Skip to content

CLI reference

Run an experiment

vllm-opt --config experiment.yaml
vllm-opt -c experiment.yaml --verbose

--verbose streams vLLM server and selected benchmark output and overrides the YAML logging level with DEBUG for that invocation.

Retry selected trials

vllm-opt retry --run runs/NAME/RUN_ID \
  --trial trial-0001 --trial trial-0004

The source run and trials are validated first. The command creates a new linked run containing only the requested fixed configurations.

Display reproduction commands

vllm-opt reproduce --run runs/NAME/RUN_ID --trial trial-0001

This reads stored artifacts and prints the vLLM and benchmark commands. It does not start a server or execute the commands.

Regenerate a report offline

vllm-opt report --run runs/NAME/RUN_ID
vllm-opt report --run runs/NAME/RUN_ID --output /tmp/vllm-opt-report

The command reads the immutable run and trial results without loading a model or invoking vLLM or GuideLLM. By default it creates a timestamped directory under RUN_ID/regenerated/; an existing destination is never overwritten.

Use vllm-opt --help or a subcommand's --help for the complete option list.