# ECDet-S

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

- Family: ECDet (Intellindust AI Lab)
- Parameters: 9.9M
- GFLOPs: 26.0
- Default input size: 640px
- Detection: detr · nms-free
- Architecture: transformer · ECViT (compact ViT) backbone
- Weight license: Apache-2.0
- Paper-reported mAP@50-95: 51.7%
- Paper: https://arxiv.org/abs/2603.18739
- Original code: https://github.com/Intellindust-AI-Lab/EdgeCrafter

## Benchmarks (COCO val2017)

| Hardware | Runtime | mAP@50-95 | FPS | Latency (ms) | VRAM (MB) |
|---|---|---|---|---|---|
| NVIDIA Jetson Orin Nano Super 8GB | ONNX Runtime FP32 | 54.3 | 5.2 | 192.3 | - |
| NVIDIA Jetson Orin Nano Super 8GB | PyTorch FP32 | 54.3 | 6.0 | 168.2 | 102 |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP16 | 54.3 | 10.6 | 94.8 | - |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP32 | 54.4 | 10.6 | 94.8 | - |
| NVIDIA RTX 5070 Ti | ONNX Runtime FP32 | 54.3 | 42.8 | 23.4 | - |
| NVIDIA RTX 5070 Ti | PyTorch FP32 | 54.4 | 27.5 | 36.4 | 103 |
| NVIDIA RTX 5070 Ti | TensorRT FP16 | 54.4 | 43.5 | 23.0 | - |
| NVIDIA RTX 5070 Ti | TensorRT FP32 | 54.4 | 63.4 | 15.8 | - |
| Raspberry Pi 5 | ONNX Runtime FP32 | 54.3 | 0.9 | 1154.9 | - |
| Raspberry Pi 5 | PyTorch FP32 | 54.3 | 0.8 | 1201.0 | - |

## Usage with LibreYOLO

```python
from libreyolo import LibreYOLO

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

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