Back to Leaderboard

YOLOv11-L

yolov11

one-stage detector with C3K2 backbone

Parameters

25.3M

FLOPs

87.0G

Input Size

640px

License

MIT

Architecture

Type

one-stage

Backbone

C3K2

Neck

SPPF

Head

Decoupled

Benchmark Results
Performance on COCO val2017 across different hardware configurations
HardwaremAP@50-95FPSLatencyVRAM
NVIDIA A100 (TensorRT FP16)53.4%95.110.5ms1168 MB
NVIDIA T4 (TensorRT FP16)53.4%37.626.6ms1252 MB
CPU (ONNX Runtime)53.5%4.6218.3ms1194 MB
Speed Breakdown (A100 TensorRT)
End-to-end latency breakdown showing preprocessing, inference, and postprocessing times
1.5ms
6.5ms
2.6ms
Preprocess
Inference
Postprocess (NMS)
Usage with LibreYOLO
from libreyolo import YOLO

# Load model
model = YOLO.from_pretrained("https://huggingface.co/Libre-YOLO/yolov11l")

# Run inference
results = model.predict("image.jpg")

# Process results
for box in results.boxes:
    print(f"Class: {box.cls}, Confidence: {box.conf:.2f}")
latesthigh-accuracy
Related Models (yolov11)