VQA SER模型推理

[StarSapph1re]

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：windows11 numpy 1.23.2 CUDA 11.6
  paddle-bfloat 0.1.7
  paddle2onnx 1.0.6
  paddlefsl 1.1.0
  paddlenlp 2.5.2
  paddleocr 2.7.0.2
  paddlepaddle-gpu 2.5.1.post116
  问题相关组件/Related components：postprocess\vqa_token_ser_layoutlm_postprocess.py
• 运行指令/Command Code： python tools/infer_vqa_token_ser.py -c configs/vqa/ser/layoutxlm.yml -o Architecture.Backbone.checkpoints=ser_LayoutXLM_xfun_zh/ Global.infer_img=doc/vqa/input/zh_val_42.jpg
• 完整报错/Complete Error Message：
• [2023/09/09 20:33:59] ppocr INFO: train with paddle 2.5.1 and device Place(gpu:0)
  W0909 20:34:02.106488 26752 gpu_resources.cc:119] Please NOTE: device: 0, GPU Compute Capability: 7.5, Driver API Version: 12.0, Runtime API Version: 11.6
  W0909 20:34:02.111487 26752 gpu_resources.cc:149] device: 0, cuDNN Version: 8.4.
  [2023/09/09 20:34:06] ppocr INFO: resume from ser_LayoutXLM_xfun_zh/
  [2023-09-09 20:34:08,903] [ INFO] - Already cached C:\Users\86152.paddlenlp\models\layoutxlm-base-uncased\sentencepiece.bpe.model
  [2023-09-09 20:34:09,402] [ INFO] - tokenizer config file saved in C:\Users\86152.paddlenlp\models\layoutxlm-base-uncased\tokenizer_config.json
  [2023-09-09 20:34:09,403] [ INFO] - Special tokens file saved in C:\Users\86152.paddlenlp\models\layoutxlm-base-uncased\special_tokens_map.json
  [2023/09/09 20:34:09] ppocr INFO: process: [0/1], save result to ./output/ser\zh_val_42_ser.jpg
  Corrupt JPEG data: premature end of data segment
  Traceback (most recent call last):
  File "E:\home\aistudio\PaddleOCR\tools\infer_vqa_token_ser.py", line 132, in
  result, _ = ser_engine(img_path)
  File "E:\home\aistudio\PaddleOCR\tools\infer_vqa_token_ser.py", line 105, in call
  post_result = self.post_process_class(
  File "E:\home\aistudio\PaddleOCR\ppocr\postprocess\vqa_token_ser_layoutlm_postprocess.py", line 50, in call
  return self._infer(preds, **kwargs)
  File "E:\home\aistudio\PaddleOCR\ppocr\postprocess\vqa_token_ser_layoutlm_postprocess.py", line 71, in _infer
  pred = np.argmax(pred, axis=1)
  File "<array_function internals>", line 180, in argmax
  File "S:\Anaconda\envs\chat\lib\site-packages\numpy\core\fromnumeric.py", line 1216, in argmax
  return _wrapfunc(a, 'argmax', axis=axis, out=out, **kwds)
  File "S:\Anaconda\envs\chat\lib\site-packages\numpy\core\fromnumeric.py", line 57, in _wrapfunc
  return bound(*args, **kwds)
  numpy.AxisError: axis 1 is out of bounds for array of dimension 1

在一些类似的推理和评估过程有也遇到过这一问题，无法解决，模型和推理图片均是github原数据，枚举了可用的numpy版本也无法正常推理