4d5b6a77ca
- Create configuration file for the custom benchmark experiment. - Implement result parsing script to handle JSONL outputs from the benchmark. - Develop run script to orchestrate the benchmark execution, including server management and health checks. - Add server start script to launch a single vLLM service on all available GPUs.