# RT-DETR-R50

> RT-DETR-R50: transformer object detector from the RT-DETR family. Available in LibreYOLO, the MIT-licensed open-source library (free for commercial use).

- Family: RT-DETR (Baidu)
- Parameters: 42.0M
- GFLOPs: 136.0
- Default input size: 640px
- Detection: detr · nms-free
- Architecture: transformer · ResNet-50 backbone
- Weight license: Apache-2.0
- Paper-reported mAP@50-95: 53.1%
- Paper: https://arxiv.org/abs/2304.08069
- Original code: https://github.com/lyuwenyu/RT-DETR

## Benchmarks (COCO val2017)

| Hardware | Runtime | mAP@50-95 | FPS | Latency (ms) | VRAM (MB) |
|---|---|---|---|---|---|
| NVIDIA A100 | PyTorch FP32 | 52.7 | 22.7 | 44.1 | 322 |
| NVIDIA Jetson Orin Nano Super 8GB | ONNX Runtime FP32 | 55.9 | 4.4 | 228.0 | - |
| NVIDIA Jetson Orin Nano Super 8GB | PyTorch FP32 | 55.9 | 4.3 | 234.6 | 323 |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP16 | 55.7 | 23.6 | 42.4 | - |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP32 | 55.9 | 12.6 | 79.1 | - |
| NVIDIA RTX 5070 Ti | ONNX Runtime FP32 | 55.9 | 62.8 | 15.9 | - |
| NVIDIA RTX 5070 Ti | PyTorch FP32 | 55.9 | 29.8 | 33.6 | 326 |
| NVIDIA RTX 5070 Ti | TensorRT FP16 | 55.8 | 97.4 | 10.3 | - |
| NVIDIA RTX 5070 Ti | TensorRT FP32 | 55.9 | 87.4 | 11.4 | - |

## Usage with LibreYOLO

```python
from libreyolo import LibreYOLO

model = LibreYOLO("LibreRTDETRr50.pt")
result = model("image.jpg", conf=0.25, iou=0.45)
```

Source: https://www.visionanalysis.org/model/rtdetr-r50. Benchmarks produced with LibreYOLO (https://github.com/Libre-YOLO/libreyolo).