Vision Analysis
Back to Leaderboard

YOLOv5su

yolov5uComing Soon

one-stage detector with CSPDarknet backbone

Parameters9.1M
GFLOPs24.0
Input Size640px
Best mAP42.4%
Licensenon-permissive

Architecture

Type

one-stage

Backbone

CSPDarknet

Neck

PAFPN

Head

Anchor-Free Decoupled

Benchmark Results

Performance on COCO val2017 across different hardware configurations

HardwareRuntimeDatasetCredit / sourcemAP@50-95FPSLatencyGPU allocation
NVIDIA Jetson Orin Nano Super 8GBONNX Runtime FP32mini500 (500 images)

Run JSON

42.4%3.6275.9ms-
NVIDIA Jetson Orin Nano Super 8GBPyTorch FP32mini500 (500 images)

Run JSON

42.4%22.844.0ms66 MB
NVIDIA Jetson Orin Nano Super 8GBTensorRT FP16mini500 (500 images)

Run JSON

42.4%42.823.4ms14 MB
Raspberry Pi 5ncnn FP32mini500 (500 images)

Run JSON

42.4%6.2162.5ms-
Raspberry Pi 5ONNX Runtime FP32mini500 (500 images)

Run JSON

42.4%2.8362.4ms-
Raspberry Pi 5PyTorch FP32mini500 (500 images)

Run JSON

42.4%2.5397.4ms-

Speed Breakdown(NVIDIA Jetson Orin Nano Super 8GB)

6.2ms
254.4ms
9.4ms
Preprocess
Inference
Postprocess (NMS)
anchor-freenmsPaper: 43% mAP

Related Models (yolov5u)

YOLOv5nuYOLOv5muYOLOv5luYOLOv5xu

Run any model with one line

LibreYOLO has the best catalogue of state-of-the-art detectors, all behind one MIT-licensed Python API.

from libreyolo import LibreYOLO, SAMPLE_IMAGE

# LibreYOLO has the best catalogue of state-of-the-art models.
model = LibreYOLO("LibreRFDETRl.pt")           # RF-DETR-L (transformer flagship)
results = model(SAMPLE_IMAGE, save=True)        # run inference, save the annotated image

# Swap in any other model, same one-line API (weights auto-download):
#   LibreYOLO("LibreYOLO9c.pt")      # YOLO9-C
#   LibreYOLO("LibreYOLOXx.pt")      # YOLOX-X
#   LibreYOLO("LibreDFINEx.pt")      # D-FINE-X
#   LibreYOLO("LibreRTDETRr50.pt")   # RT-DETR-R50