Parameters3.3M
GFLOPs8.9
Input Size640px
Best mAP44.1%
LicenseApache-2.0
Architecture
Type
one-stage
Backbone
ESNet
Neck
CSP-PAN
Head
GFL
Benchmark Results
Performance on COCO val2017 across different hardware configurations
| Hardware | Runtime | mAP@50-95 | FPS | Latency | VRAM |
|---|---|---|---|---|---|
| NVIDIA Jetson Orin Nano Super 8GB | ONNX Runtime FP32 | 43.6% | 4.3 | 232.1ms | - |
| NVIDIA Jetson Orin Nano Super 8GB | PyTorch FP32 | 44.1% | 11.1 | 90.1ms | 63 MB |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP16 | 44.0% | 19.1 | 52.3ms | - |
| NVIDIA Jetson Orin Nano Super 8GB | TensorRT FP32 | 44.1% | 16.4 | 60.8ms | - |
| NVIDIA RTX 5070 Ti | ONNX Runtime FP32 | 44.1% | 34.3 | 29.2ms | - |
| NVIDIA RTX 5070 Ti | PyTorch FP32 | 44.1% | 39.5 | 25.3ms | 72 MB |
| NVIDIA RTX 5070 Ti | TensorRT FP16 | 44.0% | 38.0 | 26.3ms | - |
| NVIDIA RTX 5070 Ti | TensorRT FP32 | 44.1% | 32.0 | 31.3ms | - |
Speed Breakdown(NVIDIA Jetson Orin Nano Super 8GB)
10.6ms
57.4ms
22.1ms
Preprocess
Inference
Postprocess (NMS)
Usage with LibreYOLO
from libreyolo import LibreYOLO
# Load model (auto-downloads from HuggingFace if not found locally)
model = LibreYOLO("LibrePICODETl.pt")
# Run inference
result = model("image.jpg", conf=0.25, iou=0.45)
# Process results
print(f"Found {len(result)} objects")
print(result.boxes.xyxy) # bounding boxes (N, 4)
print(result.boxes.conf) # confidence scores (N,)
print(result.boxes.cls) # class IDs (N,)anchor-freenmsPaper: 40.9% mAP