在写一个印章检测的测试代码时出现问题

Question

在写一个印章检测代码的时候出现了问题

```python
import os
import base64
import torch.hub

root_path = os.getcwd()
yolo_model_path = root_path + '/yolov5/models/seal_detect_best.pt'


class Infer_main:
    def __init__(self, yolo_model_path):
        self.seal_infer = self.yolo_model_load()
        self.model_path = yolo_model_path

    def yolo_model_load(self):
        model = torch.hub.load('D:/workspace/seal/yolov5', 'custom', path=self.model_path, source='local',
                               force_reload=True)
        model.conf = 0.4
        model.eval()
        return model

    def predict(self, images):
        result = {'seal_page': []}
        seal_page = []
        for i in range(len(images)):
            yolo_res = self.seal_infer(images[i], size=640)
            yolo_res_list = yolo_res.pandas().xyxy[0].values.tolist()
            if yolo_res_list:
                for a in range(len(yolo_res_list)):
                    seal_page.append(str(i+1))
        print(type(result['seal_page']))
        result['seal_page'] = seal_page
        print(type(seal_page))

if __name__ == '__main__':
    pic = open("8.png", "rb")
    yolo_model_path='D:/workspace/seal/yolov5/models'
    a = Infer_main(yolo_model_path)
    a.predict(pic.read())

我在写下面的测试代码时出现了一些错误，报错信息如下：

```python
Traceback (most recent call last):
  File "D:\workspace\seal\seal_detection.py", line 37, in <module>
    a = Infer_main(yolo_model_path)
  File "D:\workspace\seal\seal_detection.py", line 11, in __init__
    self.seal_infer = self.yolo_model_load()
  File "D:\workspace\seal\seal_detection.py", line 15, in yolo_model_load
    model = torch.hub.load('D:/workspace/seal/yolov5', 'custom', path=self.model_path, source='local',
AttributeError: 'Infer_main' object has no attribute 'model_path'

我的最终目的是正确的把图片传递给AI能力检测是否有印章

Answer 1

Infer_main中没有定义model_path属性，或者定义的不对，导致代码没有检测到。你要在init方法中把 self.model_path = yolo_model_path放在第一行，因为你先调用yolo_model_load方法的话，该方法中用到了model_path ，所以需要先调用 self.model_path = yolo_model_path，再调用 self.seal_infer = self.yolo_model_load()，这样yolo_model_load方法中才能检测到你定义的model_path ：
具体修改如下：

    def __init__(self, yolo_model_path):
        self.model_path = yolo_model_path
        self.seal_infer = self.yolo_model_load()

如有帮助，点个采纳哦