rec_r34_vd_tps_bilstm_ctc.yml 2.1 KB
Newer Older
W
WenmuZhou 已提交
1 2 3 4 5 6 7
Global:
  use_gpu: true
  epoch_num: 72
  log_smooth_window: 20
  print_batch_step: 10
  save_model_dir: ./output/rec/r34_vd_tps_bilstm_ctc/
  save_epoch_step: 3
W
WenmuZhou 已提交
8
  # evaluation is run every 2000 iterations
W
WenmuZhou 已提交
9 10 11 12 13 14
  eval_batch_step: [0, 2000]
  cal_metric_during_train: True
  pretrained_model:
  checkpoints:
  save_inference_dir:
  use_visualdl: False
W
WenmuZhou 已提交
15
  infer_img: doc/imgs_words_en/word_10.png
W
WenmuZhou 已提交
16 17 18 19 20 21
  # for data or label process
  character_dict_path: 
  character_type: en
  max_text_length: 25
  infer_mode: False
  use_space_char: False
littletomatodonkey's avatar
littletomatodonkey 已提交
22
  save_res_path: ./output/rec/predicts_r34_vd_tps_bilstm_ctc.txt
W
WenmuZhou 已提交
23 24 25 26 27 28 29 30 31 32 33 34 35

Optimizer:
  name: Adam
  beta1: 0.9
  beta2: 0.999
  lr:
    learning_rate: 0.0005
  regularizer:
    name: 'L2'
    factor: 0

Architecture:
  model_type: rec
W
WenmuZhou 已提交
36
  algorithm: STARNet
W
WenmuZhou 已提交
37 38 39 40
  Transform:
    name: TPS
    num_fiducial: 20
    loc_lr: 0.1
W
WenmuZhou 已提交
41
    model_name: large
W
WenmuZhou 已提交
42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64
  Backbone:
    name: ResNet
    layers: 34
  Neck:
    name: SequenceEncoder
    encoder_type: rnn
    hidden_size: 256
  Head:
    name: CTCHead
    fc_decay: 0

Loss:
  name: CTCLoss

PostProcess:
  name: CTCLabelDecode

Metric:
  name: RecMetric
  main_indicator: acc

Train:
  dataset:
T
tink2123 已提交
65
    name: LMDBDataSet
W
WenmuZhou 已提交
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83
    data_dir: ./train_data/data_lmdb_release/training/
    transforms:
      - DecodeImage: # load image
          img_mode: BGR
          channel_first: False
      - CTCLabelEncode: # Class handling label
      - RecResizeImg:
          image_shape: [3, 32, 100]
      - KeepKeys:
          keep_keys: ['image', 'label', 'length'] # dataloader will return list in this order
  loader:
    shuffle: True
    batch_size_per_card: 256
    drop_last: True
    num_workers: 8

Eval:
  dataset:
T
tink2123 已提交
84
    name: LMDBDataSet
W
WenmuZhou 已提交
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
    data_dir: ./train_data/data_lmdb_release/validation/
    transforms:
      - DecodeImage: # load image
          img_mode: BGR
          channel_first: False
      - CTCLabelEncode: # Class handling label
      - RecResizeImg:
          image_shape: [3, 32, 100]
      - 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