list index out of range when using PP-OCRv4

[m0hamedAit]

请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem

• 系统环境/System Environment：Google Colab

• 版本号/Version：Paddle： PaddleOCR： 问题相关组件/Related components：Paddle: 2.5.1 PaddleOCR: 2.7.0.3

• 运行指令/Command Code：

  paddleocr = PaddleOCR(use_angle_cls=True, det_model_dir='/content/drive/MyDrive/paddleOCR/ch_PP-OCRv4_det_infer', rec_model_dir='/content/drive/MyDrive/paddleOCR/ch_PP-OCRv4_rec_infer', rec_char_dict_path='/content/drive/MyDrive/paddleOCR/en_dict.txt')
  paddleocr.ocr("/content/drive/MyDrive/image.png")

  Works fine when using PP-OCRv3 instead of PP-OCRv4

• 完整报错/Complete Error Message：

  
  [2023/09/18 09:41:50] ppocr DEBUG: dt_boxes num : 47, elapsed : 1.3884902000427246
  [2023/09/18 09:41:51] ppocr DEBUG: cls num  : 47, elapsed : 0.6255166530609131
  ---------------------------------------------------------------------------
  IndexError                                Traceback (most recent call last)
  [<ipython-input-11-fb46d0b75bf0>](https://localhost:8080/#) in <cell line: 1>()
  ----> 1 text2 = ocr2.ocr("/content/drive/MyDrive/image.png")

5 frames /usr/local/lib/python3.10/dist-packages/paddleocr/paddleocr.py in ocr(self, img, det, rec, cls, bin, inv, alpha_color) 659 for idx, img in enumerate(imgs): 660 img = preprocess_image(img) --> 661 dt_boxes, recres, = self.call(img, cls) 662 if not dt_boxes and not rec_res: 663 ocr_res.append(None)

/usr/local/lib/python3.10/dist-packages/paddleocr/tools/infer/predict_system.py in call(self, img, cls) 103 len(img_crop_list), elapse)) 104 --> 105 rec_res, elapse = self.text_recognizer(img_crop_list) 106 time_dict['rec'] = elapse 107 logger.debug("rec_res num : {}, elapsed : {}".format(

/usr/local/lib/python3.10/dist-packages/paddleocr/tools/infer/predict_rec.py in call(self, img_list) 626 preds = outputs[0] 627 self.predictor.try_shrink_memory() --> 628 rec_result = self.postprocess_op(preds) 629 for rno in range(len(rec_result)): 630 rec_res[indices[beg_img_no + rno]] = rec_result[rno]

/usr/local/lib/python3.10/dist-packages/paddleocr/ppocr/postprocess/rec_postprocess.py in call(self, preds, label, *args, *kwargs) 119 def call(self, preds, label=None, args, **kwargs): 120 if isinstance(preds, tuple) or isinstance(preds, list): --> 121 preds = preds[-1] 122 if isinstance(preds, paddle.Tensor): 123 preds = preds.numpy()

/usr/local/lib/python3.10/dist-packages/paddleocr/ppocr/postprocess/rec_postprocess.py in decode(self, text_index, text_prob, is_remove_duplicate) 81 selection &= text_index[batch_idx] != ignored_token 82 83 char_list = [ ---> 84 self.character[text_id] 85 for text_id in text_index[batch_idx][selection] 86 ]

/usr/local/lib/python3.10/dist-packages/paddleocr/ppocr/postprocess/rec_postprocess.py in (.0) 82 83 char_list = [ ---> 84 self.character[text_id] 85 for text_id in text_index[batch_idx][selection] 86 ]

IndexError: list index out of range

[zzc81824130]

请问最后怎么解决的？

[m0hamedAit]

请问最后怎么解决的？

If I remember correctly, I reinstalled PaddleOCR using the following command: pip install paddleocr==2.7.0.3 --upgrade PyMuPDF==1.20.2. In my case, I had to install swig, pcre, and bison>3.5 before running this command.