Getting inhomogeneous shape error while converting stn block to inference model "x = self.stn_convnet(x)"

[BMukhtar]

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：Windows 11, 6GB GPU RTX 3060, RAM 32GB, core-I712700H
• 版本号/Version： 2.7 Paddle：2.4.2 PaddleOCR：svtr config 问题相关组件/Related components：
• 运行指令/Command Code：python3 tools/export_model.py -c ./output/kz_synthtiger_rec_hbr/config.yml -o Global.pretrained_model=./output/kz_synthtiger_rec_hbr/best_accuracy Global.save_inference_dir=./inference/kz_synthtiger_rec_hbr/
• 完整报错/Complete Error Message：

Traceback (most recent call last):
  File "/home/bma/PaddleOCR_Mukhtar/tools/export_model.py", line 289, in <module>
    main()
  File "/home/bma/PaddleOCR_Mukhtar/tools/export_model.py", line 284, in main
    export_single_model(
  File "/home/bma/PaddleOCR_Mukhtar/tools/export_model.py", line 198, in export_single_model
    paddle.jit.save(model, save_path)
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/decorator.py", line 232, in fun
    return caller(func, *(extras + args), **kw)
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/wrapped_decorator.py", line 26, in __impl__
    return wrapped_func(*args, **kwargs)
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/jit.py", line 649, in wrapper
    func(layer, path, input_spec, **configs)
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/decorator.py", line 232, in fun
    return caller(func, *(extras + args), **kw)
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/wrapped_decorator.py", line 26, in __impl__
    return wrapped_func(*args, **kwargs)
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/base.py", line 67, in __impl__
    return func(*args, **kwargs)
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/jit.py", line 927, in save
    concrete_program = static_func.concrete_program_specify_input_spec(
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/dygraph_to_static/program_translator.py", line 577, in concrete_program_specify_input_spec
    concrete_program, _ = self.get_concrete_program(
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/dygraph_to_static/program_translator.py", line 485, in get_concrete_program
    concrete_program, partial_program_layer = self._program_cache[cache_key]
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/dygraph_to_static/program_translator.py", line 955, in __getitem__
    self._caches[item_id] = self._build_once(item)
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/dygraph_to_static/program_translator.py", line 939, in _build_once
    concrete_program = ConcreteProgram.from_func_spec(
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/decorator.py", line 232, in fun
    return caller(func, *(extras + args), **kw)
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/wrapped_decorator.py", line 26, in __impl__
    return wrapped_func(*args, **kwargs)
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/base.py", line 67, in __impl__
    return func(*args, **kwargs)
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/dygraph_to_static/program_translator.py", line 895, in from_func_spec
    error_data.raise_new_exception()
  File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/dygraph_to_static/error.py", line 350, in raise_new_exception
    six.exec_("raise new_exception from None")
  File "<string>", line 1, in <module>
ValueError: In transformed code:

    File "/home/bma/PaddleOCR_Mukhtar/ppocr/modeling/architectures/base_model.py", line 83, in forward
        if self.use_transform:
    File "/home/bma/PaddleOCR_Mukhtar/ppocr/modeling/architectures/base_model.py", line 84, in forward
        x = self.transform(x)
    File "/home/bma/PaddleOCR_Mukhtar/ppocr/modeling/transforms/stn.py", line 133, in forward
        stn_img_feat, ctrl_points = self.stn_head(stn_input)
    File "/home/bma/PaddleOCR_Mukhtar/ppocr/modeling/transforms/stn.py", line 105, in forward
        def forward(self, x):
            x = self.stn_convnet(x)
            ~~~~~~~~~~~~~~~~~~~~~~~ <--- HERE
            batch_size, _, h, w = x.shape
            x = paddle.reshape(x, shape=(batch_size, -1))

    File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/dygraph_to_static/convert_call_func.py", line 178, in convert_call
        if is_builtin(func) or is_unsupported(func):
    File "/home/bma/miniforge3/envs/paddle_env/lib/python3.9/site-packages/paddle/fluid/dygraph/dygraph_to_static/convert_call_func.py", line 91, in is_unsupported
        func_in_dict = func == v

    ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (11,) + inhomogeneous part.

我们提供了AceIssueSolver来帮助你解答问题，你是否想要它来解答(请填写yes/no)?/We provide AceIssueSolver to solve issues, do you want it? (Please write yes/no):

请尽量不要包含图片在问题中/Please try to not include the image in the issue.

My config file:

Global:
  use_gpu: True
  epoch_num: 50 #was 20
  log_smooth_window: 20
  print_batch_step: 10
  save_model_dir: ./output/kz_synthtiger_rec_hbr/
  save_epoch_step: 200
  # evaluation is run every 2000 iterations after the 0th iteration
  eval_batch_step: [0, 500]
  cal_metric_during_train: True
  pretrained_model:
  checkpoints:
  save_inference_dir:
  use_visualdl: False
  infer_img: doc/imgs_words_en/word_10.png
  # for data or label process
  character_dict_path: ./ppocr/utils/dict/kz_dict_v1.txt
  character_type: en
  max_text_length: 25 #was 25
  infer_mode: False
  use_space_char: True
  save_res_path: ./output/rec/predicts_svtr_tiny.txt


Optimizer:
  name: AdamW
  beta1: 0.9
  beta2: 0.99
  epsilon: 8.e-8
  weight_decay: 0.05
  no_weight_decay_name: norm pos_embed
  one_dim_param_no_weight_decay: true
  lr:
    name: Cosine
    learning_rate: 0.0005
    warmup_epoch: 1

Architecture:
  model_type: rec
  algorithm: SVTR
  Transform:
    name: STN_ON
    tps_inputsize: [32, 64]
    tps_outputsize: [32, 100]
    num_control_points: 20
    tps_margins: [0.05,0.05]
    stn_activation: none
  Backbone:
    name: SVTRNet
    img_size: [32, 100]
    out_char_num: 25 #was 25
    out_channels: 192
    patch_merging: 'Conv'
    embed_dim: [64, 128, 256]
    depth: [3, 6, 3]
    num_heads: [2, 4, 8]
    mixer: ['Local','Local','Local','Local','Local','Local','Global','Global','Global','Global','Global','Global']
    local_mixer: [[7, 11], [7, 11], [7, 11]]
    last_stage: True
    prenorm: false
  Neck:
    name: SequenceEncoder
    encoder_type: reshape
  Head:
    name: CTCHead

Loss:
  name: CTCLoss

PostProcess:
  name: CTCLabelDecode

Metric:
  name: RecMetric
  main_indicator: acc

Train:
  dataset:
    name: SimpleDataSet
    data_dir: ../synthtiger_kz/results/train_25/
    label_file_list:
      - ../synthtiger_kz/results/train_25/gt.txt
    transforms:
      - DecodeImage: # load image
          img_mode: BGR
          channel_first: False
      - CTCLabelEncode: # Class handling label
      #- NormalizeImage:
      #    scale: 1./255.
      #    mean: [0.7273, 0.7107, 0.7026]
      #    std: [0.3534, 0.3557, 0.3638]
      #    order: 'hwc'
      - SVTRRecResizeImg:
          image_shape: [3, 64, 256]
          padding: False
      - KeepKeys:
          keep_keys: ['image', 'label', 'length'] # dataloader will return list in this order
  loader:
    shuffle: True
    batch_size_per_card: 256
    drop_last: True
    num_workers: 2

Eval:
  dataset:
    name: SimpleDataSet
    data_dir: ../synthtiger_kz/results/test_25/
    label_file_list:
      - ../synthtiger_kz/results/test_25/gt.txt
    transforms:
      - DecodeImage: # load image
          img_mode: BGR
          channel_first: False
      - CTCLabelEncode: # Class handling label
      #- NormalizeImage:
      #    scale: 1./255.
      #    mean: [0.7273, 0.7107, 0.7026]
      #    std: [0.3534, 0.3557, 0.3638]
      #    order: 'hwc'
      - SVTRRecResizeImg:
          image_shape: [3, 64, 256]
          padding: False
      - KeepKeys:
          keep_keys: ['image', 'label', 'length'] # dataloader will return list in this order
  loader:
    shuffle: False
    drop_last: False
    batch_size_per_card: 256
    num_workers: 2