Paddle_OCR Training - RecursionError: maximum recursion depth exceeded while calling a Python object #13416
Paddle_OCR Training - RecursionError: maximum recursion depth exceeded while calling a Python object
#13416
-
|
I was trying to train a paddle ocr rec mode. my train.txt file contains labeled image dataset using the tool PPOCR_label I got this error below - so how to fix this error. ? help? |
Beta Was this translation helpful? Give feedback.
Answered by
kishore-work-hard
Dec 11, 2024
Replies: 2 comments 4 replies
-
|
the dataset you used is for det model, not for rec model. |
Beta Was this translation helpful? Give feedback.
4 replies
-
SO SORRY! |
Beta Was this translation helpful? Give feedback.
-
|
Hi. I have same proplem.. Did you solve this? @kishore-work-hard |
Beta Was this translation helpful? Give feedback.
-
The datasets used for both recognition and detection models are different. For recognition models u should crop the lp part and then label it. When I used a full sized image with cars and lp I got the recursion error (this dataset is used for detection training not recognition). for the detection Label.txt, I can use the file generated by PPOCRLabel directly. But for recognition I should crop the image to fit just the text then the Labels.txt should be: filepath tabspace text Do detection training first. After that do recognition training. Don't waste time doing recognition training first |
Beta Was this translation helpful? Give feedback.
Answer selected by
GreatV
-
|
@kishore-work-hard wow! you saved my time bro! thank you so much! |
Beta Was this translation helpful? Give feedback.
-
|
Hi I'm still having this issue I'm using this config file to train my rec model (configs\rec\PP-OCRv4) Global:
debug: false
use_gpu: true
epoch_num: 500
log_smooth_window: 20
print_batch_step: 10
save_model_dir: ./output/rec_ppocr_v4
save_epoch_step: 10
eval_batch_step:
- 0
- 2000
cal_metric_during_train: true
pretrained_model: null
checkpoints: null
save_inference_dir: null
use_visualdl: false
infer_img: doc/imgs_en/A.png
character_dict_path: ppocr/utils/en_dict.txt
max_text_length: 1
infer_mode: false
use_space_char: false
distributed: true
save_res_path: ./output/rec/predicts_ppocrv3.txt
Optimizer:
name: Adam
beta1: 0.9
beta2: 0.999
lr:
name: Cosine
learning_rate: 0.0005
warmup_epoch: 5
regularizer:
name: L2
factor: 3.0e-05
Architecture:
model_type: rec
algorithm: SVTR_LCNet
Transform: null
Backbone:
name: PPLCNetV3
scale: 0.95
Head:
name: MultiHead
head_list:
- CTCHead:
Neck:
name: svtr
dims: 120
depth: 2
hidden_dims: 120
kernel_size:
- 1
- 3
use_guide: true
Head:
fc_decay: 1.0e-05
- NRTRHead:
nrtr_dim: 384
max_text_length: 25
Loss:
name: MultiLoss
loss_config_list:
- CTCLoss: null
- NRTRLoss: null
PostProcess:
name: CTCLabelDecode
Metric:
name: RecMetric
main_indicator: acc
ignore_space: false
Train:
dataset:
name: MultiScaleDataSet
# name: SimpleDataSet
ds_width: false
# data_dir: ./train_data/
data_dir: ./train_data/kaggle_rec/
ext_op_transform_idx: 1
label_file_list:
# - ./train_data/train_list.txt
- ./train_data/kaggle_rec/rec_train_symbol.txt
transforms:
- DecodeImage:
img_mode: BGR
channel_first: false
- RecConAug:
prob: 0.5
ext_data_num: 2
image_shape:
- 48
- 320
- 3
max_text_length: 25
- RecAug: null
- MultiLabelEncode:
gtc_encode: NRTRLabelEncode
- KeepKeys:
keep_keys:
- image
- label_ctc
- label_gtc
- length
- valid_ratio
sampler:
name: MultiScaleSampler
scales:
- - 320
- 32
- - 320
- 48
- - 320
- 64
first_bs: 96
fix_bs: false
divided_factor:
- 8
- 16
is_training: true
loader:
shuffle: true
# batch_size_per_card: 96
# batch_size_per_card: 124
batch_size_per_card: 62
drop_last: true
num_workers: 4
Eval:
dataset:
name: SimpleDataSet
# name: SimpleDataSet
# data_dir: ./train_data
data_dir: ./train_data/kaggle_rec/
label_file_list:
# - ./train_data/val_list.txt
- ./train_data/kaggle_rec/rec_test_symbol.txt
transforms:
- DecodeImage:
img_mode: BGR
channel_first: false
- MultiLabelEncode:
gtc_encode: NRTRLabelEncode
- RecResizeImg:
image_shape:
- 3
- 48
- 320
- KeepKeys:
keep_keys:
- image
- label_ctc
- label_gtc
- length
- valid_ratio
loader:
shuffle: false
drop_last: false
batch_size_per_card: 62
# batch_size_per_card: 62
num_workers: 4
profiler_options: nullThis is my training set txt file Result [2025/02/21 16:43:00] ppocr WARNING: Skipping import of the encryption module.
[2025/02/21 16:43:00] ppocr INFO: Architecture :
[2025/02/21 16:43:00] ppocr INFO: Backbone :
[2025/02/21 16:43:00] ppocr INFO: name : PPLCNetV3
[2025/02/21 16:43:00] ppocr INFO: scale : 0.95
[2025/02/21 16:43:00] ppocr INFO: Head :
[2025/02/21 16:43:00] ppocr INFO: head_list :
[2025/02/21 16:43:00] ppocr INFO: CTCHead :
[2025/02/21 16:43:00] ppocr INFO: Head :
[2025/02/21 16:43:00] ppocr INFO: fc_decay : 1e-05
[2025/02/21 16:43:00] ppocr INFO: Neck :
[2025/02/21 16:43:00] ppocr INFO: depth : 2
[2025/02/21 16:43:00] ppocr INFO: dims : 120
[2025/02/21 16:43:00] ppocr INFO: hidden_dims : 120
[2025/02/21 16:43:00] ppocr INFO: kernel_size : [1, 3]
[2025/02/21 16:43:00] ppocr INFO: name : svtr
[2025/02/21 16:43:00] ppocr INFO: use_guide : True
[2025/02/21 16:43:00] ppocr INFO: NRTRHead :
[2025/02/21 16:43:00] ppocr INFO: max_text_length : 25
[2025/02/21 16:43:00] ppocr INFO: nrtr_dim : 384
[2025/02/21 16:43:00] ppocr INFO: name : MultiHead
[2025/02/21 16:43:00] ppocr INFO: Transform : None
[2025/02/21 16:43:00] ppocr INFO: algorithm : SVTR_LCNet
[2025/02/21 16:43:00] ppocr INFO: model_type : rec
[2025/02/21 16:43:00] ppocr INFO: Eval :
[2025/02/21 16:43:00] ppocr INFO: dataset :
[2025/02/21 16:43:00] ppocr INFO: data_dir : ./train_data/kaggle_rec/
[2025/02/21 16:43:00] ppocr INFO: label_file_list : ['./train_data/kaggle_rec/rec_test_symbol.txt']
[2025/02/21 16:43:00] ppocr INFO: name : SimpleDataSet
[2025/02/21 16:43:00] ppocr INFO: transforms :
[2025/02/21 16:43:00] ppocr INFO: DecodeImage :
[2025/02/21 16:43:00] ppocr INFO: channel_first : False
[2025/02/21 16:43:00] ppocr INFO: img_mode : BGR
[2025/02/21 16:43:00] ppocr INFO: MultiLabelEncode :
[2025/02/21 16:43:00] ppocr INFO: gtc_encode : NRTRLabelEncode
[2025/02/21 16:43:00] ppocr INFO: RecResizeImg :
[2025/02/21 16:43:00] ppocr INFO: image_shape : [3, 48, 320]
[2025/02/21 16:43:00] ppocr INFO: KeepKeys :
[2025/02/21 16:43:00] ppocr INFO: keep_keys : ['image', 'label_ctc', 'label_gtc', 'length', 'valid_ratio']
[2025/02/21 16:43:00] ppocr INFO: loader :
[2025/02/21 16:43:00] ppocr INFO: batch_size_per_card : 62
[2025/02/21 16:43:00] ppocr INFO: drop_last : False
[2025/02/21 16:43:00] ppocr INFO: num_workers : 4
[2025/02/21 16:43:00] ppocr INFO: shuffle : False
[2025/02/21 16:43:00] ppocr INFO: Global :
[2025/02/21 16:43:00] ppocr INFO: cal_metric_during_train : True
[2025/02/21 16:43:00] ppocr INFO: character_dict_path : ppocr/utils/en_dict.txt
[2025/02/21 16:43:00] ppocr INFO: checkpoints : ./pretrain_models/en_PP-OCRv4_rec_train/best_accuracy
[2025/02/21 16:43:00] ppocr INFO: debug : False
[2025/02/21 16:43:00] ppocr INFO: distributed : False
[2025/02/21 16:43:00] ppocr INFO: epoch_num : 500
[2025/02/21 16:43:00] ppocr INFO: eval_batch_step : [0, 2000]
[2025/02/21 16:43:00] ppocr INFO: infer_img : doc/imgs_en/A.png
[2025/02/21 16:43:00] ppocr INFO: infer_mode : False
[2025/02/21 16:43:00] ppocr INFO: log_smooth_window : 20
[2025/02/21 16:43:00] ppocr INFO: max_text_length : 1
[2025/02/21 16:43:00] ppocr INFO: pretrained_model : None
[2025/02/21 16:43:00] ppocr INFO: print_batch_step : 10
[2025/02/21 16:43:00] ppocr INFO: save_epoch_step : 10
[2025/02/21 16:43:00] ppocr INFO: save_inference_dir : None
[2025/02/21 16:43:00] ppocr INFO: save_model_dir : ./output/rec_ppocr_v4
[2025/02/21 16:43:00] ppocr INFO: save_res_path : ./output/rec/predicts_ppocrv3.txt
[2025/02/21 16:43:00] ppocr INFO: use_gpu : True
[2025/02/21 16:43:00] ppocr INFO: use_space_char : False
[2025/02/21 16:43:00] ppocr INFO: use_visualdl : False
[2025/02/21 16:43:00] ppocr INFO: Loss :
[2025/02/21 16:43:00] ppocr INFO: loss_config_list :
[2025/02/21 16:43:00] ppocr INFO: CTCLoss : None
[2025/02/21 16:43:00] ppocr INFO: NRTRLoss : None
[2025/02/21 16:43:00] ppocr INFO: name : MultiLoss
[2025/02/21 16:43:00] ppocr INFO: Metric :
[2025/02/21 16:43:00] ppocr INFO: ignore_space : False
[2025/02/21 16:43:00] ppocr INFO: main_indicator : acc
[2025/02/21 16:43:00] ppocr INFO: name : RecMetric
[2025/02/21 16:43:00] ppocr INFO: Optimizer :
[2025/02/21 16:43:00] ppocr INFO: beta1 : 0.9
[2025/02/21 16:43:00] ppocr INFO: beta2 : 0.999
[2025/02/21 16:43:00] ppocr INFO: lr :
[2025/02/21 16:43:00] ppocr INFO: learning_rate : 0.0005
[2025/02/21 16:43:00] ppocr INFO: name : Cosine
[2025/02/21 16:43:00] ppocr INFO: warmup_epoch : 5
[2025/02/21 16:43:00] ppocr INFO: name : Adam
[2025/02/21 16:43:00] ppocr INFO: regularizer :
[2025/02/21 16:43:00] ppocr INFO: factor : 3e-05
[2025/02/21 16:43:00] ppocr INFO: name : L2
[2025/02/21 16:43:00] ppocr INFO: PostProcess :
[2025/02/21 16:43:00] ppocr INFO: name : CTCLabelDecode
[2025/02/21 16:43:00] ppocr INFO: Train :
[2025/02/21 16:43:00] ppocr INFO: dataset :
[2025/02/21 16:43:00] ppocr INFO: data_dir : ./train_data/kaggle_rec/
[2025/02/21 16:43:00] ppocr INFO: ds_width : False
[2025/02/21 16:43:00] ppocr INFO: ext_op_transform_idx : 1
[2025/02/21 16:43:00] ppocr INFO: label_file_list : ['./train_data/kaggle_rec/rec_train_symbol.txt']
[2025/02/21 16:43:00] ppocr INFO: name : MultiScaleDataSet
[2025/02/21 16:43:00] ppocr INFO: transforms :
[2025/02/21 16:43:00] ppocr INFO: DecodeImage :
[2025/02/21 16:43:00] ppocr INFO: channel_first : False
[2025/02/21 16:43:00] ppocr INFO: img_mode : BGR
[2025/02/21 16:43:00] ppocr INFO: RecConAug :
[2025/02/21 16:43:00] ppocr INFO: ext_data_num : 2
[2025/02/21 16:43:00] ppocr INFO: image_shape : [48, 320, 3]
[2025/02/21 16:43:00] ppocr INFO: max_text_length : 25
[2025/02/21 16:43:00] ppocr INFO: prob : 0.5
[2025/02/21 16:43:00] ppocr INFO: RecAug : None
[2025/02/21 16:43:00] ppocr INFO: MultiLabelEncode :
[2025/02/21 16:43:00] ppocr INFO: gtc_encode : NRTRLabelEncode
[2025/02/21 16:43:00] ppocr INFO: KeepKeys :
[2025/02/21 16:43:00] ppocr INFO: keep_keys : ['image', 'label_ctc', 'label_gtc', 'length', 'valid_ratio']
[2025/02/21 16:43:00] ppocr INFO: loader :
[2025/02/21 16:43:00] ppocr INFO: batch_size_per_card : 62
[2025/02/21 16:43:00] ppocr INFO: drop_last : True
[2025/02/21 16:43:00] ppocr INFO: num_workers : 4
[2025/02/21 16:43:00] ppocr INFO: shuffle : True
[2025/02/21 16:43:00] ppocr INFO: sampler :
[2025/02/21 16:43:00] ppocr INFO: divided_factor : [8, 16]
[2025/02/21 16:43:00] ppocr INFO: first_bs : 96
[2025/02/21 16:43:00] ppocr INFO: fix_bs : False
[2025/02/21 16:43:00] ppocr INFO: is_training : True
[2025/02/21 16:43:00] ppocr INFO: name : MultiScaleSampler
[2025/02/21 16:43:00] ppocr INFO: scales : [[320, 32], [320, 48], [320, 64]]
[2025/02/21 16:43:00] ppocr INFO: profiler_options : None
[2025/02/21 16:43:00] ppocr INFO: train with paddle 2.6.1 and device Place(gpu:0)
[2025/02/21 16:43:00] ppocr INFO: Initialize indexs of datasets:['./train_data/kaggle_rec/rec_train_symbol.txt']
[2025/02/21 16:43:00] ppocr INFO: Initialize indexs of datasets:['./train_data/kaggle_rec/rec_test_symbol.txt']
W0221 16:43:00.761966 47820 gpu_resources.cc:119] Please NOTE: device: 0, GPU Compute Capability: 8.9, Driver API Version: 12.6, Runtime API Version: 11.7
W0221 16:43:00.767946 47820 gpu_resources.cc:164] device: 0, cuDNN Version: 8.9.
[2025/02/21 16:43:02] ppocr INFO: train dataloader has 30 iters
[2025/02/21 16:43:02] ppocr INFO: valid dataloader has 533 iters
[2025/02/21 16:43:02] ppocr WARNING: The shape of model params head.ctc_head.fc.weight [120, 98] not matched with loaded params shape [120, 97] !
[2025/02/21 16:43:02] ppocr WARNING: The shape of model params head.ctc_head.fc.bias [98] not matched with loaded params shape [97] !
[2025/02/21 16:43:02] ppocr WARNING: The shape of model params head.gtc_head.embedding.embedding.weight [102, 384] not matched with loaded params shape [101, 384] !
[2025/02/21 16:43:02] ppocr WARNING: The shape of model params head.gtc_head.tgt_word_prj.weight [384, 102] not matched with loaded params shape [384, 101] !
[2025/02/21 16:43:02] ppocr INFO: resume from ./pretrain_models/en_PP-OCRv4_rec_train/best_accuracy
[2025/02/21 16:43:02] ppocr INFO: During the training process, after the 0th iteration, an evaluation is run every 2000 iterations
Exception in thread Thread-1 (_thread_loop):
Traceback (most recent call last):
File "C:\Python Projects\PaddleOCR\ppocr\data\simple_dataset.py", line 238, in __getitem__
outs = transform(data, self.ops[:-1])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Python Projects\PaddleOCR\ppocr\data\imaug\__init__.py", line 73, in transform
data = op(data)
^^^^^^^^
File "C:\Python Projects\PaddleOCR\ppocr\data\imaug\rec_img_aug.py", line 58, in __call__
img = tia_distort(img, random.randint(3, 6))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Python Projects\PaddleOCR\ppocr\data\imaug\text_image_aug\augment.py", line 63, in tia_distort
dst = trans.generate()
^^^^^^^^^^^^^^^^
File "C:\Python Projects\PaddleOCR\ppocr\data\imaug\text_image_aug\warp_mls.py", line 41, in generate
return self.gen_img()
^^^^^^^^^^^^^^
File "C:\Python Projects\PaddleOCR\ppocr\data\imaug\text_image_aug\warp_mls.py", line 162, in gen_img
nx = np.clip(nx, 0, src_w - 1)
^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\site-packages\numpy\core\fromnumeric.py", line 2169, in clip
return _wrapfunc(a, 'clip', a_min, a_max, out=out, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\site-packages\numpy\core\fromnumeric.py", line 59, in _wrapfunc
return bound(*args, **kwds)
^^^^^^^^^^^^^^^^^^^^
RecursionError: maximum recursion depth exceeded
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\threading.py", line 1075, in _bootstrap_inner
self.run()
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\threading.py", line 1012, in run
self._target(*self._args, **self._kwargs)
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\site-packages\paddle\io\dataloader\dataloader_iter.py", line 235, in _thread_loop
batch = self._dataset_fetcher.fetch(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\site-packages\paddle\io\dataloader\fetcher.py", line 77, in fetch
data.append(self.dataset[idx])
~~~~~~~~~~~~^^^^^
File "C:\Python Projects\PaddleOCR\ppocr\data\simple_dataset.py", line 252, in __getitem__
return self.__getitem__([img_width, img_height, rnd_idx, wh_ratio])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Python Projects\PaddleOCR\ppocr\data\simple_dataset.py", line 252, in __getitem__
return self.__getitem__([img_width, img_height, rnd_idx, wh_ratio])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Python Projects\PaddleOCR\ppocr\data\simple_dataset.py", line 252, in __getitem__
return self.__getitem__([img_width, img_height, rnd_idx, wh_ratio])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[Previous line repeated 984 more times]
File "C:\Python Projects\PaddleOCR\ppocr\data\simple_dataset.py", line 245, in __getitem__
data_line, traceback.format_exc()
^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\traceback.py", line 184, in format_exc
return "".join(format_exception(sys.exception(), limit=limit, chain=chain))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\traceback.py", line 139, in format_exception
te = TracebackException(type(value), value, tb, limit=limit, compact=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\traceback.py", line 733, in __init__
self.stack = StackSummary._extract_from_extended_frame_gen(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\traceback.py", line 438, in _extract_from_extended_frame_gen
f.line
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\traceback.py", line 323, in line
self._line = linecache.getline(self.filename, self.lineno)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\linecache.py", line 30, in getline
lines = getlines(filename, module_globals)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Lian Jiet\AppData\Local\Programs\Python\Python312\Lib\linecache.py", line 46, in getlines
return updatecache(filename, module_globals)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RecursionError: maximum recursion depth exceeded |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The datasets used for both recognition and detection models are different. For recognition models u should crop the lp part and then label it. When I used a full sized image with cars and lp I got the recursion error (this dataset is used for detection training not recognition).
for the detection Label.txt, I can use the file generated by PPOCRLabel directly.
But for recognition I should crop the image to fit just the text then the Labels.txt should be: filepath tabspace text
Do detection training first. After that do recognition training. Don't waste time doing recognition training first