0.9 --seed 42 -np 1 2>&1 | tee server_run.log