Back to Leaderboard

YOLOv10-L

yolov10

one-stage detector with CSPDarknet backbone

Parameters

24.4M

FLOPs

120.3G

Input Size

640px

License

MIT

Architecture

Type

one-stage

Backbone

CSPDarknet

Neck

PAFPN

Head

NMS-Free

Benchmark Results
Performance on COCO val2017 across different hardware configurations
HardwaremAP@50-95FPSLatencyVRAM
NVIDIA A100 (TensorRT FP16)53.3%120.98.3ms1102 MB
NVIDIA T4 (TensorRT FP16)53.1%46.621.5ms1138 MB
CPU (ONNX Runtime)53.3%5.4184.9ms1271 MB
Speed Breakdown (A100 TensorRT)
End-to-end latency breakdown showing preprocessing, inference, and postprocessing times
1.4ms
6.8ms
0.1ms
Preprocess
Inference
Postprocess (NMS)
Usage with LibreYOLO
from libreyolo import YOLO

# Load model
model = YOLO.from_pretrained("https://huggingface.co/Libre-YOLO/yolov10l")

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

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