JenZhao

#!/bin/bash
# first start it in another terminal
# vllm serve Qwen/Qwen2-VL-7B-Instruct     --swap-space 16     --disable-log-requests

# Define model variables
MODEL_NAME="Qwen/Qwen2-VL-7B-Instruct"
NUM_PROMPTS=10

# Define the output markdown file
OUTPUT_FILE="benchmark_output.md"