检索置信值Yolov5

Question

我使用此代码来加载yolov5模型model = torch.hub.load('ultralytics/yolov5', 'custom', 'C:Object Detectionyolov5_object_detectionlast.pt')，我如何在图像上检测到模型后检索置信度值?

Answer 1

加载模型，向其提供图像，循环遍历预测并提取其内容。

model = torch.hub.load('ultralytics/yolov5', 'yolov5s', pretrained=True)
model.eval()
det = model(img)  # feed image
for *xyxy, conf, cls in det:  # loop through results and extract contents
print(*xyxy)
print(conf)   # what you are looking for
print(cls)