YOLOv3

YOLOv3#

from pathlib import Path
from ultralytics import YOLO
from ultralytics import settings

temp_dir = Path(".temp")
temp_dir.mkdir(exist_ok=True)
# 修改配置
settings.update({
    "datasets_dir": "/media/pc/data/lxw/datasets", # 存储数据集的目录
    "weights_dir": f"{temp_dir}/weights", # 存储模型权重的目录
    "runs_dir": f"{temp_dir}/runs", # 存储实验运行的目录
})
model = YOLO(f"{temp_dir}/yolov3-tinyu.pt")

model.export(format="onnx")

Ultralytics 8.3.75 🚀 Python-3.12.2 torch-2.5.1 CPU (Intel Xeon E5-2678 v3 2.50GHz)
YOLOv3-tiny summary (fused): 63 layers, 12,168,784 parameters, 0 gradients, 19.0 GFLOPs

PyTorch: starting from '.temp/yolov3-tinyu.pt' with input shape (1, 3, 640, 640) BCHW and output shape(s) (1, 84, 2000) (23.3 MB)

ONNX: starting export with onnx 1.17.0 opset 19...
ONNX: slimming with onnxslim 0.1.46...
ONNX: export success ✅ 8.0s, saved as '.temp/yolov3-tinyu.onnx' (46.5 MB)

Export complete (12.0s)
Results saved to /media/pc/data/lxw/ai/torch-book/doc/ecosystem/ultralytics/.temp
Predict:         yolo predict task=detect model=.temp/yolov3-tinyu.onnx imgsz=640  
Validate:        yolo val task=detect model=.temp/yolov3-tinyu.onnx imgsz=640 data=coco.yaml  
Visualize:       https://netron.app

'.temp/yolov3-tinyu.onnx'