svtrv2_dan.yml

Global:
  device: gpu
  epoch_num: 20
  log_smooth_window: 20
  print_batch_step: 10
  output_dir: ./output/rec/u14m_filter/svtrv2_dan
  eval_epoch_step: [0, 1]
  eval_batch_step: [0, 500]
  cal_metric_during_train: True
  pretrained_model:
  checkpoints:
  use_tensorboard: false
  infer_img:
  # for data or label process
  character_dict_path: ./tools/utils/EN_symbol_dict.txt
  max_text_length: 25
  use_space_char: False
  save_res_path: ./output/rec/u14m_filter/predicts_svtrv2_dan.txt
  use_amp: True
  grad_clip_val: 20

Optimizer:
  name: AdamW
  lr: 0.00065 # 4gpus 256bs/gpu
  weight_decay: 0.05
  filter_bias_and_bn: True

LRScheduler:
  name: OneCycleLR
  warmup_epoch: 1.5 # pct_start 0.075*20 = 1.5ep
  cycle_momentum: False

Architecture:
  model_type: rec
  algorithm: DAN
  Transform:
  Encoder:
    name: SVTRv2LNConvTwo33
    use_pos_embed: False
    out_channels: 256
    dims: [128, 256, 384]
    depths: [6, 6, 6]
    num_heads: [4, 8, 12]
    mixer: [['Conv','Conv','Conv','Conv','Conv','Conv'],['Conv','Conv','FGlobal','Global','Global','Global'],['Global','Global','Global','Global','Global','Global']]
    local_k: [[5, 5], [5, 5], [-1, -1]]
    sub_k: [[1, 1], [2, 1], [-1, -1]]
    last_stage: false
    feat2d: True
  Decoder:
    name: DANDecoder
    use_cam: False
    max_len: 25

Loss:
  name: ARLoss

PostProcess:
  name: ARLabelDecode

Metric:
  name: RecMetric
  main_indicator: acc
  is_filter: True

Train:
  dataset:
    name: RatioDataSetTVResize
    ds_width: True
    padding: false
    data_dir_list: ['../Union14M-L-LMDB-Filtered/filter_filter_train_challenging',
    '../Union14M-L-LMDB-Filtered/filter_filter_train_hard',
    '../Union14M-L-LMDB-Filtered/filter_filter_train_medium',
    '../Union14M-L-LMDB-Filtered/filter_filter_train_normal',
    '../Union14M-L-LMDB-Filtered/filter_filter_train_easy',
    ]
    transforms:
      - DecodeImagePIL: # load image
          img_mode: RGB
      - PARSeqAugPIL:
      - ARLabelEncode:
      - KeepKeys:
          keep_keys: ['image', 'label', 'length'] # dataloader will return list in this order
  sampler:
    name: RatioSampler
    scales: [[128, 32]] # w, h
    # divide_factor: to ensure the width and height dimensions can be devided by downsampling multiple
    first_bs: &bs 256
    fix_bs: false
    divided_factor: [4, 16] # w, h
    is_training: True
  loader:
    shuffle: True
    batch_size_per_card: *bs
    drop_last: True
    max_ratio: &max_ratio 4
    num_workers: 4

Eval:
  dataset:
    name: RatioDataSetTVResize
    ds_width: True
    padding: False
    data_dir_list: [
      '../evaluation/CUTE80',
      '../evaluation/IC13_857',
      '../evaluation/IC15_1811',
      '../evaluation/IIIT5k',
      '../evaluation/SVT',
      '../evaluation/SVTP',
      ]
    transforms:
      - DecodeImagePIL: # load image
          img_mode: RGB
      - ARLabelEncode:
      - KeepKeys:
          keep_keys: ['image', 'label', 'length'] # dataloader will return list in this order
  sampler:
    name: RatioSampler
    scales: [[128, 32]] # w, h
    # divide_factor: to ensure the width and height dimensions can be devided by downsampling multiple
    first_bs: *bs
    fix_bs: false
    divided_factor: [4, 16] # w, h
    is_training: False
  loader:
    shuffle: False
    drop_last: False
    batch_size_per_card: *bs
    max_ratio: *max_ratio
    num_workers: 4