README.md

---
license: apache-2.0

pipeline_tag: object-detection

library_name: rfdetr

datasets:
  - Voxel51/VisDrone2019-DET

tags:
  - object-detection
  - detectionbench
  - rfdetr
  - pytorch
  - computer-vision
  - aerial-imagery
  - drone
  - visdrone
  - detr
  - roboflow
  - transformer
metrics:
  - map50
  - map50-95
  - precision
  - recall
  - f1

base_model: "Roboflow/rf-detr-medium"

model-index:
  - name: RF-DETR Medium Finetuned on VisDrone-DET
    results:
      - task:
          type: object-detection
          name: Object Detection
        dataset:
          name: VisDrone-DET
          type: visdrone
        metrics:
          - type: mAP50
            value: 36.82
            name: mAP@50 (test split)
          - type: mAP50-95
            value: 20.14
            name: mAP@50-95 (test split)
          - type: precision
            value: 64.0
            name: Precision (test split)
          - type: recall
            value: 47.05
            name: Recall (test split)
        source:
          url: https://github.com/dronefreak/DetectionBench
          name: DetectionBench
---


# RF-DETR Medium Finetuned on VisDrone-DET

Fine-tuned RF-DETR Medium object detector on the **VisDrone-DET** benchmark dataset, trained and evaluated as part of [DetectionBench](https://github.com/dronefreak/DetectionBench) -- a framework for reproducibly benchmarking modern object detectors with identical training recipes and evaluation metrics across multiple real-world datasets.

<br>

<!-- ROW 1: Identity & Tech Stack -->
<div style="display: flex; justify-content: center; align-items: center; gap: 8px; margin-bottom: 8px; flex-wrap: wrap;">
  <img src="https://img.shields.io/badge/Task-Object_Detection-blue?style=flat-square" alt="Task">
  <img src="https://img.shields.io/badge/Framework-RF--DETR-0aa1a7?style=flat-square" alt="Framework">
  <img src="https://img.shields.io/badge/Base_Model-RF--DETR_Medium-purple?style=flat-square" alt="Base Model">
</div>

<!-- ROW 2: Performance Metrics -->
<div style="display: flex; justify-content: center; align-items: center; gap: 8px; margin-bottom: 8px; flex-wrap: wrap;">
  <img src="https://img.shields.io/badge/mAP@50-36.82%25-success?style=flat-square" alt="mAP@50">
  <img src="https://img.shields.io/badge/mAP@50:95-20.14%25-orange?style=flat-square" alt="mAP@50:95">
  <img src="https://img.shields.io/badge/Params-33.7M-lightgrey?style=flat-square" alt="Params">
</div>

<!-- ROW 3: Metadata -->
<div style="display: flex; justify-content: center; align-items: center; gap: 8px; margin-bottom: 24px; flex-wrap: wrap;">
  <img src="https://img.shields.io/badge/License-Apache--2.0-lightgrey?style=flat-square" alt="License">
  <a href="https://github.com/dronefreak/DetectionBench"><img src="https://img.shields.io/badge/Source-DetectionBench-black?style=flat-square" alt="Source"></a>
</div>

---

## Detection Showcase

<p align="center">
  <img src="visdrone_rfdetr-medium_showcase.jpg" alt="VisDrone-DET Detection Demo" width="900">
</p>

---

## Performance

| Metric     | Score (%)       |
| ---------- | --------------- |
| mAP@50     | 36.82     |
| mAP@50-95  | 20.14   |
| Precision  | 64.0 |
| Recall     | 47.05    |
| F1 Score   | 54.24        |
| Parameters | 33.7M    |
| FLOPs      | N/A (not published upstream)     |

---

## Evaluation Protocol

Metrics reported in this model card are computed on the VisDrone-DET **test** split, using DetectionBench's standard evaluation pipeline (`detectionbench-evaluate`).

---

## VisDrone-DET Model Zoo

Every model DetectionBench has trained and evaluated on VisDrone-DET so far, for full transparency -- see [DetectionBench](https://github.com/dronefreak/DetectionBench) for the smaller, curated comparison set used on the project README.

| Model                | mAP@50        | mAP@50-95       | Precision         | Recall         |
| --------------------- | ------------- | --------------- | ----------------- | -------------- |
| RF-DETR Medium | 36.82 | 20.14 | 64.0 | 47.05 |
| RF-DETR Small | 33.25 | 17.88 | 62.62 | 43.51 |
| RF-DETR Nano | 25.15 | 12.77 | 58.99 | 35.0 |
---

## External VisDrone-DET Comparison

The YOLO/RT-DETR rows below were trained and evaluated on VisDrone2019-DET's test split via a separate companion codebase ([VisDrone-dataset-python-toolkit](https://github.com/dronefreak/VisDrone-dataset-python-toolkit)), not reproduced inside DetectionBench -- included here purely for context. The RF-DETR rows are this repository's own DetectionBench-trained runs (see the Model Zoo table above).

| Model | mAP@50 | mAP@50-95 | Precision | Recall |
| ----- | ------ | --------- | --------- | ------ |
| YOLOv9e | 40.02 | 23.73 | 54.78 | 42.42 |
| YOLOv11x | 38.44 | 22.6 | 52.41 | 41.43 |
| YOLOv26x | 38.33 | 22.48 | 52.91 | 41.06 |
| YOLOv11l | 37.14 | 21.85 | 51.87 | 40.33 |
| YOLOv10x | 37.24 | 21.81 | 52.59 | 39.84 |
| YOLOv26l | 37.65 | 21.75 | 51.6 | 40.42 |
| YOLOv9c | 37.22 | 21.73 | 51.99 | 39.77 |
| YOLOv8x | 36.81 | 21.52 | 51.91 | 39.78 |
| YOLOv26m | 36.67 | 21.22 | 51.03 | 39.79 |
| YOLOv10l | 35.95 | 21.09 | 52.13 | 38.48 |
| YOLOv11m | 36.35 | 21.02 | 50.24 | 39.46 |
| YOLOv9m | 36.19 | 20.95 | 51.05 | 39.12 |
| RF-DETR-Medium | 36.82 | 20.14 | 64.0 | 47.05 |
| YOLOv8m | 34.39 | 19.95 | 48.18 | 38.2 |
| YOLOv9s | 33.52 | 19.26 | 46.16 | 37.43 |
| YOLOv11s | 32.3 | 18.47 | 45.49 | 35.31 |
| YOLOv8s | 31.95 | 18.24 | 45.99 | 35.49 |
| YOLOv26s | 32.1 | 18.06 | 45.75 | 35.05 |
| RF-DETR-Small | 33.25 | 17.88 | 62.62 | 43.51 |
| YOLOv9t | 29.09 | 16.22 | 42.57 | 32.66 |
| YOLOv8n | 28.18 | 15.77 | 40.86 | 31.81 |
| YOLOv11n | 27.59 | 15.46 | 39.58 | 31.74 |
| YOLOv10n | 27.65 | 15.32 | 41.02 | 31.68 |
| YOLOv26n | 26.73 | 14.64 | 38.6 | 31.14 |
| RF-DETR-Nano | 25.15 | 12.77 | 58.99 | 35.0 |
| rt_detr_l | 21.68 | 9.34 | 35.76 | 26.3 |
Source: https://huggingface.co/collections/dronefreak/visdrone-object-detection-model-zoo

---

## Per-Class Performance

| Class                      | mAP@50          | mAP@50-95         |
| -------------------------- | --------------- | ----------------- |
| pedestrian       | 28.15 | 11.15 |
| people       | 23.02 | 8.2 |
| bicycle       | 16.46 | 6.71 |
| car       | 73.13 | 44.06 |
| van       | 39.72 | 24.55 |
| truck       | 46.54 | 29.03 |
| tricycle       | 24.14 | 12.39 |
| awning-tricycle       | 20.15 | 10.71 |
| bus       | 62.35 | 41.68 |
| motor       | 34.53 | 12.93 |
| others       | 0.0 | 0.0 |
---

## Evaluation Visualizations

This model was evaluated with [Supervision](https://github.com/roboflow/supervision)'s detection metrics, which report mAP/Precision/Recall directly but don't produce PR-curve, F1-curve, or confusion-matrix plot images the way Ultralytics' validator does. See the Performance table above for Precision/Recall/F1 and the per-class table above for the full per-class mAP breakdown.

---

## Dataset

This model was trained on **VisDrone-DET**. For the full dataset description, provenance, license, and citation, see the dataset card:

https://huggingface.co/datasets/Voxel51/VisDrone2019-DET

### Classes

* pedestrian
* people
* bicycle
* car
* van
* truck
* tricycle
* awning-tricycle
* bus
* motor
* others
---

## Usage

### Install Dependencies

```bash
pip install rfdetr huggingface_hub
```

### Load Model from Hugging Face

```python
from huggingface_hub import hf_hub_download
import rfdetr

weights = hf_hub_download(
    repo_id="dronefreak/visdrone-rfdetr-medium",
    filename="checkpoint_best_total.pth"
)

model = rfdetr.RFDETRMedium(pretrain_weights=weights)
```

### Run Inference

```python
detections = model.predict("image.jpg", threshold=0.25)
```
---

## Training Configuration

| Setting          | Value                           |
| ---------------- | -------------------------------- |
| Dataset          | VisDrone-DET       |
| Framework        | RF-DETR                  |
| Training Toolkit | DetectionBench                   |
| Epochs (configured max) | 500 |
| Epochs (actually trained) | 141 |
| Early Stopping Patience | 100 |
| Batch Size | 10 |
| Resolution | 576 |
| Optimizer | adamw |
| Learning Rate | 0.0001 |
| Seed | 42 |
---

## Repository Contents

```text
checkpoint_best_total.pth
metrics.csv
config.json
visdrone_rfdetr-medium_showcase.jpg
README.md
```

---

## Related Resources

* [VisDrone-DET dataset card](https://huggingface.co/datasets/Voxel51/VisDrone2019-DET) on Hugging Face
* [DetectionBench](https://github.com/dronefreak/DetectionBench) -- reproducible benchmarks for modern object detectors on real-world datasets

---

## Training Framework

This model was trained using [DetectionBench](https://github.com/dronefreak/DetectionBench), an open-source framework for benchmarking object detectors across multiple real-world datasets with a common pipeline.

Features include:

* A dataset-adapter registry for converting real-world datasets into a canonical format
* Identical training/evaluation recipes across model families (Ultralytics YOLO/RT-DETR, RF-DETR)
* Hardware profiling (latency, FPS, VRAM, parameters, FLOPs)
* One-command reproducibility via versioned Hydra configs

If you find this model useful, please consider starring the repository.

---

## Known Limitations

* Severe class imbalance: `car` (42.21%) and `pedestrian` (23.12%) account for two-thirds of all annotated boxes in the training set, while `awning-tricycle` (0.95%) and `tricycle` (1.40%) are rare -- the `others` class has zero annotated instances in the training set entirely and is effectively unusable (always 0 AP).
* Extreme small-object density: ~53 annotated boxes per image on average, with roughly 69% of boxes covering under 0.1% of the image area -- consistent with VisDrone's aerial small-object detection challenge (objects captured from significant altitude).
* The original authors license VisDrone under CC BY-NC-SA 3.0 -- non-commercial research use only (see the dataset's homepage); this applies to any model trained on it, not only the raw images.
* These RF-DETR checkpoints were trained/evaluated directly through DetectionBench. The YOLO/RT-DETR rows in the External VisDrone Model Zoo comparison below were trained via a separate companion codebase, not reproduced inside DetectionBench -- see that collection for their own training details and caveats.
---

## Citation

If you use this model in your research, please consider citing:

1. The VisDrone-DET dataset (see below)
2. The original RF-DETR Medium architecture (see below)
3. DetectionBench, the training/evaluation framework used to produce this checkpoint
```
@article{zhu2018vision,
  title={Vision meets drones: A challenge},
  author={Zhu, Pengfei and Wen, Longyin and Bian, Xiao and Ling, Haibin and Hu, Qinghua},
  journal={arXiv preprint arXiv:1804.07437},
  year={2018}
}
```

```bibtex
@inproceedings{robinson2026rfdetr,
  title     = {RF-DETR: Real-Time Detection Transformer},
  author    = {Robinson, Isaac and Robicheaux, Peter and Popov, Matvei and Ramanan, Deva and Peri, Neehar},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year      = {2026},
  url       = {https://arxiv.org/abs/2511.09554}
}

@article{oquab2023dinov2,
  title={DINOv2: Learning Robust Visual Features without Supervision},
  author={Oquab, Maxime and Darcet, Timoth{\'e}e and Moutakanni, Theo and Vo, Huy and Szafraniec, Marc and Khalidov, Vasil and Fernandez, Pierre and Haziza, Daniel and Massa, Francisco and El-Nouby, Alaaeldin and others},
  journal={arXiv preprint arXiv:2304.07193},
  year={2023}
}
```
```bibtex
@software{Saksena_DetectionBench_2026,
  author = {Saksena, Saumya Kumaar},
  title = {DetectionBench: Reproducible Benchmarks for Modern Object Detectors on Real-World Datasets},
  url = {https://github.com/dronefreak/DetectionBench},
  year = {2026}
}
```