# DEIMv2-S

> DEIMv2-S: transformer object detector from the DEIMv2 family. Available in LibreYOLO, the MIT-licensed open-source library (free for commercial use).

- Family: DEIMv2 (Intellindust AI Lab)
- Parameters: 9.7M
- GFLOPs: 25.6
- Default input size: 640px
- Detection: detr · nms-free
- Architecture: transformer · DINOv3-distilled ViT backbone
- Weight license: Apache-2.0
- Paper-reported mAP@50-95: 50.9%
- Paper: https://arxiv.org/abs/2509.20787
- Original code: https://github.com/Intellindust-AI-Lab/DEIMv2

## Benchmarks (COCO val2017)

| Hardware | Runtime | mAP@50-95 | FPS | Latency (ms) | VRAM (MB) |
|---|---|---|---|---|---|
| NVIDIA Jetson Orin Nano Super 8GB | ONNX Runtime FP32 | 53.0 | 5.1 | 194.4 | - |
| NVIDIA Jetson Orin Nano Super 8GB | PyTorch FP32 | 53.0 | 5.7 | 175.8 | 101 |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP16 | 53.0 | 10.5 | 95.5 | - |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP32 | 53.0 | 10.4 | 95.7 | - |
| NVIDIA RTX 5070 Ti | ONNX Runtime FP32 | 53.0 | 39.6 | 25.2 | - |
| NVIDIA RTX 5070 Ti | PyTorch FP32 | 53.0 | 24.3 | 41.2 | 104 |
| NVIDIA RTX 5070 Ti | TensorRT FP16 | 53.0 | 50.4 | 19.8 | - |
| NVIDIA RTX 5070 Ti | TensorRT FP32 | 53.0 | 52.5 | 19.0 | - |
| Raspberry Pi 5 | ONNX Runtime FP32 | 53.0 | 0.9 | 1162.8 | - |
| Raspberry Pi 5 | PyTorch FP32 | 53.0 | 0.8 | 1190.9 | - |

## Usage with LibreYOLO

```python
from libreyolo import LibreYOLO

model = LibreYOLO("LibreDEIMv2s.pt")
result = model("image.jpg", conf=0.25, iou=0.45)
```

Source: https://www.visionanalysis.org/model/deimv2-s. Benchmarks produced with LibreYOLO (https://github.com/Libre-YOLO/libreyolo).