【PaddleOCR报错】(InvalidArgument) Broadcast dimension mismatch.

项目场景：

百度飞浆OCR做识别，训练了自己的模型，在推理预测过程中报错

都是使用的v4进行训练的，CPU，PaddlePaddle2.6.1

---

问题描述

导出模型后，在推理预测时候，遇到了如下报错，有无大佬遇到过，尝试了各种方法。无果。。。

(OCRTest) D:\BaiduNetdiskDownload\PaddleOCR-dygraph\PaddleOCR-dygraph>python tools/infer/predict_rec.py --image_dir="D:/BaiduNetdiskDownload/PaddleOCR-dygraph/PaddleOCR-dygraph/train_data/2024_08_26_16_37_47.png" --rec_model_dir="D:/BaiduNetdiskDownload/PaddleOCR-dygraph/PaddleOCR-dygraph/output/db_lqq_infer/Student"
E0910 10:39:18.437500 19012 analysis_config.cc:125] Please use PaddlePaddle with GPU version.
[2024/09/10 10:39:19] ppocr INFO: In PP-OCRv3, rec_image_shape parameter defaults to '3, 48, 320', if you are using recognition model with PP-OCRv2 or an older version, please set --rec_image_shape='3,32,320
[2024/09/10 10:39:19] ppocr INFO: Traceback (most recent call last):File "tools/infer/predict_rec.py", line 722, in mainrec_res, _ = text_recognizer(img_list)File "tools/infer/predict_rec.py", line 670, in __call__self.predictor.run()
ValueError: (InvalidArgument) Broadcast dimension mismatch. Operands could not be broadcast together with the shape of X = [1, 96, 3, 20] and the shape of Y = [1, 96, 4, 20]. Received [3] in X is not equal to [4] in Y at i:2.[Hint: Expected x_dims_array[i] == y_dims_array[i] || x_dims_array[i] <= 1 || y_dims_array[i] <= 1 == true, but received x_dims_array[i] == y_dims_array[i] || x_dims_array[i] <= 1 || y_dims_array[i] <= 1:0 != true:1.] (at ..\paddle/phi/kernels/funcs/common_shape.h:86)[operator < elementwise_add > error][2024/09/10 10:39:19] ppocr INFO: (InvalidArgument) Broadcast dimension mismatch. Operands could not be broadcast together with the shape of X = [1, 96, 3, 20] and the shape of Y = [1, 96, 4, 20]. Received [3] in X is not equal to [4] in Y at i:2.[Hint: Expected x_dims_array[i] == y_dims_array[i] || x_dims_array[i] <= 1 || y_dims_array[i] <= 1 == true, but received x_dims_array[i] == y_dims_array[i] || x_dims_array[i] <= 1 || y_dims_array[i] <= 1:0 != true:1.] (at ..\paddle/phi/kernels/funcs/common_shape.h:86)[operator < elementwise_add > error]

---

问题解决

把推理命令换成下面这个

python tools/infer/predict_system.py --image_dir="D:/BaiduNetdiskDownload/PaddleOCR-dygraph/PaddleOCR-dygraph/train_data/2024_08_26_16_37_47.png" --det_model_dir="D:/BaiduNetdiskDownload/PaddleOCR-dygraph/PaddleOCR-dygraph/output/db_lqq_infer/Student" --rec_model_dir="D:/BaiduNetdiskDownload/PaddleOCR-dygraph/PaddleOCR-dygraph/inference_model/ch_PP-OCRv4_rec_infer"