OpenAI models

bash download_data.sh

python -m sglang.launch_server --model-path meta-llama/Llama-2-7b-chat-hf --port 30000

python3 bench_sglang.py --nsub 10

# OpenAI models
python3 bench_sglang.py --backend gpt-3.5-turbo --parallel 8