Using the CLI

The toolkit command-line interface is available as vot (or python -m vot). It is the main entry point for workspace setup, tracker testing, evaluation, analysis, report generation, and result packaging.

General usage

vot [global-options] <command> [command-options]

Global options:

  • –debug, -d - enable debug mode and verbose logs,

  • –registry <path> - add tracker registry path (directory or registry file).

Get command help:

vot --help
vot <command> --help

Commands overview

  • initialize (alias: configure) - initialize workspace,

  • test - run tracker integration test on synthetic (or provided) sequence,

  • evaluate (alias: run) - execute experiments for one or more trackers,

  • analysis (aliases: analyse, analyze) - compute analysis outputs,

  • report (alias: document) - generate report documents,

  • pack - package tracker results for submission.

Initialize a workspace

Initialize a workspace and configure stack/dataset.

vot initialize [--workspace <path>] [--nodownload] [<stack>]

Arguments and options:

  • <stack> - optional stack identifier,

  • –workspace <path> - workspace directory (default: current directory),

  • –nodownload - skip dataset download from stack metadata.

Examples:

vot initialize vots2025/main --workspace ./workspace
vot configure --workspace ./workspace --nodownload

Test a tracker

Run tracker integration test.

vot test [--visualize] [--sequence <path>] [--ignore <ids>] <tracker>

Arguments and options:

  • <tracker> - tracker identifier from registry,

  • –visualize, -g - show visualized test run,

  • –sequence, -s <path> - use custom sequence instead of synthetic sequence,

  • –ignore <id1,id2,…> - ignore selected object identifiers.

Example:

vot test -g mytracker

Run evaluation

Run experiment stack for one or more trackers.

vot evaluate [--workspace <path>] [--force] [--persist] [--experiments <ids>] <tracker> [<tracker> ...]

Arguments and options:

  • <tracker> [<tracker> …] - one or more tracker identifiers,

  • –workspace <path> - workspace directory,

  • –force, -f - force full rerun,

  • –persist, -p - continue even when errors are encountered,

  • –experiments <id1,id2,…> - run only selected experiments.

Example:

vot evaluate --workspace ./workspace --force mytracker

Perform analysis

Compute analysis results from evaluated tracker outputs.

vot analysis [--workspace <path>] [--format json|yaml] [--name <name>] [--sequences <ids>] [--experiments <ids>] [<tracker> ...]

Arguments and options:

  • [<tracker> …] - optional list of trackers (if omitted, available results are used),

  • –workspace <path> - workspace directory,

  • –format json|yaml - analysis output format (default: json),

  • –name <name> - output name (default: generated timestamp),

  • –sequences <id1,id2,…> - include only selected sequences,

  • –experiments <id1,id2,…> - include only selected experiments.

Example:

vot analysis --workspace ./workspace --format json --sequences bag,ball mytracker

Generate a report

Generate report files from analysis/evaluation data.

vot report [--workspace <path>] [--format html|latex|plots] [--name <name>] [--sequences <ids>] [--experiments <ids>] [<tracker> ...]

Arguments and options:

  • [<tracker> …] - optional trackers to include,

  • –workspace <path> - workspace directory,

  • –format html|latex|plots - output type (default: html),

  • –name <name> - output name,

  • –sequences <id1,id2,…> - include only selected sequences,

  • –experiments <id1,id2,…> - include only selected experiments.

Example:

vot report --workspace ./workspace --format html mytracker

Pack tracker results

Create a submission archive for a tracker. This command is needed for participation in VOT challenges, but can also be used to create a standardized output archive for any tracker.

vot pack [--workspace <path>] <tracker>

Arguments and options:

  • <tracker> - tracker identifier,

  • –workspace <path> - workspace directory.

Example:

vot pack --workspace ./workspace mytracker