Windows CPU 多次调用API检测图片，第二次就报Exception

[heepengpeng]

🔎 Search before asking

• I have searched the PaddleOCR Docs and found no similar bug report.
• I have searched the PaddleOCR Issues and found no similar bug report.
• I have searched the PaddleOCR Discussions and found no similar bug report.

🐛 Bug (问题描述)

正在处理第 2 张图片: images/weixin.png
Unknown exception
处理 images/weixin.png 出错: Unknown exception, 已跳过
所有图片处理完成！

🏃‍♂️ Environment (运行环境)

OS Windows11
Environment Python
Python 3.12.12
PaddleOCR 3.3.3
Install Wheel
RAM 32G
CPU Intel i7
CUDA None

🌰 Minimal Reproducible Example (最小可复现问题的Demo)

for idx, img_path in enumerate(image_list, 1):

    if not os.path.exists(img_path):
        print(f"文件不存在: {img_path}, 已跳过")
        continue

    print(f"正在处理第 {idx} 张图片: {img_path}")
    try:
        # 使用 OpenCV 读取图片并转换成 RGB
        img = cv2.imread(img_path)
        img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

        # 传给 OCR
        result = ocr.predict(img)
    except Exception as e:
        print(e)
        print(f"处理 {img_path} 出错: {e}, 已跳过")
        continue

[TingquanGao]

感谢反馈！请尝试以下排查步骤：

1. 直接传入文件路径而非 numpy array，确认问题是否仍存在：

result = ocr.predict(img_path)

2. 如果需要传入 numpy array，请注意 PaddleOCR 默认接收 BGR 格式（OpenCV 默认格式），无需手动转换为 RGB：

img = cv2.imread(img_path)
result = ocr.predict(img)  # 直接传 BGR，无需 cvtColor

3. 请提供完整的异常堆栈信息（加上如下代码后重新运行）：

try:
    result = ocr.predict(img)
except Exception as e:
    import traceback
    traceback.print_exc()

4. 请确认 paddleocr 和 paddlepaddle 的版本：

pip show paddleocr paddlepaddle paddlex

建议升级到最新版本（paddleocr >= 3.4.0）。

5. 如果上述方法都无法解决，能否提供用于测试的图片文件？方便我们在 Windows CPU 环境下复现和定位问题。

[benwu1980]

The same issue, paddle version 3.0.0

2026-03-17 16:57:14,310 [ERROR] OCR Unknown exception
Traceback (most recent call last):
File "D:\workfolder\project\czl_python\auto_calibration\arm_coord_handle.py", line 217, in ocr_arm_xyz_image
predict_result = ocr_model.predict(image_path)
File "D:\ProgramData\anaconda3\envs\auto_calibration\lib\site-packages\paddleocr_pipelines\ocr.py", line 191, in predict
return list(
File "D:\ProgramData\anaconda3\envs\auto_calibration\lib\site-packages\paddlex\inference\pipelines_parallel.py", line 129, in predict
yield from self._pipeline.predict(
File "D:\ProgramData\anaconda3\envs\auto_calibration\lib\site-packages\paddlex\inference\pipelines\ocr\pipeline.py", line 350, in predict
det_results = list(
File "D:\ProgramData\anaconda3\envs\auto_calibration\lib\site-packages\paddlex\inference\models\base\predictor\base_predictor.py", line 211, in call
yield from self.apply(input, **kwargs)
File "D:\ProgramData\anaconda3\envs\auto_calibration\lib\site-packages\paddlex\inference\models\base\predictor\base_predictor.py", line 267, in apply
prediction = self.process(batch_data, **kwargs)
File "D:\ProgramData\anaconda3\envs\auto_calibration\lib\site-packages\paddlex\inference\models\text_detection\predictor.py", line 105, in process
batch_preds = self.infer(x=x)
File "D:\ProgramData\anaconda3\envs\auto_calibration\lib\site-packages\paddlex\inference\models\common\static_infer.py", line 287, in call
pred = self.infer(x)
File "D:\ProgramData\anaconda3\envs\auto_calibration\lib\site-packages\paddlex\inference\models\common\static_infer.py", line 252, in call
self.predictor.run()