🔎 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)

之前自己训练的模型， 配置里是 pretrained_model: ./pretrain_models/ch_ppocr_server_v2.0_det_train/best_accuracy

首先用3.2版���导出检测推理模型会报错 （注，导出识别模型没报错）
(venv) admin-1@admin1-Rack-Server:/data/suchang/PaddleOCR$ python3 tools/export_model.py -c ../test/det_server_20220225/config.yml -o Global.pretrained_model=../test/det_server_20220225/best_accuracy Global.save_inference_dir=../my_models/dec_0225_infer
/data/suchang/test/venv/lib/python3.12/site-packages/paddle/utils/cpp_extension/extension_utils.py:718: UserWarning: No ccache found. Please be aware that recompiling all source files may be required. You can download and install ccache from: https://github.com/ccache/ccache/blob/master/doc/INSTALL.md
warnings.warn(warning_message)
Skipping import of the encryption module.
Traceback (most recent call last):
File "/data/suchang/PaddleOCR/tools/export_model.py", line 37, in
main()
File "/data/suchang/PaddleOCR/tools/export_model.py", line 33, in main
export(config)
File "/data/suchang/PaddleOCR/ppocr/utils/export_model.py", line 499, in export
model = build_model(config["Architecture"])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/suchang/PaddleOCR/ppocr/modeling/architectures/init.py", line 30, in build_model
arch = BaseModel(config)
^^^^^^^^^^^^^^^^^
File "/data/suchang/PaddleOCR/ppocr/modeling/architectures/base_model.py", line 55, in init
self.backbone = build_backbone(config["Backbone"], model_type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/suchang/PaddleOCR/ppocr/modeling/backbones/init.py", line 151, in build_backbone
module_class = eval(module_name)(**config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: ResNet.init() got an unexpected keyword argument 'disable_se'

删掉配置里的disable_se: true

(venv) admin-1@admin1-Rack-Server:/data/suchang/PaddleOCR$ python3 tools/export_model.py -c ../test/det_server_20220225/config.yml -o Global.pretrained_model=../test/det_server_20220225/best_accuracy Global.save_inference_dir=../my_models/dec_0225_infer
/data/suchang/test/venv/lib/python3.12/site-packages/paddle/utils/cpp_extension/extension_utils.py:718: UserWarning: No ccache found. Please be aware that recompiling all source files may be required. You can download and install ccache from: https://github.com/ccache/ccache/blob/master/doc/INSTALL.md
warnings.warn(warning_message)
Skipping import of the encryption module.
W1226 10:31:45.011487 155473 gpu_resources.cc:114] Please NOTE: device: 0, GPU Compute Capability: 12.0, Driver API Version: 13.0, Runtime API Version: 13.0
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_1._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_1._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_1._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_1._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_1._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_2._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_2._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_2._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_2._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_2._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_3._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_3._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_3._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_3._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.conv1_3._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.conv0._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.conv0._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.conv0._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.conv0._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.conv0._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.conv1._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.conv1._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.conv1._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.conv1._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.conv1._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.short._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.short._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.short._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.short._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_0.short._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_1.conv0._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_1.conv0._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_1.conv0._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_1.conv0._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_1.conv0._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_1.conv1._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_1.conv1._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_1.conv1._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_1.conv1._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_0_1.conv1._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.conv0._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.conv0._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.conv0._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.conv0._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.conv0._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.conv1._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.conv1._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.conv1._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.conv1._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.conv1._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.short._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.short._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.short._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.short._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_0.short._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_1.conv0._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_1.conv0._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_1.conv0._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_1.conv0._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_1.conv0._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_1.conv1._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_1.conv1._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_1.conv1._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_1.conv1._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_1_1.conv1._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.conv0._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.conv0._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.conv0._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.conv0._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.conv0._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.conv1._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.conv1._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.conv1._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.conv1._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.conv1._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.short._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.short._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.short._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.short._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_0.short._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_1.conv0._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_1.conv0._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_1.conv0._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_1.conv0._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_1.conv0._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_1.conv1._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_1.conv1._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_1.conv1._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_1.conv1._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_2_1.conv1._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.conv0._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.conv0._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.conv0._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.conv0._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.conv0._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.conv1._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.conv1._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.conv1._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.conv1._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.conv1._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.short._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.short._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.short._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.short._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_0.short._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_1.conv0._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_1.conv0._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_1.conv0._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_1.conv0._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_1.conv0._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_1.conv1._conv.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_1.conv1._batch_norm.weight not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_1.conv1._batch_norm.bias not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_1.conv1._batch_norm._mean not in model
[2025/12/26 10:31:45] ppocr WARNING: The pretrained params backbone.bb_3_1.conv1._batch_norm._variance not in model
[2025/12/26 10:31:45] ppocr INFO: load pretrain successful from ../test/det_server_20220225/best_accuracy
[2025/12/26 10:31:45] ppocr INFO: Export inference config file to ../my_models/dec_0225_infer/inference.yml
Skipping import of the encryption module
W1226 10:31:45.251539 155473 eager_utils.cc:3441] Paddle static graph(PIR) not support input out tensor for now!!!!!
[2025/12/26 10:31:45] ppocr INFO: inference model is saved to ../my_models/dec_0225_infer/inference

执行以下脚本
from paddleocr import PaddleOCR

ocr = PaddleOCR(
text_detection_model_name="my_det",
text_detection_model_dir="/data/suchang/my_models/det_0225_infer",
text_recognition_model_name="my_rec",
text_recognition_model_dir="/data/suchang/my_models/rec_0225_infer",
use_doc_orientation_classify=False,
use_doc_unwarping=False,
use_textline_orientation=False,
#ocr_version="PP-OCRv2",
)
...

报错
(venv) admin-1@admin1-Rack-Server:/data/suchang/test$ python test.py
Checking connectivity to the model hosters, this may take a while. To bypass this check, set DISABLE_MODEL_SOURCE_CHECK to True.
/data/suchang/test/venv/lib/python3.12/site-packages/paddle/utils/cpp_extension/extension_utils.py:718: UserWarning: No ccache found. Please be aware that recompiling all source files may be required. You can download and install ccache from: https://github.com/ccache/ccache/blob/master/doc/INSTALL.md
warnings.warn(warning_message)
Creating model: ('my_det', '/data/suchang/my_models/det_0225_infer')
Traceback (most recent call last):
File "/data/suchang/test/test.py", line 11, in
ocr = PaddleOCR(
^^^^^^^^^^
File "/data/suchang/test/venv/lib/python3.12/site-packages/paddleocr/_pipelines/ocr.py", line 163, in init
super().init(**base_params)
File "/data/suchang/test/venv/lib/python3.12/site-packages/paddleocr/_pipelines/base.py", line 67, in init
self.paddlex_pipeline = self._create_paddlex_pipeline()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/suchang/test/venv/lib/python3.12/site-packages/paddleocr/_pipelines/base.py", line 105, in _create_paddlex_pipeline
return create_pipeline(config=self._merged_paddlex_config, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/suchang/test/venv/lib/python3.12/site-packages/paddlex/inference/pipelines/init.py", line 167, in create_pipeline
pipeline = BasePipeline.get(pipeline_name)(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/suchang/test/venv/lib/python3.12/site-packages/paddlex/utils/deps.py", line 206, in _wrapper
return old_init_func(self, *args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/suchang/test/venv/lib/python3.12/site-packages/paddlex/inference/pipelines/_parallel.py", line 103, in init
self._pipeline = self._create_internal_pipeline(config, self.device)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/data/suchang/test/venv/lib/python3.12/site-packages/paddlex/inference/pipelines/_parallel.py", line 158, in _create_internal_pipeline
return self._pipeline_cls(
^^^^^^^^^^^^^^^^^^^
File "/data/suchang/test/venv/lib/python3.12/site-packages/paddlex/inference/pipelines/ocr/pipeline.py", line 117, in init
self.text_det_model = self.create_model(
^^^^^^^^^^^^^^^^^^
File "/data/suchang/test/venv/lib/python3.12/site-packages/paddlex/inference/pipelines/base.py", line 106, in create_model
model = create_predictor(
^^^^^^^^^^^^^^^^^
File "/data/suchang/test/venv/lib/python3.12/site-packages/paddlex/inference/models/init.py", line 79, in create_predictor
model_name == config["Global"]["model_name"]
~~~~~~^^^^^^^^^^
KeyError: 'Global'

注： 用官方默认模型是可以识别的

🏃‍♂️ Environment (运行环境)

Ubuntu 24
python 3.12
paddleocr 3.3.2
paddlepaddle-gpu 3.2.2
paddlex 3.3.12

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

det模型的配置

(venv) admin-1@admin1-Rack-Server:/data/suchang/test$ cat det_server_20220225/config.yml_bak_before_remove_disable_se
Global:
debug: false
use_gpu: true
epoch_num: 400
log_smooth_window: 20
print_batch_step: 2
save_model_dir: ./output/det_server_20220225/
save_epoch_step: 10
eval_batch_step:

• 0
• 500
  cal_metric_during_train: false
  pretrained_model: ./pretrain_models/ch_ppocr_server_v2.0_det_train/best_accuracy
  checkpoints: null
  save_inference_dir: null
  use_visualdl: false
  infer_img: doc/imgs_en/img_10.jpg
  save_res_path: ./output/det_db/predicts_db.txt
  Architecture:
  model_type: det
  algorithm: DB
  Transform: null
  Backbone:
  name: ResNet
  layers: 18
  disable_se: true
  Neck:
  name: DBFPN
  out_channels: 256
  Head:
  name: DBHead
  k: 50
  Loss:
  name: DBLoss
  balance_loss: true
  main_loss_type: DiceLoss
  alpha: 5
  beta: 10
  ohem_ratio: 3
  Optimizer:
  name: Adam
  beta1: 0.9
  beta2: 0.999
  lr:
  name: Cosine
  learning_rate: 0.0003
  warmup_epoch: 2
  regularizer:
  name: L2
  factor: 0
  PostProcess:
  name: DBPostProcess
  thresh: 0.3
  box_thresh: 0.6
  max_candidates: 1000
  unclip_ratio: 1.5
  Metric:
  name: DetMetric
  main_indicator: hmean
  Train:
  dataset:
  name: SimpleDataSet
  data_dir: ./train_data/text_localization/train_full_images_with_x/
  label_file_list:
  • ./train_data/text_localization/train_x_labels.json
    ratio_list:
  • 1.0
    transforms:
  • DecodeImage:
    img_mode: BGR
    channel_first: false
  • DetLabelEncode: null
  • IaaAugment:
    augmenter_args:
    • type: Fliplr
      args:
      p: 0.5
    • type: Affine
      args:
      rotate:
      • -10
      • 10
    • type: Resize
      args:
      size:
      • 0.5
      • 3
  • EastRandomCropData:
    size:
    • 960
    • 960
      max_tries: 50
      keep_ratio: true
  • MakeBorderMap:
    shrink_ratio: 0.4
    thresh_min: 0.3
    thresh_max: 0.7
  • MakeShrinkMap:
    shrink_ratio: 0.4
    min_text_size: 8
  • NormalizeImage:
    scale: 1./255.
    mean:
    • 0.485
    • 0.456
    • 0.406
      std:
    • 0.229
    • 0.224
    • 0.225
      order: hwc
  • ToCHWImage: null
  • KeepKeys:
    keep_keys:
    • image
    • threshold_map
    • threshold_mask
    • shrink_map
    • shrink_mask
      loader:
      shuffle: true
      drop_last: false
      batch_size_per_card: 16
      num_workers: 1
      Eval:
      dataset:
      name: SimpleDataSet
      data_dir: ./train_data/text_localization/test_full_images_with_x/
      label_file_list:
  • ./train_data/text_localization/test_x_labels.json
    transforms:
  • DecodeImage:
    img_mode: BGR
    channel_first: false
  • DetLabelEncode: null
  • DetResizeForTest: null
  • NormalizeImage:
    scale: 1./255.
    mean:
    • 0.485
    • 0.456
    • 0.406
      std:
    • 0.229
    • 0.224
    • 0.225
      order: hwc
  • ToCHWImage: null
  • KeepKeys:
    keep_keys:
    • image
    • shape
    • polys
    • ignore_tags
      loader:
      shuffle: false
      drop_last: false
      batch_size_per_card: 1
      num_workers: 1
      profiler_options: null

rec模型的配置
(venv) admin-1@admin1-Rack-Server:/data/suchang/test$ cat rec_server_20220225/config.yml
Global:
debug: false
use_gpu: true
epoch_num: 300
log_smooth_window: 20
print_batch_step: 10
save_model_dir: ./output/rec_server_20220225
save_epoch_step: 3
eval_batch_step:

• 0
• 2000
  cal_metric_during_train: true
  pretrained_model: ./pretrain_models/ch_ppocr_server_v2.0_rec_train/best_accuracy
  checkpoints: null
  save_inference_dir: null
  use_visualdl: false
  infer_img: doc/imgs_words/ch/word_1.jpg
  character_dict_path: ppocr/utils/ppocr_keys_v1.txt
  max_text_length: 25
  infer_mode: false
  use_space_char: false
  save_res_path: ./output/rec/predicts_chinese_common_v2.0.txt
  Optimizer:
  name: Adam
  beta1: 0.9
  beta2: 0.999
  lr:
  name: Cosine
  learning_rate: 0.001
  warmup_epoch: 5
  regularizer:
  name: L2
  factor: 4.0e-05
  Architecture:
  model_type: rec
  algorithm: CRNN
  Transform: null
  Backbone:
  name: ResNet
  layers: 34
  Neck:
  name: SequenceEncoder
  encoder_type: rnn
  hidden_size: 256
  Head:
  name: CTCHead
  fc_decay: 4.0e-05
  Loss:
  name: CTCLoss
  PostProcess:
  name: CTCLabelDecode
  Metric:
  name: RecMetric
  main_indicator: acc
  Train:
  dataset:
  name: SimpleDataSet
  data_dir: ./train_data/rec_all/
  label_file_list:
  • train_data/rec_all/train_label_changguisai.txt
  • train_data/rec_all/train_label_chuangxinsai.txt
  • train_data/rec_all/train_label_hecheng.txt
  • train_data/rec_all/train_list.txt
    ratio_list:
  • 1.0
  • 1.0
  • 1.0
  • 1.0
    transforms:
  • DecodeImage:
    img_mode: BGR
    channel_first: false
  • RecAug: null
  • CTCLabelEncode: null
  • RecResizeImg:
    image_shape:
    • 3
    • 32
    • 320
  • KeepKeys:
    keep_keys:
    • image
    • label
    • length
      loader:
      shuffle: true
      batch_size_per_card: 256
      drop_last: true
      num_workers: 1
      Eval:
      dataset:
      name: SimpleDataSet
      data_dir: ./train_data/rec_all/
      label_file_list:
  • train_data/rec_all/eval_label_changguisai.txt
  • train_data/rec_all/eval_label_chuangxinsai.txt
  • train_data/rec_all/eval_label_hecheng.txt
  • train_data/rec_all/val_list.txt
    ratio_list:
  • 1.0
  • 1.0
  • 1.0
  • 1.0
    transforms:
  • DecodeImage:
    img_mode: BGR
    channel_first: false
  • CTCLabelEncode: null
  • RecResizeImg:
    image_shape:
    • 3
    • 32
    • 320
  • KeepKeys:
    keep_keys:
    • image
    • label
    • length
      loader:
      shuffle: false
      drop_last: false
      batch_size_per_card: 256
      num_workers: 1
      profiler_options: null