AttributeError: 'SimpleDataSet' object has no attribute 'ds_width'

[anbo724]

W0927 21:53:55.562623 92255 gpu_context.cc:306] device: 1, cuDNN Version: 8.1.
[2023/09/27 21:53:58] ppocr INFO: Initialize indexs of datasets:['./train_data/rec/rec_gt_train.txt']
list index out of range
Traceback (most recent call last):
File "tools/train.py", line 227, in
main(config, device, logger, vdl_writer)
File "tools/train.py", line 53, in main
train_dataloader = build_dataloader(config, 'Train', device, logger)
File "/data/**/ocr/PaddleOCR/ppocr/data/init.py", line 123, in build_dataloader
batch_sampler = eval(sampler_name)(dataset, config_sampler)
File "/data//ocr/PaddleOCR/ppocr/data/multi_scale_sampler.py", line 32, in init
self.ds_width = data_source.ds_width
AttributeError: 'SimpleDataSet' object has no attribute 'ds_width'
INFO 2023-09-27 21:54:04,231 launch_utils.py:322] terminate process group gid:92255
INFO 2023-09-27 21:54:04,231 launch_utils.py:322] terminate process group gid:92255
INFO 2023-09-27 21:54:04,232 launch_utils.py:322] terminate process group gid:92260
INFO 2023-09-27 21:54:04,232 launch_utils.py:322] terminate process group gid:92260
INFO 2023-09-27 21:54:04,233 launch_utils.py:322] terminate process group gid:92265
INFO 2023-09-27 21:54:04,233 launch_utils.py:322] terminate process group gid:92265

config

Global:
debug: false
use_gpu: true
epoch_num: 10
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: refactor
  checkpoints: null
  save_inference_dir: null
  use_visualdl: false
  infer_img: train_data/rec/test/1142515.png
  character_dict_path: ppocr/utils/dict/ipa_dict.txt
  max_text_length: 25
  infer_mode: false
  use_space_char: true
  distributed: true
  save_res_path: ./output/rec/predicts_ppocrv4.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: SimpleDataSet
  ds_width: false
  data_dir: ./train_data/rec
  ext_op_transform_idx: 1
  label_file_list:
  • ./train_data/rec/rec_gt_train.txt
    transforms:
  • DecodeImage:
    img_mode: BGR
    channel_first: false
  • RecConAug:
    prob: 0.5
    ext_data_num: 2
    image_shape:
    • 300
    • 100
    • 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
    drop_last: true
    num_workers: 8
    Eval:
    dataset:
    name: SimpleDataSet
    data_dir: ./train_data/rec
    label_file_list:
  • ./train_data/rec/rec_gt_test.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: 128
      num_workers: 4
      profiler_options: null

[tschijvenaars]

Same issue here. Standard en_PP-OCRv4_rec.yml does not work out of the box:
File "C:\Users\User\Downloads\Project\training\PaddleOCR\ppocr\data\__init__.py", line 101, in build_dataloader module_name = config[mode]['dataset']['name'] KeyError: 'name'

Changing the name of the Train/Eval to SimpleDataSet gives me the error as well.

[Sincerely12]

解决了吗，我是AttributeError: 'LMDBDataSet' object has no attribute 'ds_width'

[phamkhactu]

Hi @Sincerely12 , @tschijvenaars, @anbo724

I've encountered with the same problem. Have you fixed it?

[iPengXPro]

您好，请问v4模型微调您跑通了吗？能指点一下如何操作吗？

[tschijvenaars]

@phamkhactu

Hi, I've encountered this problem when I used my GPU for PaddleOCR and another Deep Learning framework. One script initialized both at the same time, but always the second instance gave an error (also the other framework). Sadly this meant for me I could not initialize both at the same time to make use of my GPU, so I chose CPU paddle version in this case.