Skip to main content
Panlabel provides a comprehensive command-line interface for converting, validating, analyzing, and manipulating object detection annotation datasets.

​
Command Structure

All Panlabel commands follow this pattern: 
panlabel <command> [options] [arguments]

​
Getting Help

panlabel --help              # Show all available commands
panlabel <command> --help   # Show help for a specific command
panlabel -V                  # Show version

​
Available Commands

Panlabel provides six main commands for working with annotation datasets:

convert

Convert datasets between different annotation formats

validate

Validate datasets for errors and warnings

stats

Generate rich dataset statistics and reports

diff

Compare two datasets semantically

sample

Create subset datasets with various sampling strategies

list-formats

List all supported formats and their capabilities

​
Common Patterns

​
Format Specification

Most commands accept format arguments to specify input and output formats:
  • Supported formats: ir-json, coco, cvat, label-studio, tfod, yolo, voc
  • Aliases available: coco-json, cvat-xml, label-studio-json, ls, tfod-csv, ultralytics, yolov8, yolov5, pascal-voc, voc-xml
  • Auto-detection: Use --from auto to automatically detect the input format

​
Validation Modes

Many commands support validation options:
  • --strict - Treat warnings as errors
  • --no-validate - Skip validation entirely (convert only)

​
Output Formats

Commands that generate reports typically support multiple output formats:
  • --output text - Human-readable text (default)
  • --output json - Machine-readable JSON
  • --output html - Self-contained HTML report (stats only)

​
Quick Examples

​
Convert with auto-detection

panlabel convert --from auto --to coco -i /data/my_yolo -o out.json

​
Validate a dataset

panlabel validate /data/annotations.json --format coco --strict

​
Generate statistics

panlabel stats dataset.coco.json --output html > report.html

​
Compare datasets

panlabel diff a.json b.json --match-by iou --detail

​
Sample a subset

panlabel sample -i large.json -o small.json -n 100 --strategy stratified

​
Next Steps

Convert Command

Learn about format conversion options

Supported Formats

See all supported annotation formats