# RF-DETR-M

> RF-DETR-M: transformer object detector from the RF-DETR family. Available in LibreYOLO, the MIT-licensed open-source library (free for commercial use).

- Family: RF-DETR (Roboflow)
- Parameters: 33.7M
- GFLOPs: 0.0
- Default input size: 576px
- Detection: detr · nms-free
- Architecture: transformer · DINOv2 backbone
- Weight license: Apache-2.0
- Paper-reported mAP@50-95: 54.7%
- Paper: https://blog.roboflow.com/rf-detr
- Original code: https://github.com/roboflow/rf-detr

## Benchmarks (COCO val2017)

| Hardware | Runtime | mAP@50-95 | FPS | Latency (ms) | VRAM (MB) |
|---|---|---|---|---|---|
| NVIDIA Jetson Orin Nano Super 8GB | ONNX Runtime FP32 | 57.4 | 4.9 | 204.7 | - |
| NVIDIA Jetson Orin Nano Super 8GB | PyTorch FP32 | 57.4 | 5.7 | 174.5 | 193 |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP16 | 57.4 | 12.5 | 79.8 | - |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP32 | 57.4 | 11.6 | 85.9 | - |
| NVIDIA RTX 5070 Ti | ONNX Runtime FP32 | 57.4 | 59.8 | 16.7 | - |
| NVIDIA RTX 5070 Ti | PyTorch FP32 | 57.3 | 30.3 | 33.0 | 194 |
| NVIDIA RTX 5070 Ti | TensorRT FP16 | 57.4 | 62.9 | 15.9 | - |
| NVIDIA RTX 5070 Ti | TensorRT FP32 | 57.4 | 63.7 | 15.7 | - |
| Raspberry Pi 5 | ONNX Runtime FP32 | 57.4 | 0.9 | 1142.2 | - |
| Raspberry Pi 5 | PyTorch FP32 | 57.4 | 0.8 | 1281.0 | - |

## Usage with LibreYOLO

```python
from libreyolo import LibreYOLO

model = LibreYOLO("LibreRFDETRm.pt")
result = model("image.jpg", conf=0.25, iou=0.45)
```

Source: https://www.visionanalysis.org/model/rfdetr-m. Benchmarks produced with LibreYOLO (https://github.com/Libre-YOLO/libreyolo).