- 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.
