rec_mtb_nrtr.yml

Global:
  device: gpu
  epoch_num: 21
  log_smooth_window: 20
  print_batch_step: 10
  output_dir: ./output/rec/nrtr
  eval_epoch_step: [0, 1]
  cal_metric_during_train: true
  pretrained_model:
  checkpoints:
  use_tensorboard: false
  infer_mode: false
  infer_img: doc/imgs_words/en/word_1.png
  character_dict_path: &character_dict_path ppocr/utils/EN_symbol_dict.txt
  max_text_length: &max_text_length 25
  use_space_char: &use_space_char False

Export:
  export_dir:
  export_shape: [ 1, 1, 32, 100 ]
  dynamic_axes: []

Optimizer:
  name: Adam
  lr: 0.0005
  weight_decay: 0

LRScheduler:
  name: CosineAnnealingLR
  warmup_epoch: 2


Architecture:
  model_type: rec
  algorithm: NRTR
  in_channels: 1
  Transform:
  Backbone:
    name: MTB
    cnn_num: 2
  Head:
    name: Transformer
    d_model: 512
    num_encoder_layers: 6
    beam_size: -1 # When Beam size is greater than 0, it means to use beam search when evaluation.
    

Loss:
  name: CELoss
  smoothing: True

PostProcess:
  name: NRTRLabelDecode
  character_dict_path: *character_dict_path
  use_space_char: *use_space_char

Metric:
  name: RecMetric
  main_indicator: acc

Train:
  dataset:
    name: LMDBDataSet
    data_dir: ./train_data/data_lmdb_release/training/
    transforms:
      - DecodeImage: # load image
          img_mode: BGR
          channel_first: False
      - NRTRLabelEncode: # Class handling label
      - GrayRecResizeImg:
          image_shape: [100, 32] # W H
          resize_type: PIL # PIL or OpenCV
      - KeepKeys:
          keep_keys: ['image', 'label', 'length'] # dataloader will return list in this order
  loader:
    shuffle: True
    batch_size_per_card: 512
    drop_last: True
    num_workers: 8

Eval:
  dataset:
    name: LMDBDataSet
    data_dir: ./train_data/data_lmdb_release/validation/
    transforms:
      - DecodeImage: # load image
          img_mode: BGR
          channel_first: False
      - NRTRLabelEncode: # Class handling label
      - GrayRecResizeImg:
          image_shape: [100, 32] # W H
          resize_type: PIL # PIL or OpenCV
      - 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: 4