roboflow--supervision/docs/llms.txt

# supervision

> Open-source Python library for computer vision — annotate, track, count, filter, and convert.

Supervision is a Python library by Roboflow that provides a model-agnostic `Detections` class and composable tools for object detection and segmentation workflows. It includes converters for supported outputs from Ultralytics, Roboflow Inference, Hugging Face Transformers, SAM, Detectron2, MMDetection, YOLO-NAS, PaddleDet, NCNN, Azure AI Vision, and VLM parsers.

Supervision is MIT licensed, published on PyPI, and developed in public on GitHub.

## AI Access

The documentation is static HTML and open for AI consumption. `robots.txt` explicitly allows general crawlers plus selected AI crawlers.

- GPTBot: allowed
- ClaudeBot: allowed
- PerplexityBot: allowed
- CCBot: allowed
- GoogleOther: allowed

## Install

```
pip install supervision
```

Extra: `pip install supervision[metrics]` for optional metric dependencies. Sample asset utilities are included in the base package under `supervision.assets`.

## Links

- GitHub: https://github.com/roboflow/supervision
- PyPI: https://pypi.org/project/supervision
- Docs (latest stable): https://supervision.roboflow.com/latest/
- Changelog: https://supervision.roboflow.com/latest/changelog/
- Sitemap: https://supervision.roboflow.com/sitemap.xml

## Key APIs

### sv.Detections
Core data structure for bounding boxes, masks, confidence scores, class IDs, tracker IDs, and arbitrary per-detection metadata stored in a `data` dict. The lingua franca of the entire library — every connector, annotator, and tracker accepts or returns `Detections`. Supports NumPy-style boolean indexing for filtering by class, confidence, area, and spatial regions.

### sv.BoxAnnotator, sv.MaskAnnotator, sv.LabelAnnotator
Draw bounding boxes, segmentation masks, and text labels on images. Annotators expose `annotate(scene=..., detections=...)`; pass an input image and a `Detections` object to get the annotated output. `LabelAnnotator` can use explicit `labels` or fall back to `detections["class_name"]`, class IDs, then detection indices. Colors can be assigned by class or manually specified.

### sv.ByteTrack
Object tracker wrapper that assigns persistent IDs across video frames. The built-in `sv.ByteTrack` accepts `Detections` via `update_with_detections()`, but it is deprecated in favor of `ByteTrackTracker` from the external `trackers` package, where the method is named `update()`. Use tracked `Detections` with `sv.TraceAnnotator` to visualize trajectories.

### sv.PolygonZone and sv.LineZone
Zone-based counting. `PolygonZone.trigger(detections)` returns a boolean mask for detections currently inside an arbitrary polygon. `LineZone.trigger(detections)` returns `(crossed_in, crossed_out)` arrays for line crossings and requires `detections.tracker_id` so objects can be matched across frames. Both are commonly paired with zone annotators for visualization.

### sv.DetectionDataset and sv.ClassificationDataset
For detection datasets, load, merge, split, and convert between YOLO, COCO JSON, Pascal VOC, CreateML, and LabelMe formats. Classification datasets use folder-structure import and export via `ClassificationDataset.from_folder_structure()` and `as_folder_structure()`.

### sv.InferenceSlicer
SAHI-style inference slicing: split high-resolution images into overlapping tiles, run detection on each tile, merge results with non-maximum suppression or non-maximum merge. Configure tile overlap in pixels with `overlap_wh`.

### supervision.metrics.MeanAveragePrecision and sv.ConfusionMatrix
Benchmarking tools. For mAP@0.5:0.95, use `supervision.metrics.MeanAveragePrecision` with `update()` and `compute()` rather than the deprecated top-level `sv.MeanAveragePrecision.from_detections()`. `ConfusionMatrix.from_detections(predictions=predictions, targets=targets, classes=classes)` generates a confusion matrix for detection results.

### sv.CSVSink and sv.JSONSink
Export detection results to structured files. Use `CSVSink` and `JSONSink` as context managers, call `sink.append(detections, custom_data=...)`, and they write one row/object per detection with box coordinates, confidence, class ID, tracker ID, and data fields.

## How-To Guides

- Detect and annotate: https://supervision.roboflow.com/latest/how_to/detect_and_annotate/
- Track objects: https://supervision.roboflow.com/latest/how_to/track_objects/
- Detect small objects: https://supervision.roboflow.com/latest/how_to/detect_small_objects/
- Filter detections: https://supervision.roboflow.com/latest/how_to/filter_detections/
- Save detections: https://supervision.roboflow.com/latest/how_to/save_detections/
- Count in zone: https://supervision.roboflow.com/latest/how_to/count_in_zone/
- Benchmark a model: https://supervision.roboflow.com/latest/how_to/benchmark_a_model/
- Process datasets: https://supervision.roboflow.com/latest/how_to/process_datasets/

## Reference Documentation

- Detections (detection/core): https://supervision.roboflow.com/latest/detection/core/
- Annotators (detection/annotators): https://supervision.roboflow.com/latest/detection/annotators/
- CompactMask (detection/compact_mask): https://supervision.roboflow.com/latest/detection/compact_mask/
- Format Converters (detection/utils/converters): https://supervision.roboflow.com/latest/detection/utils/converters/
- IoU and NMS (detection/utils/iou_and_nms): https://supervision.roboflow.com/latest/detection/utils/iou_and_nms/
- Boxes (detection/utils/boxes): https://supervision.roboflow.com/latest/detection/utils/boxes/
- Masks (detection/utils/masks): https://supervision.roboflow.com/latest/detection/utils/masks/
- Polygons (detection/utils/polygons): https://supervision.roboflow.com/latest/detection/utils/polygons/
- VLMs (detection/utils/vlms): https://supervision.roboflow.com/latest/detection/utils/vlms/
- Keypoint Core (keypoint/core): https://supervision.roboflow.com/latest/keypoint/core/
- Keypoint Annotators (keypoint/annotators): https://supervision.roboflow.com/latest/keypoint/annotators/
- Classification Core (classification/core): https://supervision.roboflow.com/latest/classification/core/
- ByteTrack Tracker (trackers): https://supervision.roboflow.com/latest/trackers/
- Datasets Core (datasets/core): https://supervision.roboflow.com/latest/datasets/core/
- mAP (metrics/mean_average_precision): https://supervision.roboflow.com/latest/metrics/mean_average_precision/
- mAR (metrics/mean_average_recall): https://supervision.roboflow.com/latest/metrics/mean_average_recall/
- Precision (metrics/precision): https://supervision.roboflow.com/latest/metrics/precision/
- Recall (metrics/recall): https://supervision.roboflow.com/latest/metrics/recall/
- F1 Score (metrics/f1_score): https://supervision.roboflow.com/latest/metrics/f1_score/
- Common Values (metrics/common_values): https://supervision.roboflow.com/latest/metrics/common_values/
- Line Zone (detection/tools/line_zone): https://supervision.roboflow.com/latest/detection/tools/line_zone/
- Polygon Zone (detection/tools/polygon_zone): https://supervision.roboflow.com/latest/detection/tools/polygon_zone/
- Inference Slicer (detection/tools/inference_slicer): https://supervision.roboflow.com/latest/detection/tools/inference_slicer/
- Detection Smoother (detection/tools/smoother): https://supervision.roboflow.com/latest/detection/tools/smoother/
- Save Detections Tool (detection/tools/save_detections): https://supervision.roboflow.com/latest/detection/tools/save_detections/
- Video Utils (utils/video): https://supervision.roboflow.com/latest/utils/video/
- Image Utils (utils/image): https://supervision.roboflow.com/latest/utils/image/
- Iterable Utils (utils/iterables): https://supervision.roboflow.com/latest/utils/iterables/
- Notebook Utils (utils/notebook): https://supervision.roboflow.com/latest/utils/notebook/
- File Utils (utils/file): https://supervision.roboflow.com/latest/utils/file/
- Draw Utils (utils/draw): https://supervision.roboflow.com/latest/utils/draw/
- Geometry (utils/geometry): https://supervision.roboflow.com/latest/utils/geometry/
- Assets (assets): https://supervision.roboflow.com/latest/assets/

## Cookbooks

- Object tracking: https://supervision.roboflow.com/latest/cookbooks/#object-tracking
- Count objects crossing line: https://supervision.roboflow.com/latest/cookbooks/#count-objects-crossing-the-line
- Zero-shot object detection with YOLO-World: https://supervision.roboflow.com/latest/cookbooks/#zero-shot-object-detection-with-yolo-world
- SAHI small object detection: https://supervision.roboflow.com/latest/cookbooks/#small-object-detection-with-sahi

## FAQ

### What is supervision?

Supervision is an open-source Python library by Roboflow for computer vision workflows. It provides a unified `Detections` class with converters for supported detection, segmentation, and VLM outputs, plus tools for annotation, tracking, zone counting, dataset management, and model benchmarking.

### How do I install supervision?

Install with `pip install supervision`. For optional metric dependencies use `pip install supervision[metrics]`. Sample asset utilities are included in the base package under `supervision.assets`. The current package metadata requires Python 3.10+.

### What can I do with supervision?

Annotate images and video with bounding boxes, masks, and labels; track objects across frames with persistent IDs; count detections inside polygon zones or line crossings; filter and query detection results; load, split, and convert detection datasets between YOLO, COCO, Pascal VOC, and LabelMe formats; manage classification datasets with folder structures; and benchmark model performance with mAP and confusion matrices.

### Is supervision free to use?

Yes. Supervision is free and open-source under the MIT license. Source code is at https://github.com/roboflow/supervision.

### Which object detection models work with supervision?

Supervision is model-agnostic and works with supported outputs from Ultralytics YOLO, Roboflow Inference, Hugging Face Transformers, SAM, Detectron2, MMDetection, YOLO-NAS, PaddleDet, NCNN, Azure AI Vision, and VLM parsers such as Florence-2, PaliGemma, Qwen VL, Gemini, DeepSeek VL 2, and Moondream. Keypoint outputs have separate converters, including MediaPipe.

### How do I benchmark a model with supervision?

Use `supervision.metrics.mean_average_precision.MeanAveragePrecision` for mAP — accumulate predictions and ground truth with `update(...)` then call `compute()`. For confusion matrices, use `sv.ConfusionMatrix.from_detections(predictions=predictions, targets=targets, classes=classes)`. See the Benchmark a Model how-to guide for a complete walkthrough.

### How do I track objects across video frames?

Use a tracker to assign persistent IDs. The built-in `sv.ByteTrack` wrapper accepts `Detections` with `update_with_detections()`, but it is deprecated in favor of `ByteTrackTracker` from the external `trackers` package. Combine tracked detections with `sv.TraceAnnotator` to visualize trajectories.

### What dataset formats does supervision support?

For detection datasets, supervision supports YOLO, COCO JSON, Pascal VOC, CreateML, and LabelMe. Use `DetectionDataset.from_yolo()`, `from_coco()`, `from_pascal_voc()`, `from_createml()`, or `from_labelme()` to load, and `as_yolo()`, `as_coco()`, `as_pascal_voc()`, `as_createml()`, or `as_labelme()` to save. For classification datasets, use `ClassificationDataset.from_folder_structure()` and `as_folder_structure()`.

### How do I count objects in a zone?

Use `sv.PolygonZone` for arbitrary polygon zones. Use `sv.LineZone` for line-crossing counts after assigning tracker IDs, because `LineZone` needs `detections.tracker_id` to match objects across frames.

### How do I detect small objects with supervision?

Use `sv.InferenceSlicer` to split high-resolution images into overlapping tiles, run detection on each tile, and merge results with non-maximum suppression. Configure tile overlap in pixels with `overlap_wh`. See the Detect Small Objects how-to guide.

## Benchmarking

Supervision includes `supervision.metrics.MeanAveragePrecision` and `sv.ConfusionMatrix` for benchmarking object detection models. A curated [Model Leaderboard](https://leaderboard.roboflow.com/) compares YOLOv8, YOLOv11, and other models on standard datasets. The leaderboard repository is open source at https://github.com/roboflow/model-leaderboard.

## License

MIT — https://github.com/roboflow/supervision/blob/develop/LICENSE.md

## Citation

```bibtex
@software{supervision,
  author = {Roboflow},
  title = {Supervision: Computer Vision Toolkit},
  url = {https://github.com/roboflow/supervision},
  year = {2023}
}
```

## Versioning

Stable release docs: https://supervision.roboflow.com/latest/
Development branch: https://supervision.roboflow.com/develop/