# YOLOv8-X

> YOLOv8-X: one-stage object detector from the YOLOv8 family. Planned for LibreYOLO (not yet packaged).

- Family: YOLOv8 (External baseline)
- Parameters: 68.2M
- GFLOPs: 257.8
- Default input size: 640px
- Detection: anchor-free · nms
- Architecture: one-stage · CSPDarknet backbone
- Weight license: non-permissive
- Paper-reported mAP@50-95: 53.9%

## Benchmarks (COCO val2017)

| Hardware | Runtime | mAP@50-95 | FPS | Latency (ms) | VRAM (MB) |
|---|---|---|---|---|---|
| NVIDIA Jetson Orin Nano Super 8GB | ONNX Runtime FP32 | 53.0 | 0.5 | 2188.6 | - |
| NVIDIA Jetson Orin Nano Super 8GB | PyTorch FP32 | 53.0 | 5.4 | 184.8 | 379 |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP16 | 53.0 | 14.4 | 69.2 | 14 |
| Raspberry Pi 5 | HailoRT INT8 | 46.0 | 10.7 | 93.7 | - |
| Raspberry Pi 5 | ncnn FP32 | 53.0 | 1.0 | 982.0 | - |
| Raspberry Pi 5 | ONNX Runtime FP32 | 53.0 | 0.3 | 2873.9 | - |
| Raspberry Pi 5 | PyTorch FP32 | 53.0 | 0.4 | 2422.0 | - |


Source: https://www.visionanalysis.org/model/yolov8x. Benchmarks produced with LibreYOLO (https://github.com/Libre-YOLO/libreyolo).