RapidUnWrap 测试代码无法运行

[nissansz]

Python 3.8.8 (tags/v3.8.8:024d805, Feb 19 2021, 13:18:16) [MSC v.1928 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.

import cv2

from inference import DocUnwrapper
Traceback (most recent call last):
File "", line 1, in
ModuleNotFoundError: No module named 'inference'
img_path = r"F:\page4.jpg"
doc_unwrapper = DocUnwrapper(img_path)
Traceback (most recent call last):
File "", line 1, in
NameError: name 'DocUnwrapper' is not defined
unwrapped_img, elapse = doc_unwrapper(img_path)
Traceback (most recent call last):
File "", line 1, in
NameError: name 'doc_unwrapper' is not defined
print(f"doc unwrap elapse:{elapse}")
Traceback (most recent call last):
File "", line 1, in
NameError: name 'elapse' is not defined
cv2.imwrite("unwarped.png", unwrapped_img)
Traceback (most recent call last):
File "", line 1, in
NameError: name 'unwrapped_img' is not defined

[nissansz]

[Joker1212]

抱歉，写错示例了，from rapid_unwrap.inference import DocUnwrapper 这么引用

[nissansz]

doc_unwrapper = DocUnwrapper(img_path)
unwrapped_img, elapse = doc_unwrapper(img_path)
我把img_path改为 cv2的img，报了以下错误，怎么解决？

doc_unwrapper = DocUnwrapper(img)
File "L:\pythonPackaging\python-3.8.8-embed-amd64\lib\site-packages\rapid_unwrap\inference.py", line 35, in init
self.unwrap_model = UVDocPredictor(config)
File "L:\pythonPackaging\python-3.8.8-embed-amd64\lib\site-packages\rapid_unwrap\predictor.py", line 13, in init
self.session = OrtInferSession(config)
File "L:\pythonPackaging\python-3.8.8-embed-amd64\lib\site-packages\rapid_unwrap\utils\infer_engine.py", line 36, in init
EP_list = self._get_ep_list()
File "L:\pythonPackaging\python-3.8.8-embed-amd64\lib\site-packages\rapid_unwrap\utils\infer_engine.py", line 76, in _get_ep_list
self.use_cuda = self._check_cuda()
File "L:\pythonPackaging\python-3.8.8-embed-amd64\lib\site-packages\rapid_unwrap\utils\infer_engine.py", line 92, in _check_cuda
if not self.cfg_use_cuda:
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

[Joker1212]

同样是demo写错了，应该是创建的时候doc_unwrapper = DocUnwrapper()的，我马上改一下readme

[nissansz]

有没有modelscope的文本行检测模型onnx，rapidocr能用的？modelscope文本识别模型有没有训练教程？

[Joker1212]

可以去paddleOCR查看怎么训练微调

[nissansz]

不是paddle，是modelscope阿里的

[Joker1212]

应该只支持paddle的