Merge pull request #13 from PaddlePaddle/dygraph

Dygraph

MANIFEST.in

 include LICENSE.txt
 include README.md
 
-recursive-include ppocr/utils *.txt utility.py character.py check.py
-recursive-include ppocr/data/det *.py
+recursive-include ppocr/utils *.txt utility.py logging.py
+recursive-include ppocr/data/ *.py
 recursive-include ppocr/postprocess *.py
-recursive-include ppocr/postprocess/lanms *.*
-recursive-include tools/infer *.py
+recursive-include tools/infer *.py
\ No newline at end of file

configs/det/ch_ppocr_v2.0/ch_det_mv3_db_v2.0.yml

+Global:
+  use_gpu: true
+  epoch_num: 1200
+  log_smooth_window: 20
+  print_batch_step: 2
+  save_model_dir: ./output/ch_db_mv3/
+  save_epoch_step: 1200
+  # evaluation is run every 5000 iterations after the 4000th iteration
+  eval_batch_step: [3000, 2000]
+  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  load_static_weights: True
+  cal_metric_during_train: False
+  pretrained_model: ./pretrain_models/MobileNetV3_large_x0_5_pretrained
+  checkpoints: #./output/det_db_0.001_DiceLoss_256_pp_config_2.0b_4gpu/best_accuracy
+  save_inference_dir:
+  use_visualdl: False
+  infer_img: doc/imgs_en/img_10.jpg
+  save_res_path: ./output/det_db/predicts_db.txt
+
+Architecture:
+  model_type: det
+  algorithm: DB
+  Transform:
+  Backbone:
+    name: MobileNetV3
+    scale: 0.5
+    model_name: large
+    disable_se: True
+  Neck:
+    name: DBFPN
+    out_channels: 96
+  Head:
+    name: DBHead
+    k: 50
+
+Loss:
+  name: DBLoss
+  balance_loss: true
+  main_loss_type: DiceLoss
+  alpha: 5
+  beta: 10
+  ohem_ratio: 3
+
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+    name: Cosine
+    learning_rate: 0.001
+    warmup_epoch: 2
+  regularizer:
+    name: 'L2'
+    factor: 0
+
+PostProcess:
+  name: DBPostProcess
+  thresh: 0.3
+  box_thresh: 0.6
+  max_candidates: 1000
+  unclip_ratio: 1.5
+
+Metric:
+  name: DetMetric
+  main_indicator: hmean
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/icdar2015/text_localization/
+    label_file_list:
+      - ./train_data/icdar2015/text_localization/train_icdar2015_label.txt
+    ratio_list: [1.0]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - IaaAugment:
+          augmenter_args:
+            - { 'type': Fliplr, 'args': { 'p': 0.5 } }
+            - { 'type': Affine, 'args': { 'rotate': [-10, 10] } }
+            - { 'type': Resize, 'args': { 'size': [0.5, 3] } }
+      - EastRandomCropData:
+          size: [960, 960]
+          max_tries: 50
+          keep_ratio: true
+      - MakeBorderMap:
+          shrink_ratio: 0.4
+          thresh_min: 0.3
+          thresh_max: 0.7
+      - MakeShrinkMap:
+          shrink_ratio: 0.4
+          min_text_size: 8
+      - NormalizeImage:
+          scale: 1./255.
+          mean: [0.485, 0.456, 0.406]
+          std: [0.229, 0.224, 0.225]
+          order: 'hwc'
+      - ToCHWImage:
+      - KeepKeys:
+          keep_keys: ['image', 'threshold_map', 'threshold_mask', 'shrink_map', 'shrink_mask'] # the order of the dataloader list
+  loader:
+    shuffle: True
+    drop_last: False
+    batch_size_per_card: 8
+    num_workers: 4
+
+Eval:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/icdar2015/text_localization/
+    label_file_list:
+      - ./train_data/icdar2015/text_localization/test_icdar2015_label.txt
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - DetResizeForTest:
+#           image_shape: [736, 1280]
+      - NormalizeImage:
+          scale: 1./255.
+          mean: [0.485, 0.456, 0.406]
+          std: [0.229, 0.224, 0.225]
+          order: 'hwc'
+      - ToCHWImage:
+      - KeepKeys:
+          keep_keys: ['image', 'shape', 'polys', 'ignore_tags']
+  loader:
+    shuffle: False
+    drop_last: False
+    batch_size_per_card: 1 # must be 1
+    num_workers: 2

configs/det/ch_ppocr_v2.0/ch_det_res18_db_v2.0.yml

+Global:
+  use_gpu: true
+  epoch_num: 1200
+  log_smooth_window: 20
+  print_batch_step: 2
+  save_model_dir: ./output/ch_db_res18/
+  save_epoch_step: 1200
+  # evaluation is run every 5000 iterations after the 4000th iteration
+  eval_batch_step: [3000, 2000]
+  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  load_static_weights: True
+  cal_metric_during_train: False
+  pretrained_model: ./pretrain_models/ResNet18_vd_pretrained
+  checkpoints: #./output/det_db_0.001_DiceLoss_256_pp_config_2.0b_4gpu/best_accuracy
+  save_inference_dir:
+  use_visualdl: False
+  infer_img: doc/imgs_en/img_10.jpg
+  save_res_path: ./output/det_db/predicts_db.txt
+
+Architecture:
+  model_type: det
+  algorithm: DB
+  Transform:
+  Backbone:
+    name: ResNet
+    layers: 18
+    disable_se: True
+  Neck:
+    name: DBFPN
+    out_channels: 256
+  Head:
+    name: DBHead
+    k: 50
+
+Loss:
+  name: DBLoss
+  balance_loss: true
+  main_loss_type: DiceLoss
+  alpha: 5
+  beta: 10
+  ohem_ratio: 3
+
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+    name: Cosine
+    learning_rate: 0.001
+    warmup_epoch: 2
+  regularizer:
+    name: 'L2'
+    factor: 0
+
+PostProcess:
+  name: DBPostProcess
+  thresh: 0.3
+  box_thresh: 0.6
+  max_candidates: 1000
+  unclip_ratio: 1.5
+
+Metric:
+  name: DetMetric
+  main_indicator: hmean
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/icdar2015/text_localization/
+    label_file_list:
+      - ./train_data/icdar2015/text_localization/train_icdar2015_label.txt
+    ratio_list: [1.0]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - IaaAugment:
+          augmenter_args:
+            - { 'type': Fliplr, 'args': { 'p': 0.5 } }
+            - { 'type': Affine, 'args': { 'rotate': [-10, 10] } }
+            - { 'type': Resize, 'args': { 'size': [0.5, 3] } }
+      - EastRandomCropData:
+          size: [960, 960]
+          max_tries: 50
+          keep_ratio: true
+      - MakeBorderMap:
+          shrink_ratio: 0.4
+          thresh_min: 0.3
+          thresh_max: 0.7
+      - MakeShrinkMap:
+          shrink_ratio: 0.4
+          min_text_size: 8
+      - NormalizeImage:
+          scale: 1./255.
+          mean: [0.485, 0.456, 0.406]
+          std: [0.229, 0.224, 0.225]
+          order: 'hwc'
+      - ToCHWImage:
+      - KeepKeys:
+          keep_keys: ['image', 'threshold_map', 'threshold_mask', 'shrink_map', 'shrink_mask'] # the order of the dataloader list
+  loader:
+    shuffle: True
+    drop_last: False
+    batch_size_per_card: 8
+    num_workers: 4
+
+Eval:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/icdar2015/text_localization/
+    label_file_list:
+      - ./train_data/icdar2015/text_localization/test_icdar2015_label.txt
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - DetResizeForTest:
+#           image_shape: [736, 1280]
+      - NormalizeImage:
+          scale: 1./255.
+          mean: [0.485, 0.456, 0.406]
+          std: [0.229, 0.224, 0.225]
+          order: 'hwc'
+      - ToCHWImage:
+      - KeepKeys:
+          keep_keys: ['image', 'shape', 'polys', 'ignore_tags']
+  loader:
+    shuffle: False
+    drop_last: False
+    batch_size_per_card: 1 # must be 1
+    num_workers: 2

configs/det/det_mv3_east.yml

+Global:
+  use_gpu: true
+  epoch_num: 10000
+  log_smooth_window: 20
+  print_batch_step: 2
+  save_model_dir: ./output/east_mv3/
+  save_epoch_step: 1000
+  # evaluation is run every 5000 iterations after the 4000th iteration
+  eval_batch_step: [4000, 5000]
+  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  load_static_weights: True
+  cal_metric_during_train: False
+  pretrained_model: ./pretrain_models/MobileNetV3_large_x0_5_pretrained
+  checkpoints: 
+  save_inference_dir:
+  use_visualdl: False
+  infer_img: 
+  save_res_path: ./output/det_east/predicts_east.txt
+
+Architecture:
+  model_type: det
+  algorithm: EAST
+  Transform:
+  Backbone:
+    name: MobileNetV3
+    scale: 0.5
+    model_name: large
+  Neck:
+    name: EASTFPN
+    model_name: small
+  Head:
+    name: EASTHead
+    model_name: small
+
+Loss:
+  name: EASTLoss
+  
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+  #  name: Cosine
+    learning_rate: 0.001
+  #  warmup_epoch: 0
+  regularizer:
+    name: 'L2'
+    factor: 0
+
+PostProcess:
+  name: EASTPostProcess
+  score_thresh: 0.8
+  cover_thresh: 0.1
+  nms_thresh: 0.2
+
+Metric:
+  name: DetMetric
+  main_indicator: hmean
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/icdar2015/text_localization/
+    label_file_list:
+      - ./train_data/icdar2015/text_localization/train_icdar2015_label.txt
+    ratio_list: [1.0]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - EASTProcessTrain:
+          image_shape: [512, 512]
+          background_ratio: 0.125
+          min_crop_side_ratio: 0.1
+          min_text_size: 10
+      - KeepKeys:
+          keep_keys: ['image', 'score_map', 'geo_map', 'training_mask'] # dataloader will return list in this order
+  loader:
+    shuffle: True
+    drop_last: False
+    batch_size_per_card: 16
+    num_workers: 8
+
+Eval:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/icdar2015/text_localization/
+    label_file_list:
+      - ./train_data/icdar2015/text_localization/test_icdar2015_label.txt
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - DetResizeForTest:
+          limit_side_len: 2400
+          limit_type: max
+      - NormalizeImage:
+          scale: 1./255.
+          mean: [0.485, 0.456, 0.406]
+          std: [0.229, 0.224, 0.225]
+          order: 'hwc'
+      - ToCHWImage:
+      - KeepKeys:
+          keep_keys: ['image', 'shape', 'polys', 'ignore_tags']
+  loader:
+    shuffle: False
+    drop_last: False
+    batch_size_per_card: 1 # must be 1
+    num_workers: 2
\ No newline at end of file

configs/det/det_r50_vd_east.yml

+Global:
+  use_gpu: true
+  epoch_num: 10000
+  log_smooth_window: 20
+  print_batch_step: 2
+  save_model_dir: ./output/east_r50_vd/
+  save_epoch_step: 1000
+  # evaluation is run every 5000 iterations after the 4000th iteration
+  eval_batch_step: [4000, 5000]
+  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  load_static_weights: True
+  cal_metric_during_train: False
+  pretrained_model: ./pretrain_models/ResNet50_vd_pretrained/
+  checkpoints: 
+  save_inference_dir:
+  use_visualdl: False
+  infer_img: 
+  save_res_path: ./output/det_east/predicts_east.txt
+
+Architecture:
+  model_type: det
+  algorithm: EAST
+  Transform:
+  Backbone:
+    name: ResNet
+    layers: 50
+  Neck:
+    name: EASTFPN
+    model_name: large
+  Head:
+    name: EASTHead
+    model_name: large
+
+Loss:
+  name: EASTLoss
+  
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+  #  name: Cosine
+    learning_rate: 0.001
+  #  warmup_epoch: 0
+  regularizer:
+    name: 'L2'
+    factor: 0
+
+PostProcess:
+  name: EASTPostProcess
+  score_thresh: 0.8
+  cover_thresh: 0.1
+  nms_thresh: 0.2
+
+Metric:
+  name: DetMetric
+  main_indicator: hmean
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/icdar2015/text_localization/
+    label_file_list:
+      - ./train_data/icdar2015/text_localization/train_icdar2015_label.txt
+    ratio_list: [1.0]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - EASTProcessTrain:
+          image_shape: [512, 512]
+          background_ratio: 0.125
+          min_crop_side_ratio: 0.1
+          min_text_size: 10
+      - KeepKeys:
+          keep_keys: ['image', 'score_map', 'geo_map', 'training_mask'] # dataloader will return list in this order
+  loader:
+    shuffle: True
+    drop_last: False
+    batch_size_per_card: 8
+    num_workers: 8
+
+Eval:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/icdar2015/text_localization/
+    label_file_list:
+      - ./train_data/icdar2015/text_localization/test_icdar2015_label.txt
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - DetResizeForTest:
+          limit_side_len: 2400
+          limit_type: max
+      - NormalizeImage:
+          scale: 1./255.
+          mean: [0.485, 0.456, 0.406]
+          std: [0.229, 0.224, 0.225]
+          order: 'hwc'
+      - ToCHWImage:
+      - KeepKeys:
+          keep_keys: ['image', 'shape', 'polys', 'ignore_tags']
+  loader:
+    shuffle: False
+    drop_last: False
+    batch_size_per_card: 1 # must be 1
+    num_workers: 2
\ No newline at end of file

configs/det/det_r50_vd_sast_icdar15.yml

+Global:
+  use_gpu: true
+  epoch_num: 5000
+  log_smooth_window: 20
+  print_batch_step: 2
+  save_model_dir: ./output/sast_r50_vd_ic15/
+  save_epoch_step: 1000
+  # evaluation is run every 5000 iterations after the 4000th iteration
+  eval_batch_step: [4000, 5000]
+  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  load_static_weights: True
+  cal_metric_during_train: False
+  pretrained_model: ./pretrain_models/ResNet50_vd_ssld_pretrained/
+  checkpoints: 
+  save_inference_dir:
+  use_visualdl: False
+  infer_img: 
+  save_res_path: ./output/sast_r50_vd_ic15/predicts_sast.txt
+
+Architecture:
+  model_type: det
+  algorithm: SAST
+  Transform:
+  Backbone:
+    name: ResNet_SAST
+    layers: 50
+  Neck:
+    name: SASTFPN
+    with_cab: True
+  Head:
+    name: SASTHead
+
+Loss:
+  name: SASTLoss
+  
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+  #  name: Cosine
+    learning_rate: 0.001
+  #  warmup_epoch: 0
+  regularizer:
+    name: 'L2'
+    factor: 0
+
+PostProcess:
+  name: SASTPostProcess
+  score_thresh: 0.5
+  sample_pts_num: 2
+  nms_thresh: 0.2
+  expand_scale: 1.0
+  shrink_ratio_of_width: 0.3
+
+Metric:
+  name: DetMetric
+  main_indicator: hmean
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_path: [./train_data/art_latin_icdar_14pt/train_no_tt_test/train_label_json.txt, ./train_data/total_text_icdar_14pt/train_label_json.txt]
+    data_ratio_list: [0.5, 0.5]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - SASTProcessTrain:
+          image_shape: [512, 512]
+          min_crop_side_ratio: 0.3
+          min_crop_size: 24
+          min_text_size: 4
+          max_text_size: 512
+      - KeepKeys:
+          keep_keys: ['image', 'score_map', 'border_map', 'training_mask', 'tvo_map', 'tco_map'] # dataloader will return list in this order
+  loader:
+    shuffle: True
+    drop_last: False
+    batch_size_per_card: 4
+    num_workers: 4
+
+Eval:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/icdar2015/text_localization/
+    label_file_list:
+      - ./train_data/icdar2015/text_localization/test_icdar2015_label.txt
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - DetResizeForTest:
+          resize_long: 1536
+      - NormalizeImage:
+          scale: 1./255.
+          mean: [0.485, 0.456, 0.406]
+          std: [0.229, 0.224, 0.225]
+          order: 'hwc'
+      - ToCHWImage:
+      - KeepKeys:
+          keep_keys: ['image', 'shape', 'polys', 'ignore_tags']
+  loader:
+    shuffle: False
+    drop_last: False
+    batch_size_per_card: 1 # must be 1
+    num_workers: 2
\ No newline at end of file

configs/det/det_r50_vd_sast_totaltext.yml

+Global:
+  use_gpu: true
+  epoch_num: 5000
+  log_smooth_window: 20
+  print_batch_step: 2
+  save_model_dir: ./output/sast_r50_vd_tt/
+  save_epoch_step: 1000
+  # evaluation is run every 5000 iterations after the 4000th iteration
+  eval_batch_step: [4000, 5000]
+  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  load_static_weights: True
+  cal_metric_during_train: False
+  pretrained_model: ./pretrain_models/ResNet50_vd_ssld_pretrained/
+  checkpoints: 
+  save_inference_dir:
+  use_visualdl: False
+  infer_img: 
+  save_res_path: ./output/sast_r50_vd_tt/predicts_sast.txt
+
+Architecture:
+  model_type: det
+  algorithm: SAST
+  Transform:
+  Backbone:
+    name: ResNet_SAST
+    layers: 50
+  Neck:
+    name: SASTFPN
+    with_cab: True
+  Head:
+    name: SASTHead
+
+Loss:
+  name: SASTLoss
+  
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+  #  name: Cosine
+    learning_rate: 0.001
+  #  warmup_epoch: 0
+  regularizer:
+    name: 'L2'
+    factor: 0
+
+PostProcess:
+  name: SASTPostProcess
+  score_thresh: 0.5
+  sample_pts_num: 6
+  nms_thresh: 0.2
+  expand_scale: 1.2
+  shrink_ratio_of_width: 0.2
+
+Metric:
+  name: DetMetric
+  main_indicator: hmean
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    label_file_list: [./train_data/icdar2013/train_label_json.txt, ./train_data/icdar2015/train_label_json.txt, ./train_data/icdar17_mlt_latin/train_label_json.txt, ./train_data/coco_text_icdar_4pts/train_label_json.txt]
+    ratio_list: [0.1, 0.45, 0.3, 0.15]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - SASTProcessTrain:
+          image_shape: [512, 512]
+          min_crop_side_ratio: 0.3
+          min_crop_size: 24
+          min_text_size: 4
+          max_text_size: 512
+      - KeepKeys:
+          keep_keys: ['image', 'score_map', 'border_map', 'training_mask', 'tvo_map', 'tco_map'] # dataloader will return list in this order
+  loader:
+    shuffle: True
+    drop_last: False
+    batch_size_per_card: 4
+    num_workers: 4
+
+Eval:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_list: 
+        - ./train_data/total_text_icdar_14pt/test_label_json.txt
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - DetLabelEncode: # Class handling label
+      - DetResizeForTest:
+          resize_long: 768
+      - NormalizeImage:
+          scale: 1./255.
+          mean: [0.485, 0.456, 0.406]
+          std: [0.229, 0.224, 0.225]
+          order: 'hwc'
+      - ToCHWImage:
+      - KeepKeys:
+          keep_keys: ['image', 'shape', 'polys', 'ignore_tags']
+  loader:
+    shuffle: False
+    drop_last: False
+    batch_size_per_card: 1 # must be 1
+    num_workers: 2
\ No newline at end of file

configs/rec/ch_ppocr_v1.1/rec_chinese_common_train_v1.1.yaml → configs/rec/ch_ppocr_v2.0/rec_chinese_common_train_v2.0.yaml

@@ -3,7 +3,7 @@ Global:
   epoch_num: 500
   log_smooth_window: 20
   print_batch_step: 10
-  save_model_dir: ./output/rec_chinese_common_v1.1
+  save_model_dir: ./output/rec_chinese_common_v2.0
   save_epoch_step: 3
   # evaluation is run every 5000 iterations after the 4000th iteration
   eval_batch_step: [0, 2000]

configs/rec/ch_ppocr_v1.1/rec_chinese_lite_train_v1.1.yaml → configs/rec/ch_ppocr_v2.0/rec_chinese_lite_train_v2.0.yaml

@@ -3,7 +3,7 @@ Global:
   epoch_num: 500
   log_smooth_window: 20
   print_batch_step: 10
-  save_model_dir: ./output/rec_chinese_lite_v1.1
+  save_model_dir: ./output/rec_chinese_lite_v2.0
   save_epoch_step: 3
   # evaluation is run every 5000 iterations after the 4000th iteration
   eval_batch_step: [0, 2000]
@@ -19,7 +19,7 @@ Global:
   character_type: ch
   max_text_length: 25
   infer_mode: False
-  use_space_char: False
+  use_space_char: True
 
 
 Optimizer:

configs/rec/multi_language/rec_en_number_lite_train.yml

+Global:
+  use_gpu: True
+  epoch_num: 500
+  log_smooth_window: 20
+  print_batch_step: 10
+  save_model_dir: ./output/rec_en_number_lite
+  save_epoch_step: 3
+  # evaluation is run every 5000 iterations after the 4000th iteration
+  eval_batch_step: [0, 2000]
+  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  cal_metric_during_train: True
+  pretrained_model:
+  checkpoints: 
+  save_inference_dir:
+  use_visualdl: False
+  infer_img:
+  # for data or label process
+  character_dict_path: ppocr/utils/dict/en_dict.txt
+  character_type: ch
+  max_text_length: 25
+  infer_mode: False
+  use_space_char: False
+
+
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+    name: Cosine
+    learning_rate: 0.001
+  regularizer:
+    name: 'L2'
+    factor: 0.00001
+
+Architecture:
+  model_type: rec
+  algorithm: CRNN
+  Transform:
+  Backbone:
+    name: MobileNetV3
+    scale: 0.5
+    model_name: small
+    small_stride: [1, 2, 2, 2]
+  Neck:
+    name: SequenceEncoder
+    encoder_type: rnn
+    hidden_size: 48
+  Head:
+    name: CTCHead
+    fc_decay: 0.00001
+
+Loss:
+  name: CTCLoss
+
+PostProcess:
+  name: CTCLabelDecode
+
+Metric:
+  name: RecMetric
+  main_indicator: acc
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_list: ["./train_data/train_list.txt"]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - RecAug: 
+      - CTCLabelEncode: # Class handling label
+      - RecResizeImg:
+          image_shape: [3, 32, 320]
+      - 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:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_list: ["./train_data/eval_list.txt"]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - CTCLabelEncode: # Class handling label
+      - RecResizeImg:
+          image_shape: [3, 32, 320]
+      - 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: 8

configs/rec/multi_language/rec_french_lite_train.yml

+Global:
+  use_gpu: True
+  epoch_num: 500
+  log_smooth_window: 20
+  print_batch_step: 10
+  save_model_dir: ./output/rec_french_lite
+  save_epoch_step: 3
+  # evaluation is run every 5000 iterations after the 4000th iteration
+  eval_batch_step: [0, 2000]
+  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  cal_metric_during_train: True
+  pretrained_model: 
+  checkpoints: 
+  save_inference_dir: 
+  use_visualdl: False
+  infer_img:
+  # for data or label process
+  character_dict_path: ppocr/utils/dict/french_dict.txt
+  character_type: french
+  max_text_length: 25
+  infer_mode: False
+  use_space_char: False
+
+
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+    name: Cosine
+    learning_rate: 0.001
+  regularizer:
+    name: 'L2'
+    factor: 0.00001
+
+Architecture:
+  model_type: rec
+  algorithm: CRNN
+  Transform:
+  Backbone:
+    name: MobileNetV3
+    scale: 0.5
+    model_name: small
+    small_stride: [1, 2, 2, 2]
+  Neck:
+    name: SequenceEncoder
+    encoder_type: rnn
+    hidden_size: 48
+  Head:
+    name: CTCHead
+    fc_decay: 0.00001
+
+Loss:
+  name: CTCLoss
+
+PostProcess:
+  name: CTCLabelDecode
+
+Metric:
+  name: RecMetric
+  main_indicator: acc
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_list: ["./train_data/train_list.txt"]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - RecAug: 
+      - CTCLabelEncode: # Class handling label
+      - RecResizeImg:
+          image_shape: [3, 32, 320]
+      - 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:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_list: ["./train_data/eval_list.txt"]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - CTCLabelEncode: # Class handling label
+      - RecResizeImg:
+          image_shape: [3, 32, 320]
+      - 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: 8

configs/rec/multi_language/rec_german_lite_train.yml

+Global:
+  use_gpu: True
+  epoch_num: 500
+  log_smooth_window: 20
+  print_batch_step: 10
+  save_model_dir: ./output/rec_german_lite
+  save_epoch_step: 3
+  # evaluation is run every 5000 iterations after the 4000th iteration
+  eval_batch_step: [0, 2000]
+  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  cal_metric_during_train: True
+  pretrained_model:
+  checkpoints: 
+  save_inference_dir:
+  use_visualdl: False
+  infer_img:
+  # for data or label process
+  character_dict_path: ppocr/utils/dict/german_dict.txt
+  character_type: german
+  max_text_length: 25
+  infer_mode: False
+  use_space_char: False
+
+
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+    name: Cosine
+    learning_rate: 0.001
+  regularizer:
+    name: 'L2'
+    factor: 0.00001
+
+Architecture:
+  model_type: rec
+  algorithm: CRNN
+  Transform:
+  Backbone:
+    name: MobileNetV3
+    scale: 0.5
+    model_name: small
+    small_stride: [1, 2, 2, 2]
+  Neck:
+    name: SequenceEncoder
+    encoder_type: rnn
+    hidden_size: 48
+  Head:
+    name: CTCHead
+    fc_decay: 0.00001
+
+Loss:
+  name: CTCLoss
+
+PostProcess:
+  name: CTCLabelDecode
+
+Metric:
+  name: RecMetric
+  main_indicator: acc
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_list: ["./train_data/train_list.txt"]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - RecAug: 
+      - CTCLabelEncode: # Class handling label
+      - RecResizeImg:
+          image_shape: [3, 32, 320]
+      - 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:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_list: ["./train_data/eval_list.txt"]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - CTCLabelEncode: # Class handling label
+      - RecResizeImg:
+          image_shape: [3, 32, 320]
+      - 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: 8

configs/rec/multi_language/rec_japan_lite_train.yml

+Global:
+  use_gpu: True
+  epoch_num: 500
+  log_smooth_window: 20
+  print_batch_step: 10
+  save_model_dir: ./output/rec_japan_lite
+  save_epoch_step: 3
+  # evaluation is run every 5000 iterations after the 4000th iteration
+  eval_batch_step: [0, 2000]
+  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  cal_metric_during_train: True
+  pretrained_model:
+  checkpoints: 
+  save_inference_dir:
+  use_visualdl: False
+  infer_img:
+  # for data or label process
+  character_dict_path: ppocr/utils/dict/japan_dict.txt
+  character_type: japan
+  max_text_length: 25
+  infer_mode: False
+  use_space_char: False
+
+
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+    name: Cosine
+    learning_rate: 0.001
+  regularizer:
+    name: 'L2'
+    factor: 0.00001
+
+Architecture:
+  model_type: rec
+  algorithm: CRNN
+  Transform:
+  Backbone:
+    name: MobileNetV3
+    scale: 0.5
+    model_name: small
+    small_stride: [1, 2, 2, 2]
+  Neck:
+    name: SequenceEncoder
+    encoder_type: rnn
+    hidden_size: 48
+  Head:
+    name: CTCHead
+    fc_decay: 0.00001
+
+Loss:
+  name: CTCLoss
+
+PostProcess:
+  name: CTCLabelDecode
+
+Metric:
+  name: RecMetric
+  main_indicator: acc
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_list: ["./train_data/train_list.txt"]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - RecAug: 
+      - CTCLabelEncode: # Class handling label
+      - RecResizeImg:
+          image_shape: [3, 32, 320]
+      - 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:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_list: ["./train_data/eval_list.txt"]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - CTCLabelEncode: # Class handling label
+      - RecResizeImg:
+          image_shape: [3, 32, 320]
+      - 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: 8

configs/rec/multi_language/rec_korean_lite_train.yml

+Global:
+  use_gpu: True
+  epoch_num: 500
+  log_smooth_window: 20
+  print_batch_step: 10
+  save_model_dir: ./output/rec_korean_lite
+  save_epoch_step: 3
+  # evaluation is run every 5000 iterations after the 4000th iteration
+  eval_batch_step: [0, 2000]
+  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  cal_metric_during_train: True
+  pretrained_model:
+  checkpoints: 
+  save_inference_dir:
+  use_visualdl: False
+  infer_img:
+  # for data or label process
+  character_dict_path: ppocr/utils/dict/korean_dict.txt
+  character_type: korean
+  max_text_length: 25
+  infer_mode: False
+  use_space_char: False
+
+
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+    name: Cosine
+    learning_rate: 0.001
+  regularizer:
+    name: 'L2'
+    factor: 0.00001
+
+Architecture:
+  model_type: rec
+  algorithm: CRNN
+  Transform:
+  Backbone:
+    name: MobileNetV3
+    scale: 0.5
+    model_name: small
+    small_stride: [1, 2, 2, 2]
+  Neck:
+    name: SequenceEncoder
+    encoder_type: rnn
+    hidden_size: 48
+  Head:
+    name: CTCHead
+    fc_decay: 0.00001
+
+Loss:
+  name: CTCLoss
+
+PostProcess:
+  name: CTCLabelDecode
+
+Metric:
+  name: RecMetric
+  main_indicator: acc
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_list: ["./train_data/train_list.txt"]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - RecAug: 
+      - CTCLabelEncode: # Class handling label
+      - RecResizeImg:
+          image_shape: [3, 32, 320]
+      - 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:
+    name: SimpleDataSet
+    data_dir: ./train_data/
+    label_file_list: ["./train_data/eval_list.txt"]
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+      - CTCLabelEncode: # Class handling label
+      - RecResizeImg:
+          image_shape: [3, 32, 320]
+      - 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: 8

deploy/cpp_infer/src/ocr_cls.cpp

@@ -81,7 +81,7 @@ cv::Mat Classifier::Run(cv::Mat &img) {
 
 void Classifier::LoadModel(const std::string &model_dir) {
   AnalysisConfig config;
-  config.SetModel(model_dir + "/model", model_dir + "/params");
+  config.SetModel(model_dir + ".pdmodel", model_dir + ".pdiparams");
 
   if (this->use_gpu_) {
     config.EnableUseGpu(this->gpu_mem_, this->gpu_id_);

deploy/cpp_infer/src/ocr_det.cpp

@@ -18,7 +18,7 @@ namespace PaddleOCR {
 
 void DBDetector::LoadModel(const std::string &model_dir) {
   AnalysisConfig config;
-  config.SetModel(model_dir + "/model", model_dir + "/params");
+  config.SetModel(model_dir + ".pdmodel", model_dir + ".pdiparams");
 
   if (this->use_gpu_) {
     config.EnableUseGpu(this->gpu_mem_, this->gpu_id_);

deploy/cpp_infer/src/ocr_rec.cpp

@@ -103,7 +103,7 @@ void CRNNRecognizer::Run(std::vector<std::vector<std::vector<int>>> boxes,
 
 void CRNNRecognizer::LoadModel(const std::string &model_dir) {
   AnalysisConfig config;
-  config.SetModel(model_dir + "/model", model_dir + "/params");
+  config.SetModel(model_dir + ".pdmodel", model_dir + ".pdiparams");
 
   if (this->use_gpu_) {
     config.EnableUseGpu(this->gpu_mem_, this->gpu_id_);
@@ -186,4 +186,4 @@ cv::Mat CRNNRecognizer::GetRotateCropImage(const cv::Mat &srcimage,
   }
 }
 
-} // namespace PaddleOCR
\ No newline at end of file
+} // namespace PaddleOCR

deploy/docker/hubserving/README.md

+English | [简体中文](README_cn.md)
+
+## Introduction
+Many users hope package the PaddleOCR service into a docker image, so that it can be quickly released and used in the docker or k8s environment.
+
+This page provides some standardized code to achieve this goal. You can quickly publish the PaddleOCR project into a callable Restful API service through the following steps. (At present, the deployment based on the HubServing mode is implemented first, and author plans to increase the deployment of the PaddleServing mode in the futrue)
+
+## 1. Prerequisites
+
+You need to install the following basic components first：
+a. Docker
+b. Graphics driver and CUDA 10.0+（GPU）
+c. NVIDIA Container Toolkit（GPU，Docker 19.03+ can skip this）
+d. cuDNN 7.6+（GPU）
+
+## 2. Build Image
+a. Goto Dockerfile directory（ps：Need to distinguish between cpu and gpu version, the following takes cpu as an example, gpu version needs to replace the keyword）
+```
+cd deploy/docker/hubserving/cpu
+```
+c. Build image
+```
+docker build -t paddleocr:cpu .
+```
+
+## 3. Start container
+a. CPU version
+```
+sudo docker run -dp 8868:8868 --name paddle_ocr paddleocr:cpu
+```
+b. GPU version (base on NVIDIA Container Toolkit)
+```
+sudo nvidia-docker run -dp 8868:8868 --name paddle_ocr paddleocr:gpu
+```
+c. GPU version (Docker 19.03++)
+```
+sudo docker run -dp 8868:8868 --gpus all --name paddle_ocr paddleocr:gpu
+```
+d. Check service status（If you can see the following statement then it means completed：Successfully installed ocr_system && Running on http://0.0.0.0:8868/）
+```
+docker logs -f paddle_ocr
+```
+
+## 4. Test
+a. Calculate the Base64 encoding of the picture to be recognized (if you just test, you can use a free online tool, like：https://freeonlinetools24.com/base64-image/）
+b. Post a service request（sample request in sample_request.txt）
+
+```
+curl -H "Content-Type:application/json" -X POST --data "{\"images\": [\"Input image Base64 encode(need to delete the code 'data:image/jpg;base64,'）\"]}" http://localhost:8868/predict/ocr_system
+```
+c. Get resposne（If the call is successful, the following result will be returned）
+```
+{"msg":"","results":[[{"confidence":0.8403433561325073,"text":"约定","text_region":[[345,377],[641,390],[634,540],[339,528]]},{"confidence":0.8131805658340454,"text":"最终相遇","text_region":[[356,532],[624,530],[624,596],[356,598]]}]],"status":"0"}
+```

deploy/docker/hubserving/README_cn.md

+[English](README.md) | 简体中文
+
+## Docker化部署服务
+在日常项目应用中，相信大家一般都会希望能通过Docker技术，把PaddleOCR服务打包成一个镜像，以便在Docker或k8s环境里，快速发布上线使用。
+
+本文将提供一些标准化的代码来实现这样的目标。大家通过如下步骤可以把PaddleOCR项目快速发布成可调用的Restful API服务。（目前暂时先实现了基于HubServing模式的部署，后续作者计划增加PaddleServing模式的部署）
+
+## 1.实施前提准备
+
+需要先完成如下基本组件的安装：
+a. Docker环境
+b. 显卡驱动和CUDA 10.0+（GPU）
+c. NVIDIA Container Toolkit（GPU，Docker 19.03以上版本可以跳过此步）
+d. cuDNN 7.6+（GPU）
+
+## 2.制作镜像
+a.切换至Dockerfile目录（注：需要区分cpu或gpu版本，下文以cpu为例，gpu版本需要替换一下关键字即可）
+```
+cd deploy/docker/hubserving/cpu
+```
+c.生成镜像
+```
+docker build -t paddleocr:cpu .
+```
+
+## 3.启动Docker容器
+a. CPU 版本
+```
+sudo docker run -dp 8868:8868 --name paddle_ocr paddleocr:cpu
+```
+b. GPU 版本 (通过NVIDIA Container Toolkit)
+```
+sudo nvidia-docker run -dp 8868:8868 --name paddle_ocr paddleocr:gpu
+```
+c. GPU 版本 (Docker 19.03以上版本，可以直接用如下命令)
+```
+sudo docker run -dp 8868:8869 --gpus all --name paddle_ocr paddleocr:gpu
+```
+d. 检查服务运行情况（出现：Successfully installed ocr_system和Running on http://0.0.0.0:8868 等信息，表示运行成功）
+```
+docker logs -f paddle_ocr
+```
+
+## 4.测试服务
+a. 计算待识别图片的Base64编码（如果只是测试一下效果，可以通过免费的在线工具实现，如：http://tool.chinaz.com/tools/imgtobase/）
+b. 发送服务请求（可参见sample_request.txt中的值）
+```
+curl -H "Content-Type:application/json" -X POST --data "{\"images\": [\"填入图片Base64编码(需要删除'data:image/jpg;base64,'）\"]}" http://localhost:8868/predict/ocr_system
+```
+c. 返回结果（如果调用成功，会返回如下结果）
+```
+{"msg":"","results":[[{"confidence":0.8403433561325073,"text":"约定","text_region":[[345,377],[641,390],[634,540],[339,528]]},{"confidence":0.8131805658340454,"text":"最终相遇","text_region":[[356,532],[624,530],[624,596],[356,598]]}]],"status":"0"}
+```

deploy/docker/hubserving/cpu/Dockerfile

+# Version: 1.0.0
+FROM hub.baidubce.com/paddlepaddle/paddle:latest-gpu-cuda10.0-cudnn7-dev
+
+# PaddleOCR base on Python3.7
+RUN pip3.7 install --upgrade pip -i https://mirror.baidu.com/pypi/simple
+
+RUN python3.7 -m pip install paddlepaddle==2.0.0rc0 -i https://mirror.baidu.com/pypi/simple
+
+RUN pip3.7 install paddlehub --upgrade -i https://mirror.baidu.com/pypi/simple
+
+RUN git clone https://github.com/PaddlePaddle/PaddleOCR.git /PaddleOCR
+
+WORKDIR /PaddleOCR
+
+RUN pip3.7 install -r requirements.txt -i https://mirror.baidu.com/pypi/simple
+
+RUN mkdir -p /PaddleOCR/inference/
+# Download orc detect model(light version). if you want to change normal version, you can change ch_ppocr_mobile_v1.1_det_infer to ch_ppocr_server_v1.1_det_infer, also remember change det_model_dir in deploy/hubserving/ocr_system/params.py）
+ADD {link} /PaddleOCR/inference/
+RUN tar xf /PaddleOCR/inference/{file} -C /PaddleOCR/inference/
+
+# Download direction classifier(light version). If you want to change normal version, you can change ch_ppocr_mobile_v1.1_cls_infer to ch_ppocr_mobile_v1.1_cls_infer, also remember change cls_model_dir in deploy/hubserving/ocr_system/params.py）
+ADD {link} /PaddleOCR/inference/
+RUN tar xf /PaddleOCR/inference/{file}.tar -C /PaddleOCR/inference/
+
+# Download orc recognition model(light version). If you want to change normal version, you can change ch_ppocr_mobile_v1.1_rec_infer to ch_ppocr_server_v1.1_rec_infer, also remember change rec_model_dir in deploy/hubserving/ocr_system/params.py）
+ADD {link} /PaddleOCR/inference/
+RUN tar xf /PaddleOCR/inference/{file}.tar -C /PaddleOCR/inference/
+
+EXPOSE 8868
+
+CMD ["/bin/bash","-c","hub install deploy/hubserving/ocr_system/ && hub serving start -m ocr_system"]
\ No newline at end of file

deploy/docker/hubserving/gpu/Dockerfile

+# Version: 1.0.0
+FROM hub.baidubce.com/paddlepaddle/paddle:latest-gpu-cuda10.0-cudnn7-dev
+
+# PaddleOCR base on Python3.7
+RUN pip3.7 install --upgrade pip -i https://mirror.baidu.com/pypi/simple
+
+RUN python3.7 -m pip install paddlepaddle-gpu==2.0.0rc0 -i https://mirror.baidu.com/pypi/simple
+
+RUN pip3.7 install paddlehub --upgrade -i https://mirror.baidu.com/pypi/simple
+
+RUN git clone https://github.com/PaddlePaddle/PaddleOCR.git /PaddleOCR
+
+WORKDIR /PaddleOCR
+
+RUN pip3.7 install -r requirements.txt -i https://mirror.baidu.com/pypi/simple
+
+RUN mkdir -p /PaddleOCR/inference/
+# Download orc detect model(light version). if you want to change normal version, you can change ch_ppocr_mobile_v1.1_det_infer to ch_ppocr_server_v1.1_det_infer, also remember change det_model_dir in deploy/hubserving/ocr_system/params.py）
+ADD {link} /PaddleOCR/inference/
+RUN tar xf /PaddleOCR/inference/{file}.tar -C /PaddleOCR/inference/
+
+# Download direction classifier(light version). If you want to change normal version, you can change ch_ppocr_mobile_v1.1_cls_infer to ch_ppocr_mobile_v1.1_cls_infer, also remember change cls_model_dir in deploy/hubserving/ocr_system/params.py）
+ADD {link} /PaddleOCR/inference/
+RUN tar xf /PaddleOCR/inference/{file} -C /PaddleOCR/inference/
+
+# Download orc recognition model(light version). If you want to change normal version, you can change ch_ppocr_mobile_v1.1_rec_infer to ch_ppocr_server_v1.1_rec_infer, also remember change rec_model_dir in deploy/hubserving/ocr_system/params.py）
+ADD {link} /PaddleOCR/inference/
+RUN tar xf /PaddleOCR/inference/{file}.tar -C /PaddleOCR/inference/
+
+EXPOSE 8868
+
+CMD ["/bin/bash","-c","hub install deploy/hubserving/ocr_system/ && hub serving start -m ocr_system"]
\ No newline at end of file

doc/doc_ch/add_new_algorithm.md

+# 添加新算法
+
+PaddleOCR将一个算法分解为以下几个部分，并对各部分进行模块化处理，方便快速组合出新的算法。
+
+* 数据加载和处理
+* 网络
+* 后处理
+* 损失函数
+* 指标评估
+* 优化器
+
+下面将分别对每个部分进行介绍，并介绍如何在该部分里添加新算法所需模块。
+
+## 数据加载和处理
+
+数据加载和处理由不同的模块(module)组成，其完成了图片的读取、数据增强和label的制作。这一部分在[ppocr/data](../../ppocr/data)下。 各个文件及文件夹作用说明如下:
+
+```bash
+ppocr/data/
+├── imaug             # 图片的读取、数据增强和label制作相关的文件
+│   ├── label_ops.py  # 对label进行变换的modules
+│   ├── operators.py  # 对image进行变换的modules
+│   ├──.....
+├── __init__.py
+├── lmdb_dataset.py   # 读取lmdb的数据集的dataset
+└── simple_dataset.py # 读取以`image_path\tgt`形式保存的数据集的dataset
+```
+
+PaddleOCR内置了大量图像操作相关模块，对于没有没有内置的模块可通过如下步骤添加:
+
+1. 在 [ppocr/data/imaug](../../ppocr/data/imaug) 文件夹下新建文件，如my_module.py。
+2. 在 my_module.py 文件内添加相关代码，示例代码如下:
+
+```python
+class MyModule:
+    def __init__(self, *args, **kwargs):
+        # your init code
+        pass
+
+    def __call__(self, data):
+        img = data['image']
+        label = data['label']
+        # your process code
+
+        data['image'] = img
+        data['label'] = label
+        return data
+```
+
+3. 在 [ppocr/data/imaug/\__init\__.py](../../ppocr/data/imaug/__init__.py) 文件内导入添加的模块。
+
+数据处理的所有处理步骤由不同的模块顺序执行而成，在config文件中按照列表的形式组合并执行。如:
+
+```yaml
+# angle class data process
+transforms:
+  - DecodeImage: # load image
+      img_mode: BGR
+      channel_first: False
+  - MyModule:
+      args1: args1
+      args2: args2
+  - KeepKeys:
+      keep_keys: [ 'image', 'label' ] # dataloader will return list in this order
+```
+
+## 网络
+
+网络部分完成了网络的组网操作，PaddleOCR将网络划分为四部分，这一部分在[ppocr/modeling](../../ppocr/modeling)下。 进入网络的数据将按照顺序(transforms->backbones->
+necks->heads)依次通过这四个部分。
+
+```bash
+├── architectures # 网络的组网代码
+├── transforms    # 网络的图像变换模块
+├── backbones     # 网络的特征提取模块
+├── necks         # 网络的特征增强模块
+└── heads         # 网络的输出模块
+```
+
+PaddleOCR内置了DB,EAST,SAST,CRNN和Attention等算法相关的常用模块，对于没有内置的模块可通过如下步骤添加，四个部分添加步骤一致，以backbones为例:
+
+1. 在 [ppocr/modeling/backbones](../../ppocr/modeling/backbones) 文件夹下新建文件，如my_backbone.py。
+2. 在 my_backbone.py 文件内添加相关代码，示例代码如下:
+
+```python
+import paddle
+import paddle.nn as nn
+import paddle.nn.functional as F
+
+
+class MyBackbone(nn.Layer):
+    def __init__(self, *args, **kwargs):
+        super(MyBackbone, self).__init__()
+        # your init code
+        self.conv = nn.xxxx
+
+    def forward(self, inputs):
+        # your necwork forward
+        y = self.conv(inputs)
+        return y
+```
+
+3. 在 [ppocr/modeling/backbones/\__init\__.py](../../ppocr/modeling/backbones/__init__.py)文件内导入添加的模块。
+
+在完成网络的四部分模块添加之后，只需要配置文件中进行配置即可使用，如:
+
+```yaml
+Architecture:
+  model_type: rec
+  algorithm: CRNN
+  Transform:
+    name: MyTransform
+    args1: args1
+    args2: args2
+  Backbone:
+    name: MyBackbone
+    args1: args1
+  Neck:
+    name: MyNeck
+    args1: args1
+  Head:
+    name: MyHead
+    args1: args1
+```
+
+## 后处理
+
+后处理实现解码网络输出获得文本框或者识别到的文字。这一部分在[ppocr/postprocess](../../ppocr/postprocess)下。
+PaddleOCR内置了DB,EAST,SAST,CRNN和Attention等算法相关的后处理模块，对于没有内置的组件可通过如下步骤添加:
+
+1. 在 [ppocr/postprocess](../../ppocr/postprocess) 文件夹下新建文件，如 my_postprocess.py。
+2. 在 my_postprocess.py 文件内添加相关代码，示例代码如下:
+
+```python
+import paddle
+
+
+class MyPostProcess:
+    def __init__(self, *args, **kwargs):
+        # your init code
+        pass
+
+    def __call__(self, preds, label=None, *args, **kwargs):
+        if isinstance(preds, paddle.Tensor):
+            preds = preds.numpy()
+        # you preds decode code
+        preds = self.decode_preds(preds)
+        if label is None:
+            return preds
+        # you label decode code
+        label = self.decode_label(label)
+        return preds, label
+
+    def decode_preds(self, preds):
+        # you preds decode code
+        pass
+
+    def decode_label(self, preds):
+        # you label decode code
+        pass
+```
+
+3. 在 [ppocr/postprocess/\__init\__.py](../../ppocr/postprocess/__init__.py)文件内导入添加的模块。
+
+在后处理模块添加之后，只需要配置文件中进行配置即可使用，如:
+
+```yaml
+PostProcess:
+  name: MyPostProcess
+  args1: args1
+  args2: args2
+```
+
+## 损失函数
+
+损失函数用于计算网络输出和label之间的距离。这一部分在[ppocr/losses](../../ppocr/losses)下。
+PaddleOCR内置了DB,EAST,SAST,CRNN和Attention等算法相关的损失函数模块，对于没有内置的模块可通过如下步骤添加:
+
+1. 在 [ppocr/losses](../../ppocr/losses) 文件夹下新建文件，如 my_loss.py。
+2. 在 my_loss.py 文件内添加相关代码，示例代码如下:
+
+```python
+import paddle
+from paddle import nn
+
+
+class MyLoss(nn.Layer):
+    def __init__(self, **kwargs):
+        super(MyLoss, self).__init__()
+        # you init code
+        pass
+
+    def __call__(self, predicts, batch):
+        label = batch[1]
+        # your loss code
+        loss = self.loss(input=predicts, label=label)
+        return {'loss': loss}
+```
+
+3. 在 [ppocr/losses/\__init\__.py](../../ppocr/losses/__init__.py)文件内导入添加的模块。
+
+在损失函数添加之后，只需要配置文件中进行配置即可使用，如:
+
+```yaml
+Loss:
+  name: MyLoss
+  args1: args1
+  args2: args2
+```
+
+## 指标评估
+
+指标评估用于计算网络在当前batch上的性能。这一部分在[ppocr/metrics](../../ppocr/metrics)下。 PaddleOCR内置了检测，分类和识别等算法相关的指标评估模块，对于没有内置的模块可通过如下步骤添加:
+
+1. 在 [ppocr/metrics](../../ppocr/metrics) 文件夹下新建文件，如my_metric.py。
+2. 在 my_metric.py 文件内添加相关代码，示例代码如下:
+
+```python
+
+class MyMetric(object):
+    def __init__(self, main_indicator='acc', **kwargs):
+        # main_indicator is used for select best model
+        self.main_indicator = main_indicator
+        self.reset()
+
+    def __call__(self, preds, batch, *args, **kwargs):
+        # preds is out of postprocess
+        # batch is out of dataloader
+        labels = batch[1]
+        cur_correct_num = 0
+        cur_all_num = 0
+        # you metric code
+        self.correct_num += cur_correct_num
+        self.all_num += cur_all_num
+        return {'acc': cur_correct_num / cur_all_num, }
+
+    def get_metric(self):
+        """
+        return metircs {
+                 'acc': 0,
+                 'norm_edit_dis': 0,
+            }
+        """
+        acc = self.correct_num / self.all_num
+        self.reset()
+        return {'acc': acc}
+
+    def reset(self):
+        # reset metric
+        self.correct_num = 0
+        self.all_num = 0
+
+```
+
+3. 在 [ppocr/metrics/\__init\__.py](../../ppocr/metrics/__init__.py)文件内导入添加的模块。
+
+在指标评估模块添加之后，只需要配置文件中进行配置即可使用，如:
+
+```yaml
+Metric:
+  name: MyMetric
+  main_indicator: acc
+```
+
+## 优化器
+
+优化器用于训练网络。优化器内部还包含了网络正则化和学习率衰减模块。 这一部分在[ppocr/optimizer](../../ppocr/optimizer)下。 PaddleOCR内置了`Momentum`,`Adam`
+和`RMSProp`等常用的优化器模块，`Linear`,`Cosine`,`Step`和`Piecewise`等常用的正则化模块与`L1Decay`和`L2Decay`等常用的学习率衰减模块。
+对于没有内置的模块可通过如下步骤添加，以`optimizer`为例:
+
+1. 在 [ppocr/optimizer/optimizer.py](../../ppocr/optimizer/optimizer.py) 文件内创建自己的优化器，示例代码如下:
+
+```python
+from paddle import optimizer as optim
+
+
+class MyOptim(object):
+    def __init__(self, learning_rate=0.001, *args, **kwargs):
+        self.learning_rate = learning_rate
+
+    def __call__(self, parameters):
+        # It is recommended to wrap the built-in optimizer of paddle
+        opt = optim.XXX(
+            learning_rate=self.learning_rate,
+            parameters=parameters)
+        return opt
+
+```
+
+在优化器模块添加之后，只需要配置文件中进行配置即可使用，如:
+
+```yaml
+Optimizer:
+  name: MyOptim
+  args1: args1
+  args2: args2
+  lr:
+    name: Cosine
+    learning_rate: 0.001
+  regularizer:
+    name: 'L2'
+    factor: 0
+```
\ No newline at end of file

doc/doc_ch/angle_class.md

@@ -45,7 +45,7 @@ train_data/cls/word_002.jpg   180
 ```
 |-train_data
     |-cls
-        |- 和一个cls_gt_test.txt
+        |- cls_gt_test.txt
         |- test
             |- word_001.jpg
             |- word_002.jpg
@@ -62,29 +62,36 @@ PaddleOCR提供了训练脚本、评估脚本和预测脚本。
 *如果您安装的是cpu版本，请将配置文件中的 `use_gpu` 字段修改为false*
 
 ```
-# 设置PYTHONPATH路径
-export PYTHONPATH=$PYTHONPATH:.
-# GPU训练 支持单卡，多卡训练，通过CUDA_VISIBLE_DEVICES指定卡号
-export CUDA_VISIBLE_DEVICES=0,1,2,3
-# 启动训练
-python3 tools/train.py -c configs/cls/cls_mv3.yml
+# GPU训练 支持单卡，多卡训练，通过selected_gpus指定卡号
+# 启动训练，下面的命令已经写入train.sh文件中，只需修改文件里的配置文件路径即可
+python3 -m paddle.distributed.launch --selected_gpus '0,1,2,3,4,5,6,7'  tools/train.py -c configs/cls/cls_mv3.yml
 ```
 
 - 数据增强
 
-PaddleOCR提供了多种数据增强方式，如果您希望在训练时加入扰动，请在配置文件中设置 `distort: true`。
+PaddleOCR提供了多种数据增强方式，如果您希望在训练时加入扰动，请在配置文件中取消`Train.dataset.transforms`下的`RecAug`和`RandAugment`字段的注释。
 
 默认的扰动方式有：颜色空间转换(cvtColor)、模糊(blur)、抖动(jitter)、噪声(Gasuss noise)、随机切割(random crop)、透视(perspective)、颜色反转(reverse),随机数据增强(RandAugment)。
 
 训练过程中除随机数据增强外每种扰动方式以50%的概率被选择，具体代码实现请参考：
-[randaugment.py](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/ppocr/data/cls/randaugment.py)
-[img_tools.py](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/ppocr/data/rec/img_tools.py)
+[rec_img_aug.py](../../ppocr/data/imaug/rec_img_aug.py) 
+[randaugment.py](../../ppocr/data/imaug/randaugment.py)
 
 *由于OpenCV的兼容性问题，扰动操作暂时只支持linux*
 
 ### 训练
 
-PaddleOCR支持训练和评估交替进行, 可以在 `configs/cls/cls_mv3.yml` 中修改 `eval_batch_step` 设置评估频率，默认每500个iter评估一次。评估过程中默认将最佳acc模型，保存为 `output/cls_mv3/best_accuracy` 。
+PaddleOCR支持训练和评估交替进行, 可以在 `configs/cls/cls_mv3.yml` 中修改 `eval_batch_step` 设置评估频率，默认每1000个iter评估一次。训练过程中将会保存如下内容：
+```bash
+├── best_accuracy.pdopt # 最佳模型的优化器参数
+├── best_accuracy.pdparams # 最佳模型的参数
+├── best_accuracy.states # 最佳模型的指标和epoch等信息
+├── config.yml # 本次实验的配置文件
+├── latest.pdopt # 最新模型的优化器参数
+├── latest.pdparams # 最新模型的参数
+├── latest.states # 最新模型的指标和epoch等信息
+└── train.log # 训练日志
+```
 
 如果验证集很大，测试将会比较耗时，建议减少评估次数，或训练完再进行评估。
 
@@ -92,9 +99,8 @@ PaddleOCR支持训练和评估交替进行, 可以在 `configs/cls/cls_mv3.yml`
 
 ### 评估
 
-评估数据集可以通过`configs/cls/cls_reader.yml`  修改EvalReader中的 `label_file_path` 设置。
+评估数据集可以通过修改`configs/cls/cls_mv3.yml`文件里的`Eval.dataset.label_file_list` 字段设置。
 
-*注意* 评估时必须确保配置文件中 infer_img 字段为空
 ```
 export CUDA_VISIBLE_DEVICES=0
 # GPU 评估， Global.checkpoints 为待测权重
@@ -107,21 +113,20 @@ python3 tools/eval.py -c configs/cls/cls_mv3.yml -o Global.checkpoints={path/to/
 
 使用 PaddleOCR 训练好的模型，可以通过以下脚本进行快速预测。
 
-默认预测图片存储在 `infer_img` 里，通过 `-o Global.checkpoints` 指定权重：
+通过 `Global.infer_img` 指定预测图片或文件夹路径，通过 `Global.checkpoints` 指定权重：
 
 ```
 # 预测分类结果
-python3 tools/infer_cls.py -c configs/cls/cls_mv3.yml -o Global.checkpoints={path/to/weights}/best_accuracy Global.infer_img=doc/imgs_words/en/word_1.png
+python3 tools/infer_cls.py -c configs/cls/cls_mv3.yml -o Global.checkpoints={path/to/weights}/best_accuracy Global.infer_img=doc/imgs_words/ch/word_1.jpg
 ```
 
 预测图片：
 
-![](../imgs_words/en/word_1.png)
+![](../imgs_words/ch/word_1.jpg)
 
 得到输入图像的预测结果：
 
 ```
-infer_img: doc/imgs_words/en/word_1.png
-    scores: [[0.93161047 0.06838956]]
-    label: [0]
+infer_img: doc/imgs_words/ch/word_1.jpg
+     result: ('0', 0.9998784)
 ```

doc/doc_ch/config.md

-# 可选参数列表
+## 可选参数列表
 
 以下列表可以通过`--help`查看
 
@@ -8,65 +8,115 @@
 |          -o              |      ALL       |  设置配置文件里的参数内容  |  None  |  使用-o配置相较于-c选择的配置文件具有更高的优先级。例如：`-o Global.use_gpu=false`  |  
 
 
-## 配置文件 Global 参数介绍
+## 配置文件参数介绍
 
 以 `rec_chinese_lite_train_v1.1.yml ` 为例
-
+### Global 
 
 |         字段             |            用途                |      默认值       |            备注            |
 | :----------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
-|      algorithm           |    设置算法                    |  与配置文件同步   |     选择模型，支持模型请参考[简介](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/README.md) |
-|      use_gpu             |    设置代码运行场所            |       true        |                \                 |
-|      epoch_num           |    最大训练epoch数             |       3000        |                \                 |
+|      use_gpu             |    设置代码是否在gpu运行           |       true        |                \                 |
+|      epoch_num           |    最大训练epoch数             |       500        |                \                 |
 |      log_smooth_window   |    滑动窗口大小            |       20          |                \                 |
 |      print_batch_step    |    设置打印log间隔         |       10          |                \                 |
 |      save_model_dir      |    设置模型保存路径        |  output/{算法名称}  |                \                 |
 |      save_epoch_step     |    设置模型保存间隔        |       3           |                \                 |
 |      eval_batch_step     |    设置模型评估间隔        | 2000 或 [1000, 2000]        | 2000 表示每2000次迭代评估一次，[1000， 2000]表示从1000次迭代开始，每2000次评估一次   |
-|train_batch_size_per_card |  设置训练时单卡batch size    |         256         |                \                 |
-| test_batch_size_per_card |  设置评估时单卡batch size    |         256         |                \                 |
-|      image_shape         |    设置输入图片尺寸        |   [3, 32, 100]    |                \                 |
+|      cal_metric_during_train     |    设置是否在训练过程中评估指标，此时评估的是模型在当前batch下的指标        |       true         |                \                 |
+|      load_static_weights     |   设置预训练模型是否是静态图模式保存(目前仅检测算法需要)        |       true         |                \                 |
+|      pretrained_model    |    设置加载预训练模型路径      |  ./pretrain_models/CRNN/best_accuracy  |  \          |
+|      checkpoints         |    加载模型参数路径            |       None        |    用于中断后加载参数继续训练 |
+|      use_visualdl  |    设置是否启用visualdl进行可视化log展示 |          False        |    [教程地址](https://www.paddlepaddle.org.cn/paddle/visualdl) |
+|      infer_img            |    设置预测图像路径或文件夹路径     |       ./infer_img | \|
+|      character_dict_path |    设置字典路径            |  ./ppocr/utils/ppocr_keys_v1.txt  |    \                 |
 |      max_text_length     |    设置文本最大长度        |       25          |                \                 |
 |      character_type      |    设置字符类型            |       ch          |    en/ch, en时将使用默认dict，ch时使用自定义dict|
-|      character_dict_path |    设置字典路径            |  ./ppocr/utils/ic15_dict.txt  |    \                 |
-|      loss_type           |    设置 loss 类型              |       ctc         |    支持两种loss： ctc / attention |
-|       distort            |    设置是否使用数据增强          |       false       |  设置为true时，将在训练时随机进行扰动，支持的扰动操作可阅读[img_tools.py](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/ppocr/data/rec/img_tools.py)                 |
-|       use_space_char     |    设置是否识别空格             |        false      |          仅在 character_type=ch 时支持空格                 |
+|      use_space_char     |    设置是否识别空格             |        True      |          仅在 character_type=ch 时支持空格                 |
 |      label_list          |    设置方向分类器支持的角度       |    ['0','180']    |     仅在方向分类器中生效 |
-|      average_window      |    ModelAverage优化器中的窗口长度计算比例 |  0.15       |       目前仅应用与SRN |
-|      max_average_window  |    平均值计算窗口长度的最大值   |   15625              | 推荐设置为一轮训练中mini-batchs的数目|
-|      min_average_window  |    平均值计算窗口长度的最小值  |    10000              |      \          |
-|      reader_yml          |    设置reader配置文件          |  ./configs/rec/rec_icdar15_reader.yml  |  \          |
-|      pretrain_weights    |    加载预训练模型路径      |  ./pretrain_models/CRNN/best_accuracy  |  \          |
-|      checkpoints         |    加载模型参数路径            |       None        |    用于中断后加载参数继续训练 |
-|      save_inference_dir  |    inference model 保存路径 |          None        |    用于保存inference model |
+|      save_res_path          |    设置检测模型的结果保存地址       |    ./output/det_db/predicts_db.txt    |     仅在检测模型中生效 |
 
-## 配置文件 Reader 系列参数介绍
+### Optimizer ([ppocr/optimizer](../../ppocr/optimizer))
 
-以 `rec_chinese_reader.yml` 为例
+|         字段             |            用途            |      默认值        |            备注             |
+| :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
+|      name        |         优化器类名          |  Adam  |  目前支持`Momentum`,`Adam`,`RMSProp`, 见[ppocr/optimizer/optimizer.py](../../ppocr/optimizer/optimizer.py)  |
+|      beta1           |    设置一阶矩估计的指数衰减率  |       0.9         |               \             |
+|      beta2           |    设置二阶矩估计的指数衰减率  |     0.999         |               \             |
+|      **lr**                |         设置学习率decay方式       |   -    |       \  |
+|        name    |      学习率decay类名   |         Cosine       | 目前支持`Linear`,`Cosine`,`Step`,`Piecewise`, 见[ppocr/optimizer/learning_rate.py](../../ppocr/optimizer/learning_rate.py) |
+|        learning_rate      |    基础学习率        |       0.001      |  \        |
+|      **regularizer**      |  设置网络正则化方式        |       -      | \        |
+|        name      |    正则化类名      |       L2     | 目前支持`L1`,`L2`, 见[ppocr/optimizer/regularizer.py](../../ppocr/optimizer/regularizer.py)        |
+|        factor      |    学习率衰减系数       |       0.00004     |  \        |
 
-|         字段             |            用途                |      默认值       |            备注            |
-| :----------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
-|      reader_function     |    选择数据读取方式        |  ppocr.data.rec.dataset_traversal,SimpleReader  | 支持SimpleReader / LMDBReader 两种数据读取方式 |
-|      num_workers             |    设置数据读取线程数            |       8        |                \                 |
-|      img_set_dir          |    数据集路径             |       ./train_data        |                \                 |
-|      label_file_path      |    数据标签路径           |       ./train_data/rec_gt_train.txt| \    |
-|      infer_img            |    预测图像文件夹路径     |       ./infer_img | \|
 
-## 配置文件 Optimizer 系列参数介绍
+### Architecture ([ppocr/modeling](../../ppocr/modeling))
+在ppocr中，网络被划分为Transform,Backbone,Neck和Head四个阶段
+
+|         字段             |            用途            |      默认值        |            备注             |
+| :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
+|      model_type        |         网络类型          |  rec  |  目前支持`rec`,`det`,`cls`  |
+|      algorithm           |    模型名称  |       CRNN         |               支持列表见[algorithm_overview](./algorithm_overview.md)             |
+|      **Transform**           |    设置变换方式  |       -       |               目前仅rec类型的算法支持, 具体见[ppocr/modeling/transform](../../ppocr/modeling/transform)              |
+|        name    |      变换方式类名   |         TPS       | 目前支持`TPS` |
+|        num_fiducial      |    TPS控制点数        |       20      |  上下边各十个       |
+|        loc_lr      |    定位网络学习率        |       0.1      |  \      |
+|        model_name      |    定位网络大小        |       small      |  目前支持`small`,`large`       |
+|      **Backbone**      |  设置网络backbone类名        |       -      | 具体见[ppocr/modeling/backbones](../../ppocr/modeling/backbones)        |
+|        name      |    backbone类名       |       ResNet     | 目前支持`MobileNetV3`,`ResNet`        |
+|        layers      |    resnet层数       |       34     |  支持18,34,50,101,152,200       |
+|        model_name      |    MobileNetV3 网络大小       |       small     |  支持`small`,`large`       |
+|      **Neck**      |  设置网络neck        |       -      | 具体见[ppocr/modeling/necks](../../ppocr/modeling/necks)        |
+|        name      |    neck类名       |       SequenceEncoder     | 目前支持`SequenceEncoder`,`DBFPN`        |
+|        encoder_type      |    SequenceEncoder编码器类型       |       rnn     |  支持`reshape`,`fc`,`rnn`       |
+|        hidden_size      |   rnn内部单元数       |       48     |  \      |
+|        out_channels      |   DBFPN输出通道数       |       256     |  \      |
+|      **Head**      |  设置网络Head        |       -      | 具体见[ppocr/modeling/heads](../../ppocr/modeling/heads)        |
+|        name      |    head类名       |       CTCHead     | 目前支持`CTCHead`,`DBHead`,`ClsHead`        |
+|        fc_decay      |    CTCHead正则化系数       |       0.0004     |  \      |
+|        k      |   DBHead二值化系数       |       50     |  \      |
+|        class_dim      |   ClsHead输出分类数       |       2     |  \      |
+
 
-以 `rec_icdar15_train.yml` 为例
+### Loss ([ppocr/losses](../../ppocr/losses))
+
+|         字段             |            用途            |      默认值        |            备注             |
+| :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
+|      name        |         网络loss类名          |  CTCLoss  |  目前支持`CTCLoss`,`DBLoss`,`ClsLoss`  |
+|      balance_loss        |        DBLossloss中是否对正负样本数量进行均衡(使用OHEM)         |  True  |  \  |
+|      ohem_ratio        |        DBLossloss中的OHEM的负正样本比例         |  3  |  \  |
+|      main_loss_type        |        DBLossloss中shrink_map所采用的的loss        |  DiceLoss  |  支持`DiceLoss`,`BCELoss`  |
+|      alpha        |        DBLossloss中shrink_map_loss的系数       |  5  |  \  |
+|      beta        |        DBLossloss中threshold_map_loss的系数       |  10  |  \  |
+
+### PostProcess ([ppocr/postprocess](../../ppocr/postprocess))
+
+|         字段             |            用途            |      默认值        |            备注             |
+| :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
+|      name        |         后处理类名          |  CTCLabelDecode  |  目前支持`CTCLoss`,`AttnLabelDecode`,`DBPostProcess`,`ClsPostProcess`  |
+|      thresh        |        DBPostProcess中分割图进行二值化的阈值         |  0.3  |  \  |
+|      box_thresh        |        DBPostProcess中对输出框进行过滤的阈值，低于此阈值的框不会输出         |  0.7  |  \  |
+|      max_candidates        |        DBPostProcess中输出的最大文本框数量        |  1000  |   |
+|      unclip_ratio        |        DBPostProcess中对文本框进行放大的比例       |  2.0  |  \  |
+
+### Metric ([ppocr/metrics](../../ppocr/metrics))
+
+|         字段             |            用途            |      默认值        |            备注             |
+| :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
+|      name        |         指标评估方法名称          |  CTCLabelDecode  |  目前支持`DetMetric`,`RecMetric`,`ClsMetric`  |
+|      main_indicator        |        主要指标,用于选取最优模型         |  acc |  对于检测方法为hmean，识别和分类方法为acc  |
 
+### Dataset  ([ppocr/data](../../ppocr/data))
 |         字段             |            用途            |      默认值        |            备注             |
 | :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
-|         function        |         选择优化器          |  pocr.optimizer,AdamDecay  |  目前只支持Adam方式  |
-|         base_lr         |      设置初始学习率          |       0.0005      |               \             |
-|         beta1           |    设置一阶矩估计的指数衰减率  |       0.9         |               \             |
-|         beta2           |    设置二阶矩估计的指数衰减率  |     0.999         |               \             |
-|         decay           |         是否使用decay       |    \              |               \             |
-|      function(decay)    |         设置decay方式       |   -    |       目前支持cosine_decay, cosine_decay_warmup与piecewise_decay  |
-|      step_each_epoch    |      每个epoch包含多少次迭代, cosine_decay/cosine_decay_warmup时有效   |         20       | 计算方式：total_image_num / (batch_size_per_card * card_size) |
-|        total_epoch      |    总共迭代多少个epoch, cosine_decay/cosine_decay_warmup时有效        |       1000      | 与Global.epoch_num 一致        |
-|        warmup_minibatch      |  线性warmup的迭代次数, cosine_decay_warmup时有效        |       1000      | \        |
-|        boundaries      |    学习率下降时的迭代次数间隔, piecewise_decay时有效       |       -      | 参数为列表形式        |
-|        decay_rate      |    学习率衰减系数, piecewise_decay时有效       |       -      |  \        |
+|      **dataset**        |         每次迭代返回一个样本          |  -  |  -  |
+|      name        |        dataset类名         |  SimpleDataSet |  目前支持`SimpleDataSet`和`LMDBDateSet`  |
+|      data_dir        |        数据集图片存放路径         |  ./train_data |  \  |
+|      label_file_list        |        数据标签路径         |  ["./train_data/train_list.txt"] | dataset为LMDBDateSet时不需要此参数   |
+|      ratio_list        |        数据集的比例         |  [1.0] | 若label_file_list中有两个train_list，且ratio_list为[0.4,0.6]，则从train_list1中采样40%，从train_list2中采样60%组合整个dataset   |
+|      transforms        |        对图片和标签进行变换的方法列表         |  [DecodeImage,CTCLabelEncode,RecResizeImg,KeepKeys] |   见[ppocr/data/imaug](../../ppocr/data/imaug)  |
+|      **loader**        |        dataloader相关         |  - |   |
+|      shuffle        |        每个epoch是否将数据集顺序打乱         |  True | \  |
+|      batch_size_per_card        |        训练时单卡batch size         |  256 | \  |
+|      drop_last        |        是否丢弃因数据集样本数不能被 batch_size 整除而产生的最后一个不完整的mini-batch        |  True | \  |
+|      num_workers        |        用于加载数据的子进程个数，若为0即为不开启子进程，在主进程中进行数据加载        |  8 | \  |
\ No newline at end of file

doc/doc_ch/whl.md

@@ -261,6 +261,61 @@ im_show.save('result.jpg')
 paddleocr --image_dir PaddleOCR/doc/imgs/11.jpg --det_model_dir {your_det_model_dir} --rec_model_dir {your_rec_model_dir} --rec_char_dict_path {your_rec_char_dict_path} --cls_model_dir {your_cls_model_dir} --use_angle_cls true --cls true
 ```
 
+### 使用网络图片或者numpy数组作为输入
+
+1. 网络图片
+
+代码使用
+```python
+from paddleocr import PaddleOCR, draw_ocr
+# Paddleocr目前支持中英文、英文、法语、德语、韩语、日语，可以通过修改lang参数进行切换
+# 参数依次为`ch`, `en`, `french`, `german`, `korean`, `japan`。
+ocr = PaddleOCR(use_angle_cls=True, lang="ch") # need to run only once to download and load model into memory
+img_path = 'http://n.sinaimg.cn/ent/transform/w630h933/20171222/o111-fypvuqf1838418.jpg'
+result = ocr.ocr(img_path, cls=True)
+for line in result:
+    print(line)
+
+# 显示结果
+from PIL import Image
+image = Image.open(img_path).convert('RGB')
+boxes = [line[0] for line in result]
+txts = [line[1][0] for line in result]
+scores = [line[1][1] for line in result]
+im_show = draw_ocr(image, boxes, txts, scores, font_path='/path/to/PaddleOCR/doc/simfang.ttf')
+im_show = Image.fromarray(im_show)
+im_show.save('result.jpg')
+```
+命令行模式
+```bash
+paddleocr --image_dir http://n.sinaimg.cn/ent/transform/w630h933/20171222/o111-fypvuqf1838418.jpg --use_angle_cls=true
+```
+
+2. numpy数组
+仅通过代码使用时支持numpy数组作为输入
+```python
+from paddleocr import PaddleOCR, draw_ocr
+# Paddleocr目前支持中英文、英文、法语、德语、韩语、日语，可以通过修改lang参数进行切换
+# 参数依次为`ch`, `en`, `french`, `german`, `korean`, `japan`。
+ocr = PaddleOCR(use_angle_cls=True, lang="ch") # need to run only once to download and load model into memory
+img_path = 'PaddleOCR/doc/imgs/11.jpg'
+img = cv2.imread(img_path)
+# img = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY), 如果你自己训练的模型支持灰度图，可以将这句话的注释取消
+result = ocr.ocr(img_path, cls=True)
+for line in result:
+    print(line)
+
+# 显示结果
+from PIL import Image
+image = Image.open(img_path).convert('RGB')
+boxes = [line[0] for line in result]
+txts = [line[1][0] for line in result]
+scores = [line[1][1] for line in result]
+im_show = draw_ocr(image, boxes, txts, scores, font_path='/path/to/PaddleOCR/doc/simfang.ttf')
+im_show = Image.fromarray(im_show)
+im_show.save('result.jpg')
+```
+
 ## 参数说明
 
 | 字段                    | 说明                                                                                                                                                                                                                 | 默认值                  |
@@ -285,6 +340,7 @@ paddleocr --image_dir PaddleOCR/doc/imgs/11.jpg --det_model_dir {your_det_model_
 | max_text_length         | 识别算法能识别的最大文字长度                                                                                                                                                                                         | 25                      |
 | rec_char_dict_path      | 识别模型字典路径，当rec_model_dir使用方式2传参时需要修改为自己的字典路径                                                                                                                                                | ./ppocr/utils/ppocr_keys_v1.txt                        |
 | use_space_char          | 是否识别空格                                                                                                                                                                                                         | TRUE                    |
+| drop_score          | 对输出按照分数(来自于识别模型)进行过滤，低于此分数的不返回                                                                                                                                                                                                         | 0.5                    |
 | use_angle_cls          | 是否加载分类模型                                                                                                                                                                                                         | FALSE                    |
 | cls_model_dir          | 分类模型所在文件夹。传参方式有两种，1. None: 自动下载内置模型到 `~/.paddleocr/cls`；2.自己转换好的inference模型路径，模型路径下必须包含model和params文件                                                                                 | None                    |
 | cls_image_shape          | 分类算法的输入图片尺寸                                                                           | "3, 48, 192"                    |
@@ -295,4 +351,4 @@ paddleocr --image_dir PaddleOCR/doc/imgs/11.jpg --det_model_dir {your_det_model_
 | lang                     | 模型语言类型,目前支持 中文(ch)和英文(en)                                                                                                                                                                                                  | ch                    |
 | det                     | 前向时使用启动检测                                                                                                                                                                                                   | TRUE                    |
 | rec                     | 前向时是否启动识别                                                                                                                                                                                                   | TRUE                    |
-| cls                     | 前向时是否启动分类                                                                                                                                                                                                 | FALSE                    |
+| cls                     | 前向时是否启动分类 (命令行模式下使用use_angle_cls控制前向是否启动分类)                                                                                                                                                                                                | FALSE                    |

doc/doc_en/add_new_algorithm_en.md

+# Add new algorithm
+
+PaddleOCR decomposes an algorithm into the following parts, and modularizes each part to make it more convenient to develop new algorithms.
+
+* Data loading and processing
+* Network
+* Post-processing
+* Loss
+* Metric
+* Optimizer
+
+The following will introduce each part separately, and introduce how to add the modules required for the new algorithm.
+
+
+## Data loading and processing
+
+Data loading and processing are composed of different modules, which complete the image reading, data augment and label production. This part is under [ppocr/data](../../ppocr/data). The explanation of each file and folder are as follows:
+
+```bash
+ppocr/data/
+├── imaug             # Scripts for image reading, data augment and label production
+│   ├── label_ops.py  # Modules that transform the label
+│   ├── operators.py  # Modules that transform the image
+│   ├──.....
+├── __init__.py
+├── lmdb_dataset.py   # The dataset that reads the lmdb
+└── simple_dataset.py # Read the dataset saved in the form of `image_path\tgt`
+```
+
+PaddleOCR has a large number of built-in image operation related modules. For modules that are not built-in, you can add them through the following steps:
+
+1. Create a new file under the [ppocr/data/imaug](../../ppocr/data/imaug) folder, such as my_module.py.
+2. Add code in the my_module.py file, the sample code is as follows:
+
+```python
+class MyModule:
+    def __init__(self, *args, **kwargs):
+        # your init code
+        pass
+
+    def __call__(self, data):
+        img = data['image']
+        label = data['label']
+        # your process code
+
+        data['image'] = img
+        data['label'] = label
+        return data
+```
+
+3. Import the added module in the [ppocr/data/imaug/\__init\__.py](../../ppocr/data/imaug/__init__.py) file.
+
+All different modules of data processing are executed by sequence, combined and executed in the form of a list in the config file. Such as:
+
+```yaml
+# angle class data process
+transforms:
+  - DecodeImage: # load image
+      img_mode: BGR
+      channel_first: False
+  - MyModule:
+      args1: args1
+      args2: args2
+  - KeepKeys:
+      keep_keys: [ 'image', 'label' ] # dataloader will return list in this order
+```
+
+## Network
+
+The network part completes the construction of the network, and PaddleOCR divides the network into four parts, which are under [ppocr/modeling](../../ppocr/modeling). The data entering the network will pass through these four parts in sequence(transforms->backbones->
+necks->heads).
+
+```bash
+├── architectures # Code for building network
+├── transforms    # Image Transformation Module
+├── backbones     # Feature extraction module
+├── necks         # Feature enhancement module
+└── heads         # Output module
+```
+
+PaddleOCR has built-in commonly used modules related to algorithms such as DB, EAST, SAST, CRNN and Attention. For modules that do not have built-in, you can add them through the following steps, the four parts are added in the same steps, take backbones as an example:
+
+1. Create a new file under the [ppocr/modeling/backbones](../../ppocr/modeling/backbones) folder, such as my_backbone.py.
+2. Add code in the my_backbone.py file, the sample code is as follows:
+
+```python
+import paddle
+import paddle.nn as nn
+import paddle.nn.functional as F
+
+
+class MyBackbone(nn.Layer):
+    def __init__(self, *args, **kwargs):
+        super(MyBackbone, self).__init__()
+        # your init code
+        self.conv = nn.xxxx
+
+    def forward(self, inputs):
+        # your necwork forward
+        y = self.conv(inputs)
+        return y
+```
+
+3. Import the added module in the [ppocr/modeling/backbones/\__init\__.py](../../ppocr/modeling/backbones/__init__.py) file.
+
+After adding the four-part modules of the network, you only need to configure them in the configuration file to use, such as:
+
+```yaml
+Architecture:
+  model_type: rec
+  algorithm: CRNN
+  Transform:
+    name: MyTransform
+    args1: args1
+    args2: args2
+  Backbone:
+    name: MyBackbone
+    args1: args1
+  Neck:
+    name: MyNeck
+    args1: args1
+  Head:
+    name: MyHead
+    args1: args1
+```
+
+## Post-processing
+
+Post-processing realizes decoding network output to obtain text box or recognized text. This part is under [ppocr/postprocess](../../ppocr/postprocess).
+PaddleOCR has built-in post-processing modules related to algorithms such as DB, EAST, SAST, CRNN and Attention. For components that are not built-in, they can be added through the following steps:
+
+1. Create a new file under the [ppocr/postprocess](../../ppocr/postprocess) folder, such as my_postprocess.py.
+2. Add code in the my_postprocess.py file, the sample code is as follows:
+
+```python
+import paddle
+
+
+class MyPostProcess:
+    def __init__(self, *args, **kwargs):
+        # your init code
+        pass
+
+    def __call__(self, preds, label=None, *args, **kwargs):
+        if isinstance(preds, paddle.Tensor):
+            preds = preds.numpy()
+        # you preds decode code
+        preds = self.decode_preds(preds)
+        if label is None:
+            return preds
+        # you label decode code
+        label = self.decode_label(label)
+        return preds, label
+
+    def decode_preds(self, preds):
+        # you preds decode code
+        pass
+
+    def decode_label(self, preds):
+        # you label decode code
+        pass
+```
+
+3. Import the added module in the [ppocr/postprocess/\__init\__.py](../../ppocr/postprocess/__init__.py) file.
+
+After the post-processing module is added, you only need to configure it in the configuration file to use, such as:
+
+```yaml
+PostProcess:
+  name: MyPostProcess
+  args1: args1
+  args2: args2
+```
+
+## Loss
+
+The loss function is used to calculate the distance between the network output and the label. This part is under [ppocr/losses](../../ppocr/losses).
+PaddleOCR has built-in loss function modules related to algorithms such as DB, EAST, SAST, CRNN and Attention. For modules that do not have built-in modules, you can add them through the following steps:
+
+1. Create a new file in the [ppocr/losses](../../ppocr/losses) folder, such as my_loss.py.
+2. Add code in the my_loss.py file, the sample code is as follows:
+
+```python
+import paddle
+from paddle import nn
+
+
+class MyLoss(nn.Layer):
+    def __init__(self, **kwargs):
+        super(MyLoss, self).__init__()
+        # you init code
+        pass
+
+    def __call__(self, predicts, batch):
+        label = batch[1]
+        # your loss code
+        loss = self.loss(input=predicts, label=label)
+        return {'loss': loss}
+```
+
+3. Import the added module in the [ppocr/losses/\__init\__.py](../../ppocr/losses/__init__.py) file.
+
+After the loss function module is added, you only need to configure it in the configuration file to use it, such as:
+
+```yaml
+Loss:
+  name: MyLoss
+  args1: args1
+  args2: args2
+```
+
+## Metric
+
+Metric is used to calculate the performance of the network on the current batch. This part is under [ppocr/metrics](../../ppocr/metrics). PaddleOCR has built-in evaluation modules related to algorithms such as detection, classification and recognition. For modules that do not have built-in modules, you can add them through the following steps:
+
+1. Create a new file under the [ppocr/metrics](../../ppocr/metrics) folder, such as my_metric.py.
+2. Add code in the my_metric.py file, the sample code is as follows:
+
+```python
+
+class MyMetric(object):
+    def __init__(self, main_indicator='acc', **kwargs):
+        # main_indicator is used for select best model
+        self.main_indicator = main_indicator
+        self.reset()
+
+    def __call__(self, preds, batch, *args, **kwargs):
+        # preds is out of postprocess
+        # batch is out of dataloader
+        labels = batch[1]
+        cur_correct_num = 0
+        cur_all_num = 0
+        # you metric code
+        self.correct_num += cur_correct_num
+        self.all_num += cur_all_num
+        return {'acc': cur_correct_num / cur_all_num, }
+
+    def get_metric(self):
+        """
+        return metircs {
+                 'acc': 0,
+                 'norm_edit_dis': 0,
+            }
+        """
+        acc = self.correct_num / self.all_num
+        self.reset()
+        return {'acc': acc}
+
+    def reset(self):
+        # reset metric
+        self.correct_num = 0
+        self.all_num = 0
+
+```
+
+3. Import the added module in the [ppocr/metrics/\__init\__.py](../../ppocr/metrics/__init__.py) file.
+
+After the metric module is added, you only need to configure it in the configuration file to use it, such as:
+
+```yaml
+Metric:
+  name: MyMetric
+  main_indicator: acc
+```
+
+## 优化器
+
+The optimizer is used to train the network. The optimizer also contains network regularization and learning rate decay modules. This part is under [ppocr/optimizer](../../ppocr/optimizer). PaddleOCR has built-in
+Commonly used optimizer modules such as `Momentum`, `Adam` and `RMSProp`, common regularization modules such as `Linear`, `Cosine`, `Step` and `Piecewise`, and common learning rate decay modules such as `L1Decay` and `L2Decay`.
+Modules without built-in can be added through the following steps, take `optimizer` as an example:
+
+1. Create your own optimizer in the [ppocr/optimizer/optimizer.py](../../ppocr/optimizer/optimizer.py) file, the sample code is as follows:
+
+```python
+from paddle import optimizer as optim
+
+
+class MyOptim(object):
+    def __init__(self, learning_rate=0.001, *args, **kwargs):
+        self.learning_rate = learning_rate
+
+    def __call__(self, parameters):
+        # It is recommended to wrap the built-in optimizer of paddle
+        opt = optim.XXX(
+            learning_rate=self.learning_rate,
+            parameters=parameters)
+        return opt
+
+```
+
+After the optimizer module is added, you only need to configure it in the configuration file to use, such as:
+
+```yaml
+Optimizer:
+  name: MyOptim
+  args1: args1
+  args2: args2
+  lr:
+    name: Cosine
+    learning_rate: 0.001
+  regularizer:
+    name: 'L2'
+    factor: 0
+```
\ No newline at end of file

doc/doc_en/angle_class_en.md

@@ -65,26 +65,35 @@ Start training:
 ```
 # Set PYTHONPATH path
 export PYTHONPATH=$PYTHONPATH:.
-# GPU training Support single card and multi-card training, specify the card number through CUDA_VISIBLE_DEVICES
-export CUDA_VISIBLE_DEVICES=0,1,2,3
-# Training icdar15 English data
-python3 tools/train.py -c configs/cls/cls_mv3.yml
+# GPU training Support single card and multi-card training, specify the card number through selected_gpus
+# Start training, the following command has been written into the train.sh file, just modify the configuration file path in the file
+python3 -m paddle.distributed.launch --selected_gpus '0,1,2,3,4,5,6,7'  tools/train.py -c configs/cls/cls_mv3.yml
 ```
 
 - Data Augmentation
 
-PaddleOCR provides a variety of data augmentation methods. If you want to add disturbance during training, please set `distort: true` in the configuration file.
+PaddleOCR provides a variety of data augmentation methods. If you want to add disturbance during training, Please uncomment the `RecAug` and `RandAugment` fields under `Train.dataset.transforms` in the configuration file.
 
 The default perturbation methods are: cvtColor, blur, jitter, Gasuss noise, random crop, perspective, color reverse, RandAugment.
 
 Except for RandAugment, each disturbance method is selected with a 50% probability during the training process. For specific code implementation, please refer to:
-[randaugment.py](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/ppocr/data/cls/randaugment.py)
-[img_tools.py](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/ppocr/data/rec/img_tools.py)
+[rec_img_aug.py](../../ppocr/data/imaug/rec_img_aug.py) 
+[randaugment.py](../../ppocr/data/imaug/randaugment.py)
 
 
 - Training
 
-PaddleOCR supports alternating training and evaluation. You can modify `eval_batch_step` in `configs/cls/cls_mv3.yml` to set the evaluation frequency. By default, it is evaluated every 500 iter and the best acc model is saved under `output/cls_mv3/best_accuracy` during the evaluation process.
+PaddleOCR supports alternating training and evaluation. You can modify `eval_batch_step` in `configs/cls/cls_mv3.yml` to set the evaluation frequency. By default, it is evaluated every 1000 iter. The following content will be saved during training:
+```bash
+├── best_accuracy.pdopt # Optimizer parameters for the best model
+├── best_accuracy.pdparams # Parameters of the best model
+├── best_accuracy.states # Metric info and epochs of the best model
+├── config.yml # Configuration file for this experiment
+├── latest.pdopt # Optimizer parameters for the latest model
+├── latest.pdparams # Parameters of the latest model
+├── latest.states # Metric info and epochs of the latest model
+└── train.log # Training log
+```
 
 If the evaluation set is large, the test will be time-consuming. It is recommended to reduce the number of evaluations, or evaluate after training.
 
@@ -92,7 +101,7 @@ If the evaluation set is large, the test will be time-consuming. It is recommend
 
 ### EVALUATION
 
-The evaluation data set can be modified via `configs/cls/cls_reader.yml` setting of `label_file_path` in EvalReader.
+The evaluation dataset can be set by modifying the `Eval.dataset.label_file_list` field in the `configs/cls/cls_mv3.yml` file.
 
 ```
 export CUDA_VISIBLE_DEVICES=0
@@ -106,21 +115,20 @@ python3 tools/eval.py -c configs/cls/cls_mv3.yml -o Global.checkpoints={path/to/
 
 Using the model trained by paddleocr, you can quickly get prediction through the following script.
 
-The default prediction picture is stored in `infer_img`, and the weight is specified via `-o Global.checkpoints`:
+Use `Global.infer_img` to specify the path of the predicted picture or folder, and use `Global.checkpoints` to specify the weight:
 
 ```
 # Predict English results
-python3 tools/infer_rec.py -c configs/cls/cls_mv3.yml -o Global.checkpoints={path/to/weights}/best_accuracy TestReader.infer_img=doc/imgs_words/en/word_1.jpg
+python3 tools/infer_cls.py -c configs/cls/cls_mv3.yml -o Global.checkpoints={path/to/weights}/best_accuracy Global.infer_img=doc/imgs_words_en/word_10.png
 ```
 
 Input image:
 
-![](../imgs_words/en/word_1.png)
+![](../imgs_words_en/word_10.png)
 
 Get the prediction result of the input image:
 
 ```
-infer_img: doc/imgs_words/en/word_1.png
-    scores: [[0.93161047 0.06838956]]
-    label: [0]
+infer_img: doc/imgs_words_en/word_10.png
+     result: ('0', 0.9999995)
 ```

doc/doc_en/config_en.md

-# OPTIONAL PARAMETERS LIST
+## Optional parameter list
 
-The following list can be viewed via `--help`
+The following list can be viewed through `--help`
 
 |         FLAG             |     Supported script    |        Use        |      Defaults       |         Note         |
 | :----------------------: | :------------: | :---------------: | :--------------: | :-----------------: |
-|          -c              |      ALL       |  Specify configuration file to use |  None  |  **Please refer to the parameter introduction for configuration file usage** |
-|          -o              |      ALL       |  set configuration options  |  None  |  Configuration using -o has higher priority than the configuration file selected with -c. E.g: `-o Global.use_gpu=false`  |  
-
+|          -c              |      ALL       |  Specify configuration file to use  |  None  |  **Please refer to the parameter introduction for configuration file usage** |
+|          -o              |      ALL       |  set configuration options  |  None  |  Configuration using -o has higher priority than the configuration file selected with -c. E.g: -o Global.use_gpu=false |
 
 ## INTRODUCTION TO GLOBAL PARAMETERS OF CONFIGURATION FILE
 
-Take `rec_chinese_lite_train_v1.1.yml` as an example
-
+Take rec_chinese_lite_train_v1.1.yml as an example
+### Global 
 
-|         Parameter             |            Use                |      Default       |            Note            |
+|         Parameter             |            Use                |      Defaults       |            Note            |
 | :----------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
-|      algorithm           |    Select algorithm to use                    |  Synchronize with configuration file   |     For selecting model, please refer to the supported model [list](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/README_en.md) |
-|      use_gpu             |    Set using GPU or not            |       true        |                \                 |
-|      epoch_num           |    Maximum training epoch number             |       3000        |                \                 |
+|      use_gpu             |    Set using GPU or not           |       true        |                \                 |
+|      epoch_num           |    Maximum training epoch number             |       500        |                \                 |
 |      log_smooth_window   |    Sliding window size            |       20          |                \                 |
 |      print_batch_step    |    Set print log interval         |       10          |                \                 |
-|      save_model_dir      |    Set model save path        |  output/{model_name}  |                \                 |
+|      save_model_dir      |    Set model save path        |  output/{算法名称}  |                \                 |
 |      save_epoch_step     |    Set model save interval        |       3           |                \                 |
-|      eval_batch_step     |    Set the model evaluation interval        |2000 or [1000, 2000] |runing evaluation every 2000 iters or evaluation is run every 2000 iterations after the 1000th iteration  |
-|train_batch_size_per_card |  Set the batch size during training   |         256         |                \                 |
-| test_batch_size_per_card |  Set the batch size during testing    |         256         |                \                 |
-|      image_shape         |    Set input image size        |   [3, 32, 100]    |                \                 |
-|      max_text_length     |    Set the maximum text length        |       25          |                \                 |
-|      character_type      |    Set character type            |       ch          |    en/ch, the default dict will be used for en, and the custom dict will be used for ch|
-|      character_dict_path |    Set dictionary path            |  ./ppocr/utils/ic15_dict.txt  |    \                 |
-|      loss_type           |    Set loss type              |       ctc         |    Supports two types of loss: ctc / attention |
-|       distort            |    Set use distort          |       false       |  Support distort type ,read [img_tools.py](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/ppocr/data/rec/img_tools.py)                 |
-|      use_space_char          |    Wether to recognize space             |        false      |         Only support in character_type=ch mode                 |
-     label_list          | Set the angle supported by the direction classifier | ['0','180'] | Only valid in the direction classifier |
-|      reader_yml          |    Set the reader configuration file          |  ./configs/rec/rec_icdar15_reader.yml  |  \          |
-|      pretrain_weights    |    Load pre-trained model path      |  ./pretrain_models/CRNN/best_accuracy  |  \          |
-|      checkpoints         |    Load saved model path            |       None        |    Used to load saved parameters to continue training after interruption |
-|      save_inference_dir  |   path to save model for inference |          None        |   Use to save inference model |
-
-## INTRODUCTION TO READER PARAMETERS OF CONFIGURATION FILE
-
-Take `rec_chinese_reader.yml` as an example:
-
-|         Parameter             |            Use                |      Default       |            Note            |
-| :----------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
-|      reader_function     |    Select data reading method        |  ppocr.data.rec.dataset_traversal,SimpleReader  | Support two data reading methods: SimpleReader / LMDBReader  |
-|      num_workers             |    Set the number of data reading threads            |       8        |                \                 |
-|      img_set_dir          |    Image folder path             |       ./train_data        |                \                 |
-|      label_file_path      |    Groundtruth file path           |       ./train_data/rec_gt_train.txt| \    |
-|      infer_img            |    Result folder path     |       ./infer_img | \|
+|      eval_batch_step     |    Set the model evaluation interval        | 2000 or [1000, 2000]        | runing evaluation every 2000 iters or evaluation is run every 2000 iterations after the 1000th iteration   |
+|      cal_metric_during_train     |    Set whether to evaluate the metric during the training process. At this time, the metric of the model under the current batch is evaluated        |       true         |                \                 |
+|      load_static_weights     |   Set whether the pre-training model is saved in static graph mode (currently only required by the detection algorithm)        |       true         |                \                 |
+|      pretrained_model    |    Set the path of the pre-trained model      |  ./pretrain_models/CRNN/best_accuracy  |  \          |
+|      checkpoints         |    set model parameter path            |       None        |   Used to load parameters after interruption to continue training|
+|      use_visualdl  |    Set whether to enable visualdl for visual log display |          False        |    [Tutorial](https://www.paddlepaddle.org.cn/paddle/visualdl) |
+|      infer_img            |    Set inference image path or folder path     |       ./infer_img | \|
+|      character_dict_path |    Set dictionary path            |  ./ppocr/utils/ppocr_keys_v1.txt  |    \                 |
+|      max_text_length     |    Set the maximum length of text        |       25          |                \                 |
+|      character_type      |    Set character type            |       ch          |    en/ch, the default dict will be used for en, and the custom dict will be used for ch |
+|      use_space_char     |    Set whether to recognize spaces             |        True      |          Only support in character_type=ch mode                 |
+|      label_list          |    Set the angle supported by the direction classifier       |    ['0','180']    |     Only valid in angle classifier model |
+|      save_res_path          |    Set the save address of the test model results       |    ./output/det_db/predicts_db.txt    |     Only valid in the text detection model |
+
+### Optimizer ([ppocr/optimizer](../../ppocr/optimizer))
+
+|         Parameter             |            Use            |      Defaults        |            Note             |
+| :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
+|      name        |         Optimizer class name          |  Adam  |  Currently supports`Momentum`,`Adam`,`RMSProp`, see [ppocr/optimizer/optimizer.py](../../ppocr/optimizer/optimizer.py)  |
+|      beta1           |    Set the exponential decay rate for the 1st moment estimates  |       0.9         |               \             |
+|      beta2           |    Set the exponential decay rate for the 2nd moment estimates  |     0.999         |               \             |
+|      **lr**                |         Set the learning rate decay method       |   -    |       \  |
+|        name    |      Learning rate decay class name   |         Cosine       | Currently supports`Linear`,`Cosine`,`Step`,`Piecewise`, see[ppocr/optimizer/learning_rate.py](../../ppocr/optimizer/learning_rate.py) |
+|        learning_rate      |    Set the base learning rate        |       0.001      |  \        |
+|      **regularizer**      |  Set network regularization method        |       -      | \        |
+|        name      |    Regularizer class name      |       L2     |  Currently support`L1`,`L2`, see[ppocr/optimizer/regularizer.py](../../ppocr/optimizer/regularizer.py)        |
+|        factor      |    Learning rate decay coefficient       |       0.00004     |  \        |
+
+
+### Architecture ([ppocr/modeling](../../ppocr/modeling))
+In ppocr, the network is divided into four stages: Transform, Backbone, Neck and Head
+
+|         Parameter             |            Use            |      Defaults        |            Note             |
+| :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
+|      model_type        |         Network Type          |  rec  |  Currently support`rec`,`det`,`cls`  |
+|      algorithm           |    Model name  |       CRNN         |               See [algorithm_overview](./algorithm_overview.md) for the support list             |
+|      **Transform**           |    Set the transformation method  |       -       |               Currently only recognition algorithms are supported, see [ppocr/modeling/transform](../../ppocr/modeling/transform) for details            |
+|        name    |      Transformation class name   |         TPS       | Currently supports `TPS` |
+|        num_fiducial      |   Number of TPS control points        |       20      |  Ten on the top and bottom       |
+|        loc_lr      |    Localization network learning rate        |       0.1      |  \      |
+|        model_name      |    Localization network size        |       small      |  Currently support`small`,`large`       |
+|      **Backbone**      |  Set the network backbone class name        |       -      | see [ppocr/modeling/backbones](../../ppocr/modeling/backbones)        |
+|        name      |    backbone class name       |       ResNet     | Currently support`MobileNetV3`,`ResNet`        |
+|        layers      |    resnet layers       |       34     |  Currently support18,34,50,101,152,200       |
+|        model_name      |    MobileNetV3 network size       |       small     |  Currently support`small`,`large`       |
+|      **Neck**      |  Set network neck        |       -      | see[ppocr/modeling/necks](../../ppocr/modeling/necks)        |
+|        name      |    neck class name       |       SequenceEncoder     | Currently support`SequenceEncoder`,`DBFPN`        |
+|        encoder_type      |    SequenceEncoder encoder type       |       rnn     |  Currently support`reshape`,`fc`,`rnn`       |
+|        hidden_size      |   rnn number of internal units       |       48     |  \      |
+|        out_channels      |   Number of DBFPN output channels       |       256     |  \      |
+|      **Head**      |  Set the network head        |       -      | see[ppocr/modeling/heads](../../ppocr/modeling/heads)        |
+|        name      |    head class name       |       CTCHead     | Currently support`CTCHead`,`DBHead`,`ClsHead`        |
+|        fc_decay      |    CTCHead regularization coefficient       |       0.0004     |  \      |
+|        k      |   DBHead binarization coefficient       |       50     |  \      |
+|        class_dim      |   ClsHead output category number       |       2     |  \      |
+
+
+### Loss ([ppocr/losses](../../ppocr/losses))
+
+|         Parameter             |            Use            |      Defaults        |            Note             |
+| :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
+|      name        |         loss class name          |  CTCLoss  |  Currently support`CTCLoss`,`DBLoss`,`ClsLoss`  |
+|      balance_loss        |        Whether to balance the number of positive and negative samples in DBLossloss (using OHEM)         |  True  |  \  |
+|      ohem_ratio        |        The negative and positive sample ratio of OHEM in DBLossloss         |  3  |  \  |
+|      main_loss_type        |        The loss used by shrink_map in DBLossloss        |  DiceLoss  |  Currently support`DiceLoss`,`BCELoss`  |
+|      alpha        |        The coefficient of shrink_map_loss in DBLossloss       |  5  |  \  |
+|      beta        |        The coefficient of threshold_map_loss in DBLossloss       |  10  |  \  |
 
-## INTRODUCTION TO OPTIMIZER PARAMETERS OF CONFIGURATION FILE
+### PostProcess ([ppocr/postprocess](../../ppocr/postprocess))
 
-Take `rec_icdar15_train.yml` as an example:
+|         Parameter             |            Use            |      Defaults        |            Note             |
+| :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
+|      name        |         Post-processing class name          |  CTCLabelDecode  |  Currently support`CTCLoss`,`AttnLabelDecode`,`DBPostProcess`,`ClsPostProcess`  |
+|      thresh        |        The threshold for binarization of the segmentation map in DBPostProcess         |  0.3  |  \  |
+|      box_thresh        |        The threshold for filtering output boxes in DBPostProcess. Boxes below this threshold will not be output         |  0.7  |  \  |
+|      max_candidates        |        The maximum number of text boxes output in DBPostProcess        |  1000  |   |
+|      unclip_ratio        |        The unclip ratio of the text box in DBPostProcess       |  2.0  |  \  |
+
+### Metric ([ppocr/metrics](../../ppocr/metrics))
+
+|         Parameter             |            Use            |      Defaults        |            Note             |
+| :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
+|      name        |         Metric method name          |  CTCLabelDecode  |  Currently support`DetMetric`,`RecMetric`,`ClsMetric`  |
+|      main_indicator        |        Main indicators, used to select the best model        |  acc |  For the detection method is hmean, the recognition and classification method is acc  |
 
-|         Parameter             |            Use          |      Default        |            None             |
+### Dataset  ([ppocr/data](../../ppocr/data))
+|         Parameter             |            Use            |      Defaults        |            Note             |
 | :---------------------: |  :---------------------:   | :--------------:  |   :--------------------:   |
-|         function        |         Select Optimizer function          |  pocr.optimizer,AdamDecay  |  Only support Adam  |
-|         base_lr         |      Set the base lr          |       0.0005      |               \             |
-|         beta1           |    Set the exponential decay rate for the 1st moment estimates  |       0.9         |               \             |
-|         beta2           |    Set the exponential decay rate for the 2nd moment estimates  |     0.999         |               \             |
-|         decay           |         Whether to use decay       |    \              |               \             |
-|      function(decay)    |         Set the decay function       |   cosine_decay    |         Support cosine_decay, cosine_decay_warmup and piecewise_decay            |
-|      step_each_epoch    |      The number of steps in an epoch. Used in cosine_decay/cosine_decay_warmup  |         20       | Calculation: total_image_num / (batch_size_per_card * card_size) |
-|        total_epoch      |    The number of epochs. Used in cosine_decay/cosine_decay_warmup      |       1000      | Consistent with Global.epoch_num      |
-|        warmup_minibatch      |  Number of steps for linear warmup. Used in cosine_decay_warmup        |       1000      | \        |
-|        boundaries      |    The step intervals to reduce learning rate. Used in piecewise_decay       |       -      |  The format is list        |
-|        decay_rate      |    Learning rate decay rate. Used in piecewise_decay       |       -      |  \        |
+|      **dataset**        |         Return one sample per iteration          |  -  |  -  |
+|      name        |        dataset class name         |  SimpleDataSet |   Currently support`SimpleDataSet`,`LMDBDateSet`  |
+|      data_dir        |        Image folder path        |  ./train_data |  \  |
+|      label_file_list        |        Groundtruth file path         |  ["./train_data/train_list.txt"] | This parameter is not required when dataset is LMDBDateSet   |
+|      ratio_list        |        Ratio of data set         |  [1.0] | If there are two train_lists in label_file_list and ratio_list is [0.4,0.6], 40% will be sampled from train_list1, and 60% will be sampled from train_list2 to combine the entire dataset   |
+|      transforms        |        List of methods to transform images and labels         |  [DecodeImage,CTCLabelEncode,RecResizeImg,KeepKeys] |   see[ppocr/data/imaug](../../ppocr/data/imaug)  |
+|      **loader**        |        dataloader related         |  - |   |
+|      shuffle        |        Does each epoch disrupt the order of the data set         |  True | \  |
+|      batch_size_per_card        |        Single card batch size during training         |  256 | \  |
+|      drop_last        |        Whether to discard the last incomplete mini-batch because the number of samples in the data set cannot be divisible by batch_size        |  True | \  |
+|      num_workers        |        The number of sub-processes used to load data, if it is 0, the sub-process is not started, and the data is loaded in the main process       |  8 | \  |
\ No newline at end of file

doc/doc_en/whl_en.md

@@ -271,6 +271,59 @@ im_show.save('result.jpg')
 paddleocr --image_dir PaddleOCR/doc/imgs/11.jpg --det_model_dir {your_det_model_dir} --rec_model_dir {your_rec_model_dir} --rec_char_dict_path {your_rec_char_dict_path} --cls_model_dir {your_cls_model_dir} --use_angle_cls true --cls true
 ```
 
+### Use web images or numpy array as input
+
+1. Web image
+
+Use by code
+```python
+from paddleocr import PaddleOCR, draw_ocr
+ocr = PaddleOCR(use_angle_cls=True, lang="ch") # need to run only once to download and load model into memory
+img_path = 'http://n.sinaimg.cn/ent/transform/w630h933/20171222/o111-fypvuqf1838418.jpg'
+result = ocr.ocr(img_path, cls=True)
+for line in result:
+    print(line)
+
+# show result
+from PIL import Image
+image = Image.open(img_path).convert('RGB')
+boxes = [line[0] for line in result]
+txts = [line[1][0] for line in result]
+scores = [line[1][1] for line in result]
+im_show = draw_ocr(image, boxes, txts, scores, font_path='/path/to/PaddleOCR/doc/simfang.ttf')
+im_show = Image.fromarray(im_show)
+im_show.save('result.jpg')
+```
+Use by command line
+```bash
+paddleocr --image_dir http://n.sinaimg.cn/ent/transform/w630h933/20171222/o111-fypvuqf1838418.jpg --use_angle_cls=true
+```
+
+2. Numpy array
+Support numpy array as input only when used by code
+
+```python
+from paddleocr import PaddleOCR, draw_ocr
+ocr = PaddleOCR(use_angle_cls=True, lang="ch") # need to run only once to download and load model into memory
+img_path = 'PaddleOCR/doc/imgs/11.jpg'
+img = cv2.imread(img_path)
+# img = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY), If your own training model supports grayscale images, you can uncomment this line
+result = ocr.ocr(img_path, cls=True)
+for line in result:
+    print(line)
+
+# show result
+from PIL import Image
+image = Image.open(img_path).convert('RGB')
+boxes = [line[0] for line in result]
+txts = [line[1][0] for line in result]
+scores = [line[1][1] for line in result]
+im_show = draw_ocr(image, boxes, txts, scores, font_path='/path/to/PaddleOCR/doc/simfang.ttf')
+im_show = Image.fromarray(im_show)
+im_show.save('result.jpg')
+```
+
+
 ## Parameter Description
 
 | Parameter                    | Description                                                                                                                                                                                                                 | Default value                  |
@@ -295,6 +348,7 @@ paddleocr --image_dir PaddleOCR/doc/imgs/11.jpg --det_model_dir {your_det_model_
 | max_text_length         | The maximum text length that the recognition algorithm can recognize                                                                                                                                                                                         | 25                      |
 | rec_char_dict_path      | the alphabet path which needs to be modified to your own path when `rec_model_Name` use mode 2                                                                                                                                              | ./ppocr/utils/ppocr_keys_v1.txt                        |
 | use_space_char          | Whether to recognize spaces                                                                                                                                                                                                         | TRUE                    |
+| drop_score          | Filter the output by score (from the recognition model), and those below this score will not be returned                                                                                                                                                                                                        | 0.5                    |
 | use_angle_cls          | Whether to load classification model                                                                                                                                                                                                       | FALSE                    |
 | cls_model_dir           | the classification inference model folder. There are two ways to transfer parameters, 1. None: Automatically download the built-in model to `~/.paddleocr/cls`; 2. The path of the inference model converted by yourself, the model and params files must be included in the model path | None |
 | cls_image_shape         | image shape of classification algorithm                                                                                                                                                                                            | "3,48,192"              |
@@ -305,4 +359,4 @@ paddleocr --image_dir PaddleOCR/doc/imgs/11.jpg --det_model_dir {your_det_model_
 | lang                     | The support language, now only Chinese(ch)、English(en)、French(french)、German(german)、Korean(korean)、Japanese(japan) are supported                                                                                                                                                                                                  | ch                    |
 | det                     | Enable detction when `ppocr.ocr` func exec                                                                                                                                                                                                   | TRUE                    |
 | rec                     | Enable recognition when `ppocr.ocr` func exec                                                                                                                                                                                                   | TRUE                    |
-| cls                     | Enable classification when `ppocr.ocr` func exec                                                                                                                                                                                                   | FALSE                    |
+| cls                     | Enable classification when `ppocr.ocr` func exec((Use use_angle_cls in command line mode to control whether to start classification in the forward direction)                                                                                                                                                                                                   | FALSE                    |

paddleocr.py

@@ -26,17 +26,50 @@ import requests
 from tqdm import tqdm
 
 from tools.infer import predict_system
-from ppocr.utils.utility import initial_logger
+from ppocr.utils.logging import get_logger
 
-logger = initial_logger()
+logger = get_logger()
 from ppocr.utils.utility import check_and_read_gif, get_image_file_list
 
 __all__ = ['PaddleOCR']
 
-model_params = {
-    'det': 'https://paddleocr.bj.bcebos.com/ch_models/ch_det_mv3_db_infer.tar',
-    'rec':
-    'https://paddleocr.bj.bcebos.com/ch_models/ch_rec_mv3_crnn_enhance_infer.tar',
+model_urls = {
+    'det':
+        'https://paddleocr.bj.bcebos.com/20-09-22/mobile/det/ch_ppocr_mobile_v1.1_det_infer.tar',
+    'rec': {
+        'ch': {
+            'url':
+                'https://paddleocr.bj.bcebos.com/20-09-22/mobile/rec/ch_ppocr_mobile_v1.1_rec_infer.tar',
+            'dict_path': './ppocr/utils/ppocr_keys_v1.txt'
+        },
+        'en': {
+            'url':
+                'https://paddleocr.bj.bcebos.com/20-09-22/mobile/en/en_ppocr_mobile_v1.1_rec_infer.tar',
+            'dict_path': './ppocr/utils/ic15_dict.txt'
+        },
+        'french': {
+            'url':
+                'https://paddleocr.bj.bcebos.com/20-09-22/mobile/fr/french_ppocr_mobile_v1.1_rec_infer.tar',
+            'dict_path': './ppocr/utils/dict/french_dict.txt'
+        },
+        'german': {
+            'url':
+                'https://paddleocr.bj.bcebos.com/20-09-22/mobile/ge/german_ppocr_mobile_v1.1_rec_infer.tar',
+            'dict_path': './ppocr/utils/dict/german_dict.txt'
+        },
+        'korean': {
+            'url':
+                'https://paddleocr.bj.bcebos.com/20-09-22/mobile/kr/korean_ppocr_mobile_v1.1_rec_infer.tar',
+            'dict_path': './ppocr/utils/dict/korean_dict.txt'
+        },
+        'japan': {
+            'url':
+                'https://paddleocr.bj.bcebos.com/20-09-22/mobile/jp/japan_ppocr_mobile_v1.1_rec_infer.tar',
+            'dict_path': './ppocr/utils/dict/japan_dict.txt'
+        }
+    },
+    'cls':
+        'https://paddleocr.bj.bcebos.com/20-09-22/cls/ch_ppocr_mobile_v1.1_cls_infer.tar'
 }
 
 SUPPORT_DET_MODEL = ['DB']
@@ -54,8 +87,8 @@ def download_with_progressbar(url, save_path):
             progress_bar.update(len(data))
             file.write(data)
     progress_bar.close()
-    if total_size_in_bytes != 0 and progress_bar.n != total_size_in_bytes:
-        logger.error("ERROR, something went wrong")
+    if total_size_in_bytes == 0 or progress_bar.n != total_size_in_bytes:
+        logger.error("Something went wrong while downloading models")
         sys.exit(0)
 
 
@@ -63,7 +96,7 @@ def maybe_download(model_storage_directory, url):
     # using custom model
     if not os.path.exists(os.path.join(
             model_storage_directory, 'model')) or not os.path.exists(
-                os.path.join(model_storage_directory, 'params')):
+        os.path.join(model_storage_directory, 'params')):
         tmp_path = os.path.join(model_storage_directory, url.split('/')[-1])
         print('download {} to {}'.format(url, tmp_path))
         os.makedirs(model_storage_directory, exist_ok=True)
@@ -84,53 +117,102 @@ def maybe_download(model_storage_directory, url):
         os.remove(tmp_path)
 
 
-def parse_args():
+def parse_args(mMain=True, add_help=True):
     import argparse
 
     def str2bool(v):
         return v.lower() in ("true", "t", "1")
 
-    parser = argparse.ArgumentParser()
-    # params for prediction engine
-    parser.add_argument("--use_gpu", type=str2bool, default=True)
-    parser.add_argument("--ir_optim", type=str2bool, default=True)
-    parser.add_argument("--use_tensorrt", type=str2bool, default=False)
-    parser.add_argument("--gpu_mem", type=int, default=8000)
-
-    # params for text detector
-    parser.add_argument("--image_dir", type=str)
-    parser.add_argument("--det_algorithm", type=str, default='DB')
-    parser.add_argument("--det_model_dir", type=str, default=None)
-    parser.add_argument("--det_max_side_len", type=float, default=960)
-
-    # DB parmas
-    parser.add_argument("--det_db_thresh", type=float, default=0.3)
-    parser.add_argument("--det_db_box_thresh", type=float, default=0.5)
-    parser.add_argument("--det_db_unclip_ratio", type=float, default=2.0)
-
-    # EAST parmas
-    parser.add_argument("--det_east_score_thresh", type=float, default=0.8)
-    parser.add_argument("--det_east_cover_thresh", type=float, default=0.1)
-    parser.add_argument("--det_east_nms_thresh", type=float, default=0.2)
-
-    # params for text recognizer
-    parser.add_argument("--rec_algorithm", type=str, default='CRNN')
-    parser.add_argument("--rec_model_dir", type=str, default=None)
-    parser.add_argument("--rec_image_shape", type=str, default="3, 32, 320")
-    parser.add_argument("--rec_char_type", type=str, default='ch')
-    parser.add_argument("--rec_batch_num", type=int, default=30)
-    parser.add_argument("--max_text_length", type=int, default=25)
-    parser.add_argument(
-        "--rec_char_dict_path",
-        type=str,
-        default="./ppocr/utils/ppocr_keys_v1.txt")
-    parser.add_argument("--use_space_char", type=bool, default=True)
-    parser.add_argument("--enable_mkldnn", type=bool, default=False)
-
-    parser.add_argument("--det", type=str2bool, default=True)
-    parser.add_argument("--rec", type=str2bool, default=True)
-    parser.add_argument("--use_zero_copy_run", type=bool, default=False)
-    return parser.parse_args()
+    if mMain:
+        parser = argparse.ArgumentParser(add_help=add_help)
+        # params for prediction engine
+        parser.add_argument("--use_gpu", type=str2bool, default=True)
+        parser.add_argument("--ir_optim", type=str2bool, default=True)
+        parser.add_argument("--use_tensorrt", type=str2bool, default=False)
+        parser.add_argument("--gpu_mem", type=int, default=8000)
+
+        # params for text detector
+        parser.add_argument("--image_dir", type=str)
+        parser.add_argument("--det_algorithm", type=str, default='DB')
+        parser.add_argument("--det_model_dir", type=str, default=None)
+        parser.add_argument("--det_limit_side_len", type=float, default=960)
+        parser.add_argument("--det_limit_type", type=str, default='max')
+
+        # DB parmas
+        parser.add_argument("--det_db_thresh", type=float, default=0.3)
+        parser.add_argument("--det_db_box_thresh", type=float, default=0.5)
+        parser.add_argument("--det_db_unclip_ratio", type=float, default=2.0)
+
+        # EAST parmas
+        parser.add_argument("--det_east_score_thresh", type=float, default=0.8)
+        parser.add_argument("--det_east_cover_thresh", type=float, default=0.1)
+        parser.add_argument("--det_east_nms_thresh", type=float, default=0.2)
+
+        # params for text recognizer
+        parser.add_argument("--rec_algorithm", type=str, default='CRNN')
+        parser.add_argument("--rec_model_dir", type=str, default=None)
+        parser.add_argument("--rec_image_shape", type=str, default="3, 32, 320")
+        parser.add_argument("--rec_char_type", type=str, default='ch')
+        parser.add_argument("--rec_batch_num", type=int, default=30)
+        parser.add_argument("--max_text_length", type=int, default=25)
+        parser.add_argument("--rec_char_dict_path", type=str, default=None)
+        parser.add_argument("--use_space_char", type=bool, default=True)
+        parser.add_argument("--drop_score", type=float, default=0.5)
+
+        # params for text classifier
+        parser.add_argument("--cls_model_dir", type=str, default=None)
+        parser.add_argument("--cls_image_shape", type=str, default="3, 48, 192")
+        parser.add_argument("--label_list", type=list, default=['0', '180'])
+        parser.add_argument("--cls_batch_num", type=int, default=30)
+        parser.add_argument("--cls_thresh", type=float, default=0.9)
+
+        parser.add_argument("--enable_mkldnn", type=bool, default=False)
+        parser.add_argument("--use_zero_copy_run", type=bool, default=False)
+        parser.add_argument("--use_pdserving", type=str2bool, default=False)
+
+        parser.add_argument("--lang", type=str, default='ch')
+        parser.add_argument("--det", type=str2bool, default=True)
+        parser.add_argument("--rec", type=str2bool, default=True)
+        parser.add_argument("--use_angle_cls", type=str2bool, default=False)
+        return parser.parse_args()
+    else:
+        return argparse.Namespace(use_gpu=True,
+                                  ir_optim=True,
+                                  use_tensorrt=False,
+                                  gpu_mem=8000,
+                                  image_dir='',
+                                  det_algorithm='DB',
+                                  det_model_dir=None,
+                                  det_limit_side_len=960,
+                                  det_limit_type='max',
+                                  det_db_thresh=0.3,
+                                  det_db_box_thresh=0.5,
+                                  det_db_unclip_ratio=2.0,
+                                  det_east_score_thresh=0.8,
+                                  det_east_cover_thresh=0.1,
+                                  det_east_nms_thresh=0.2,
+                                  rec_algorithm='CRNN',
+                                  rec_model_dir=None,
+                                  rec_image_shape="3, 32, 320",
+                                  rec_char_type='ch',
+                                  rec_batch_num=30,
+                                  max_text_length=25,
+                                  rec_char_dict_path=None,
+                                  use_space_char=True,
+                                  drop_score=0.5,
+                                  cls_model_dir=None,
+                                  cls_image_shape="3, 48, 192",
+                                  label_list=['0', '180'],
+                                  cls_batch_num=30,
+                                  cls_thresh=0.9,
+                                  enable_mkldnn=False,
+                                  use_zero_copy_run=False,
+                                  use_pdserving=False,
+                                  lang='ch',
+                                  det=True,
+                                  rec=True,
+                                  use_angle_cls=False
+                                  )
 
 
 class PaddleOCR(predict_system.TextSystem):
@@ -140,18 +222,31 @@ class PaddleOCR(predict_system.TextSystem):
         args:
             **kwargs: other params show in paddleocr --help
         """
-        postprocess_params = parse_args()
+        postprocess_params = parse_args(mMain=False, add_help=False)
         postprocess_params.__dict__.update(**kwargs)
+        self.use_angle_cls = postprocess_params.use_angle_cls
+        lang = postprocess_params.lang
+        assert lang in model_urls[
+            'rec'], 'param lang must in {}, but got {}'.format(
+            model_urls['rec'].keys(), lang)
+        if postprocess_params.rec_char_dict_path is None:
+            postprocess_params.rec_char_dict_path = model_urls['rec'][lang][
+                'dict_path']
 
         # init model dir
         if postprocess_params.det_model_dir is None:
             postprocess_params.det_model_dir = os.path.join(BASE_DIR, 'det')
         if postprocess_params.rec_model_dir is None:
-            postprocess_params.rec_model_dir = os.path.join(BASE_DIR, 'rec')
+            postprocess_params.rec_model_dir = os.path.join(
+                BASE_DIR, 'rec/{}'.format(lang))
+        if postprocess_params.cls_model_dir is None:
+            postprocess_params.cls_model_dir = os.path.join(BASE_DIR, 'cls')
         print(postprocess_params)
         # download model
-        maybe_download(postprocess_params.det_model_dir, model_params['det'])
-        maybe_download(postprocess_params.rec_model_dir, model_params['rec'])
+        maybe_download(postprocess_params.det_model_dir, model_urls['det'])
+        maybe_download(postprocess_params.rec_model_dir,
+                       model_urls['rec'][lang]['url'])
+        maybe_download(postprocess_params.cls_model_dir, model_urls['cls'])
 
         if postprocess_params.det_algorithm not in SUPPORT_DET_MODEL:
             logger.error('det_algorithm must in {}'.format(SUPPORT_DET_MODEL))
@@ -166,7 +261,7 @@ class PaddleOCR(predict_system.TextSystem):
         # init det_model and rec_model
         super().__init__(postprocess_params)
 
-    def ocr(self, img, det=True, rec=True):
+    def ocr(self, img, det=True, rec=True, cls=False):
         """
         ocr with paddleocr
         args：
@@ -175,7 +270,16 @@ class PaddleOCR(predict_system.TextSystem):
             rec: use text recognition or not, if false, only det will be exec. default is True
         """
         assert isinstance(img, (np.ndarray, list, str))
+        if isinstance(img, list) and det == True:
+            logger.error('When input a list of images, det must be false')
+            exit(0)
+
+        self.use_angle_cls = cls
         if isinstance(img, str):
+            # download net image
+            if img.startswith('http'):
+                download_with_progressbar(img, 'tmp.jpg')
+                img = 'tmp.jpg'
             image_file = img
             img, flag = check_and_read_gif(image_file)
             if not flag:
@@ -183,6 +287,8 @@ class PaddleOCR(predict_system.TextSystem):
             if img is None:
                 logger.error("error in loading image:{}".format(image_file))
                 return None
+        if isinstance(img, np.ndarray) and len(img.shape) == 2:
+            img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
         if det and rec:
             dt_boxes, rec_res = self.__call__(img)
             return [[box.tolist(), res] for box, res in zip(dt_boxes, rec_res)]
@@ -194,20 +300,34 @@ class PaddleOCR(predict_system.TextSystem):
         else:
             if not isinstance(img, list):
                 img = [img]
+            if self.use_angle_cls:
+                img, cls_res, elapse = self.text_classifier(img)
+                if not rec:
+                    return cls_res
             rec_res, elapse = self.text_recognizer(img)
             return rec_res
 
 
 def main():
-    # for com
-    args = parse_args()
-    image_file_list = get_image_file_list(args.image_dir)
+    # for cmd
+    args = parse_args(mMain=True)
+    image_dir = args.image_dir
+    if image_dir.startswith('http'):
+        download_with_progressbar(image_dir, 'tmp.jpg')
+        image_file_list = ['tmp.jpg']
+    else:
+        image_file_list = get_image_file_list(args.image_dir)
     if len(image_file_list) == 0:
         logger.error('no images find in {}'.format(args.image_dir))
         return
-    ocr_engine = PaddleOCR()
+
+    ocr_engine = PaddleOCR(**(args.__dict__))
     for img_path in image_file_list:
-        print(img_path)
-        result = ocr_engine.ocr(img_path, det=args.det, rec=args.rec)
-        for line in result:
-            print(line)
\ No newline at end of file
+        logger.info('{}{}{}'.format('*' * 10, img_path, '*' * 10))
+        result = ocr_engine.ocr(img_path,
+                                det=args.det,
+                                rec=args.rec,
+                                cls=args.use_angle_cls)
+        if result is not None:
+            for line in result:
+                logger.info(line)

ppocr/data/imaug/__init__.py

@@ -26,6 +26,9 @@ from .randaugment import RandAugment
 from .operators import *
 from .label_ops import *
 
+from .east_process import *
+from .sast_process import *
+
 
 def transform(data, ops=None):
     """ transform """

ppocr/data/imaug/east_process.py

+#copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
+#
+#Licensed under the Apache License, Version 2.0 (the "License");
+#you may not use this file except in compliance with the License.
+#You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+#Unless required by applicable law or agreed to in writing, software
+#distributed under the License is distributed on an "AS IS" BASIS,
+#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#See the License for the specific language governing permissions and
+#limitations under the License.
+
+import math
+import cv2
+import numpy as np
+import json
+import sys
+import os
+
+__all__ = ['EASTProcessTrain']
+
+
+class EASTProcessTrain(object):
+    def __init__(self,
+                 image_shape = [512, 512],
+                 background_ratio = 0.125,
+                 min_crop_side_ratio = 0.1,
+                 min_text_size = 10,
+                 **kwargs):
+        self.input_size = image_shape[1]
+        self.random_scale = np.array([0.5, 1, 2.0, 3.0])
+        self.background_ratio = background_ratio
+        self.min_crop_side_ratio = min_crop_side_ratio
+        self.min_text_size = min_text_size
+
+    def preprocess(self, im):
+        input_size = self.input_size
+        im_shape = im.shape
+        im_size_min = np.min(im_shape[0:2])
+        im_size_max = np.max(im_shape[0:2])
+        im_scale = float(input_size) / float(im_size_max)
+        im = cv2.resize(im, None, None, fx=im_scale, fy=im_scale)
+        img_mean = [0.485, 0.456, 0.406]
+        img_std = [0.229, 0.224, 0.225]
+        # im = im[:, :, ::-1].astype(np.float32)
+        im = im / 255
+        im -= img_mean
+        im /= img_std
+        new_h, new_w, _ = im.shape
+        im_padded = np.zeros((input_size, input_size, 3), dtype=np.float32)
+        im_padded[:new_h, :new_w, :] = im
+        im_padded = im_padded.transpose((2, 0, 1))
+        im_padded = im_padded[np.newaxis, :]
+        return im_padded, im_scale
+
+    def rotate_im_poly(self, im, text_polys):
+        """
+        rotate image with 90 / 180 / 270 degre
+        """
+        im_w, im_h = im.shape[1], im.shape[0]
+        dst_im = im.copy()
+        dst_polys = []
+        rand_degree_ratio = np.random.rand()
+        rand_degree_cnt = 1
+        if 0.333 < rand_degree_ratio < 0.666:
+            rand_degree_cnt = 2
+        elif rand_degree_ratio > 0.666:
+            rand_degree_cnt = 3
+        for i in range(rand_degree_cnt):
+            dst_im = np.rot90(dst_im)
+        rot_degree = -90 * rand_degree_cnt
+        rot_angle = rot_degree * math.pi / 180.0
+        n_poly = text_polys.shape[0]
+        cx, cy = 0.5 * im_w, 0.5 * im_h
+        ncx, ncy = 0.5 * dst_im.shape[1], 0.5 * dst_im.shape[0]
+        for i in range(n_poly):
+            wordBB = text_polys[i]
+            poly = []
+            for j in range(4):
+                sx, sy = wordBB[j][0], wordBB[j][1]
+                dx = math.cos(rot_angle) * (sx - cx)\
+                    - math.sin(rot_angle) * (sy - cy) + ncx
+                dy = math.sin(rot_angle) * (sx - cx)\
+                    + math.cos(rot_angle) * (sy - cy) + ncy
+                poly.append([dx, dy])
+            dst_polys.append(poly)
+        dst_polys = np.array(dst_polys, dtype=np.float32)
+        return dst_im, dst_polys
+
+    def polygon_area(self, poly):
+        """
+        compute area of a polygon
+        :param poly:
+        :return:
+        """
+        edge = [(poly[1][0] - poly[0][0]) * (poly[1][1] + poly[0][1]),
+                (poly[2][0] - poly[1][0]) * (poly[2][1] + poly[1][1]),
+                (poly[3][0] - poly[2][0]) * (poly[3][1] + poly[2][1]),
+                (poly[0][0] - poly[3][0]) * (poly[0][1] + poly[3][1])]
+        return np.sum(edge) / 2.
+
+    def check_and_validate_polys(self, polys, tags, img_height, img_width):
+        """
+        check so that the text poly is in the same direction,
+        and also filter some invalid polygons
+        :param polys:
+        :param tags:
+        :return:
+        """
+        h, w = img_height, img_width
+        if polys.shape[0] == 0:
+            return polys
+        polys[:, :, 0] = np.clip(polys[:, :, 0], 0, w - 1)
+        polys[:, :, 1] = np.clip(polys[:, :, 1], 0, h - 1)
+
+        validated_polys = []
+        validated_tags = []
+        for poly, tag in zip(polys, tags):
+            p_area = self.polygon_area(poly)
+            #invalid poly
+            if abs(p_area) < 1:
+                continue
+            if p_area > 0:
+                #'poly in wrong direction'
+                if not tag:
+                    tag = True  #reversed cases should be ignore
+                poly = poly[(0, 3, 2, 1), :]
+            validated_polys.append(poly)
+            validated_tags.append(tag)
+        return np.array(validated_polys), np.array(validated_tags)
+
+    def draw_img_polys(self, img, polys):
+        if len(img.shape) == 4:
+            img = np.squeeze(img, axis=0)
+        if img.shape[0] == 3:
+            img = img.transpose((1, 2, 0))
+            img[:, :, 2] += 123.68
+            img[:, :, 1] += 116.78
+            img[:, :, 0] += 103.94
+        cv2.imwrite("tmp.jpg", img)
+        img = cv2.imread("tmp.jpg")
+        for box in polys:
+            box = box.astype(np.int32).reshape((-1, 1, 2))
+            cv2.polylines(img, [box], True, color=(255, 255, 0), thickness=2)
+        import random
+        ino = random.randint(0, 100)
+        cv2.imwrite("tmp_%d.jpg" % ino, img)
+        return
+
+    def shrink_poly(self, poly, r):
+        """
+        fit a poly inside the origin poly, maybe bugs here...
+        used for generate the score map
+        :param poly: the text poly
+        :param r: r in the paper
+        :return: the shrinked poly
+        """
+        # shrink ratio
+        R = 0.3
+        # find the longer pair
+        dist0 = np.linalg.norm(poly[0] - poly[1])
+        dist1 = np.linalg.norm(poly[2] - poly[3])
+        dist2 = np.linalg.norm(poly[0] - poly[3])
+        dist3 = np.linalg.norm(poly[1] - poly[2])
+        if dist0 + dist1 > dist2 + dist3:
+            # first move (p0, p1), (p2, p3), then (p0, p3), (p1, p2)
+            ## p0, p1
+            theta = np.arctan2((poly[1][1] - poly[0][1]),
+                               (poly[1][0] - poly[0][0]))
+            poly[0][0] += R * r[0] * np.cos(theta)
+            poly[0][1] += R * r[0] * np.sin(theta)
+            poly[1][0] -= R * r[1] * np.cos(theta)
+            poly[1][1] -= R * r[1] * np.sin(theta)
+            ## p2, p3
+            theta = np.arctan2((poly[2][1] - poly[3][1]),
+                               (poly[2][0] - poly[3][0]))
+            poly[3][0] += R * r[3] * np.cos(theta)
+            poly[3][1] += R * r[3] * np.sin(theta)
+            poly[2][0] -= R * r[2] * np.cos(theta)
+            poly[2][1] -= R * r[2] * np.sin(theta)
+            ## p0, p3
+            theta = np.arctan2((poly[3][0] - poly[0][0]),
+                               (poly[3][1] - poly[0][1]))
+            poly[0][0] += R * r[0] * np.sin(theta)
+            poly[0][1] += R * r[0] * np.cos(theta)
+            poly[3][0] -= R * r[3] * np.sin(theta)
+            poly[3][1] -= R * r[3] * np.cos(theta)
+            ## p1, p2
+            theta = np.arctan2((poly[2][0] - poly[1][0]),
+                               (poly[2][1] - poly[1][1]))
+            poly[1][0] += R * r[1] * np.sin(theta)
+            poly[1][1] += R * r[1] * np.cos(theta)
+            poly[2][0] -= R * r[2] * np.sin(theta)
+            poly[2][1] -= R * r[2] * np.cos(theta)
+        else:
+            ## p0, p3
+            # print poly
+            theta = np.arctan2((poly[3][0] - poly[0][0]),
+                               (poly[3][1] - poly[0][1]))
+            poly[0][0] += R * r[0] * np.sin(theta)
+            poly[0][1] += R * r[0] * np.cos(theta)
+            poly[3][0] -= R * r[3] * np.sin(theta)
+            poly[3][1] -= R * r[3] * np.cos(theta)
+            ## p1, p2
+            theta = np.arctan2((poly[2][0] - poly[1][0]),
+                               (poly[2][1] - poly[1][1]))
+            poly[1][0] += R * r[1] * np.sin(theta)
+            poly[1][1] += R * r[1] * np.cos(theta)
+            poly[2][0] -= R * r[2] * np.sin(theta)
+            poly[2][1] -= R * r[2] * np.cos(theta)
+            ## p0, p1
+            theta = np.arctan2((poly[1][1] - poly[0][1]),
+                               (poly[1][0] - poly[0][0]))
+            poly[0][0] += R * r[0] * np.cos(theta)
+            poly[0][1] += R * r[0] * np.sin(theta)
+            poly[1][0] -= R * r[1] * np.cos(theta)
+            poly[1][1] -= R * r[1] * np.sin(theta)
+            ## p2, p3
+            theta = np.arctan2((poly[2][1] - poly[3][1]),
+                               (poly[2][0] - poly[3][0]))
+            poly[3][0] += R * r[3] * np.cos(theta)
+            poly[3][1] += R * r[3] * np.sin(theta)
+            poly[2][0] -= R * r[2] * np.cos(theta)
+            poly[2][1] -= R * r[2] * np.sin(theta)
+        return poly
+
+    def generate_quad(self, im_size, polys, tags):
+        """
+        Generate quadrangle.
+        """
+        h, w = im_size
+        poly_mask = np.zeros((h, w), dtype=np.uint8)
+        score_map = np.zeros((h, w), dtype=np.uint8)
+        # (x1, y1, ..., x4, y4, short_edge_norm)
+        geo_map = np.zeros((h, w, 9), dtype=np.float32)
+        # mask used during traning, to ignore some hard areas
+        training_mask = np.ones((h, w), dtype=np.uint8)
+        for poly_idx, poly_tag in enumerate(zip(polys, tags)):
+            poly = poly_tag[0]
+            tag = poly_tag[1]
+
+            r = [None, None, None, None]
+            for i in range(4):
+                dist1 = np.linalg.norm(poly[i] - poly[(i + 1) % 4])
+                dist2 = np.linalg.norm(poly[i] - poly[(i - 1) % 4])
+                r[i] = min(dist1, dist2)
+            # score map
+            shrinked_poly = self.shrink_poly(
+                poly.copy(), r).astype(np.int32)[np.newaxis, :, :]
+            cv2.fillPoly(score_map, shrinked_poly, 1)
+            cv2.fillPoly(poly_mask, shrinked_poly, poly_idx + 1)
+            # if the poly is too small, then ignore it during training
+            poly_h = min(
+                np.linalg.norm(poly[0] - poly[3]),
+                np.linalg.norm(poly[1] - poly[2]))
+            poly_w = min(
+                np.linalg.norm(poly[0] - poly[1]),
+                np.linalg.norm(poly[2] - poly[3]))
+            if min(poly_h, poly_w) < self.min_text_size:
+                cv2.fillPoly(training_mask,
+                             poly.astype(np.int32)[np.newaxis, :, :], 0)
+
+            if tag:
+                cv2.fillPoly(training_mask,
+                             poly.astype(np.int32)[np.newaxis, :, :], 0)
+
+            xy_in_poly = np.argwhere(poly_mask == (poly_idx + 1))
+            # geo map.
+            y_in_poly = xy_in_poly[:, 0]
+            x_in_poly = xy_in_poly[:, 1]
+            poly[:, 0] = np.minimum(np.maximum(poly[:, 0], 0), w)
+            poly[:, 1] = np.minimum(np.maximum(poly[:, 1], 0), h)
+            for pno in range(4):
+                geo_channel_beg = pno * 2
+                geo_map[y_in_poly, x_in_poly, geo_channel_beg] =\
+                    x_in_poly - poly[pno, 0]
+                geo_map[y_in_poly, x_in_poly, geo_channel_beg+1] =\
+                    y_in_poly - poly[pno, 1]
+            geo_map[y_in_poly, x_in_poly, 8] = \
+                1.0 / max(min(poly_h, poly_w), 1.0)
+        return score_map, geo_map, training_mask
+
+    def crop_area(self,
+                  im,
+                  polys,
+                  tags,
+                  crop_background=False,
+                  max_tries=50):
+        """
+        make random crop from the input image
+        :param im:
+        :param polys:
+        :param tags:
+        :param crop_background:
+        :param max_tries:
+        :return:
+        """
+        h, w, _ = im.shape
+        pad_h = h // 10
+        pad_w = w // 10
+        h_array = np.zeros((h + pad_h * 2), dtype=np.int32)
+        w_array = np.zeros((w + pad_w * 2), dtype=np.int32)
+        for poly in polys:
+            poly = np.round(poly, decimals=0).astype(np.int32)
+            minx = np.min(poly[:, 0])
+            maxx = np.max(poly[:, 0])
+            w_array[minx + pad_w:maxx + pad_w] = 1
+            miny = np.min(poly[:, 1])
+            maxy = np.max(poly[:, 1])
+            h_array[miny + pad_h:maxy + pad_h] = 1
+        # ensure the cropped area not across a text
+        h_axis = np.where(h_array == 0)[0]
+        w_axis = np.where(w_array == 0)[0]
+        if len(h_axis) == 0 or len(w_axis) == 0:
+            return im, polys, tags
+
+        for i in range(max_tries):
+            xx = np.random.choice(w_axis, size=2)
+            xmin = np.min(xx) - pad_w
+            xmax = np.max(xx) - pad_w
+            xmin = np.clip(xmin, 0, w - 1)
+            xmax = np.clip(xmax, 0, w - 1)
+            yy = np.random.choice(h_axis, size=2)
+            ymin = np.min(yy) - pad_h
+            ymax = np.max(yy) - pad_h
+            ymin = np.clip(ymin, 0, h - 1)
+            ymax = np.clip(ymax, 0, h - 1)
+            if xmax - xmin < self.min_crop_side_ratio * w or \
+               ymax - ymin < self.min_crop_side_ratio * h:
+                # area too small
+                continue
+            if polys.shape[0] != 0:
+                poly_axis_in_area = (polys[:, :, 0] >= xmin)\
+                    & (polys[:, :, 0] <= xmax)\
+                    & (polys[:, :, 1] >= ymin)\
+                    & (polys[:, :, 1] <= ymax)
+                selected_polys = np.where(
+                    np.sum(poly_axis_in_area, axis=1) == 4)[0]
+            else:
+                selected_polys = []
+
+            if len(selected_polys) == 0:
+                # no text in this area
+                if crop_background:
+                    im = im[ymin:ymax + 1, xmin:xmax + 1, :]
+                    polys = []
+                    tags = []
+                    return im, polys, tags
+                else:
+                    continue
+
+            im = im[ymin:ymax + 1, xmin:xmax + 1, :]
+            polys = polys[selected_polys]
+            tags = tags[selected_polys]
+            polys[:, :, 0] -= xmin
+            polys[:, :, 1] -= ymin
+            return im, polys, tags
+        return im, polys, tags
+
+    def crop_background_infor(self, im, text_polys, text_tags):
+        im, text_polys, text_tags = self.crop_area(
+            im, text_polys, text_tags, crop_background=True)
+
+        if len(text_polys) > 0:
+            return None
+        # pad and resize image
+        input_size = self.input_size
+        im, ratio = self.preprocess(im)
+        score_map = np.zeros((input_size, input_size), dtype=np.float32)
+        geo_map = np.zeros((input_size, input_size, 9), dtype=np.float32)
+        training_mask = np.ones((input_size, input_size), dtype=np.float32)
+        return im, score_map, geo_map, training_mask
+
+    def crop_foreground_infor(self, im, text_polys, text_tags):
+        im, text_polys, text_tags = self.crop_area(
+            im, text_polys, text_tags, crop_background=False)
+
+        if text_polys.shape[0] == 0:
+            return None
+        #continue for all ignore case
+        if np.sum((text_tags * 1.0)) >= text_tags.size:
+            return None
+        # pad and resize image
+        input_size = self.input_size
+        im, ratio = self.preprocess(im)
+        text_polys[:, :, 0] *= ratio
+        text_polys[:, :, 1] *= ratio
+        _, _, new_h, new_w = im.shape
+        #         print(im.shape)
+        #         self.draw_img_polys(im, text_polys)
+        score_map, geo_map, training_mask = self.generate_quad(
+            (new_h, new_w), text_polys, text_tags)
+        return im, score_map, geo_map, training_mask
+
+    def __call__(self, data):
+        im = data['image']
+        text_polys = data['polys']
+        text_tags = data['ignore_tags']
+        if im is None:
+            return None
+        if text_polys.shape[0] == 0:
+            return None
+
+        #add rotate cases
+        if np.random.rand() < 0.5:
+            im, text_polys = self.rotate_im_poly(im, text_polys)
+        h, w, _ = im.shape
+        text_polys, text_tags = self.check_and_validate_polys(text_polys,
+                                                              text_tags, h, w)
+        if text_polys.shape[0] == 0:
+            return None
+
+        # random scale this image
+        rd_scale = np.random.choice(self.random_scale)
+        im = cv2.resize(im, dsize=None, fx=rd_scale, fy=rd_scale)
+        text_polys *= rd_scale
+        if np.random.rand() < self.background_ratio:
+            outs = self.crop_background_infor(im, text_polys, text_tags)
+        else:
+            outs = self.crop_foreground_infor(im, text_polys, text_tags)
+
+        if outs is None:
+            return None
+        im, score_map, geo_map, training_mask = outs
+        score_map = score_map[np.newaxis, ::4, ::4].astype(np.float32)
+        geo_map = np.swapaxes(geo_map, 1, 2)
+        geo_map = np.swapaxes(geo_map, 1, 0)
+        geo_map = geo_map[:, ::4, ::4].astype(np.float32)
+        training_mask = training_mask[np.newaxis, ::4, ::4]
+        training_mask = training_mask.astype(np.float32)
+
+        data['image'] = im[0]
+        data['score_map'] = score_map
+        data['geo_map'] = geo_map
+        data['training_mask'] = training_mask
+        # print(im.shape, score_map.shape, geo_map.shape, training_mask.shape)
+        return data
\ No newline at end of file

ppocr/data/imaug/label_ops.py

@@ -52,6 +52,7 @@ class DetLabelEncode(object):
                 txt_tags.append(True)
             else:
                 txt_tags.append(False)
+        boxes = self.expand_points_num(boxes)
         boxes = np.array(boxes, dtype=np.float32)
         txt_tags = np.array(txt_tags, dtype=np.bool)
 
@@ -70,6 +71,17 @@ class DetLabelEncode(object):
         rect[3] = pts[np.argmax(diff)]
         return rect
 
+    def expand_points_num(self, boxes):
+        max_points_num = 0
+        for box in boxes:
+            if len(box) > max_points_num:
+                max_points_num = len(box)
+        ex_boxes = []
+        for box in boxes:
+            ex_box = box + [box[-1]] * (max_points_num - len(box))
+            ex_boxes.append(ex_box)
+        return ex_boxes
+
 
 class BaseRecLabelEncode(object):
     """ Convert between text-label and text-index """
@@ -79,15 +91,17 @@ class BaseRecLabelEncode(object):
                  character_dict_path=None,
                  character_type='ch',
                  use_space_char=False):
-        support_character_type = ['ch', 'en', 'en_sensitive']
+        support_character_type = [
+            'ch', 'en', 'en_sensitive', 'french', 'german', 'japan', 'korean'
+        ]
         assert character_type in support_character_type, "Only {} are supported now but get {}".format(
-            support_character_type, self.character_str)
+            support_character_type, character_type)
 
         self.max_text_len = max_text_length
         if character_type == "en":
             self.character_str = "0123456789abcdefghijklmnopqrstuvwxyz"
             dict_character = list(self.character_str)
-        elif character_type == "ch":
+        elif character_type in ["ch", "french", "german", "japan", "korean"]:
             self.character_str = ""
             assert character_dict_path is not None, "character_dict_path should not be None when character_type is ch"
             with open(character_dict_path, "rb") as fin:

ppocr/data/imaug/operators.py

@@ -42,6 +42,8 @@ class DecodeImage(object):
                 img) > 0, "invalid input 'img' in DecodeImage"
         img = np.frombuffer(img, dtype='uint8')
         img = cv2.imdecode(img, 1)
+        if img is None:
+            return None
         if self.img_mode == 'GRAY':
             img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
         elif self.img_mode == 'RGB':
@@ -120,26 +122,37 @@ class DetResizeForTest(object):
         if 'limit_side_len' in kwargs:
             self.limit_side_len = kwargs['limit_side_len']
             self.limit_type = kwargs.get('limit_type', 'min')
+        if 'resize_long' in kwargs:
+            self.resize_type = 2
+            self.resize_long = kwargs.get('resize_long', 960)
         else:
             self.limit_side_len = 736
             self.limit_type = 'min'
 
     def __call__(self, data):
         img = data['image']
+        src_h, src_w, _ = img.shape
 
         if self.resize_type == 0:
-            img, shape = self.resize_image_type0(img)
+            # img, shape = self.resize_image_type0(img)
+            img, [ratio_h, ratio_w] = self.resize_image_type0(img)
+        elif self.resize_type == 2:
+            img, [ratio_h, ratio_w] = self.resize_image_type2(img)
         else:
-            img, shape = self.resize_image_type1(img)
+            # img, shape = self.resize_image_type1(img)
+            img, [ratio_h, ratio_w] = self.resize_image_type1(img)
         data['image'] = img
-        data['shape'] = shape
+        data['shape'] = np.array([src_h, src_w, ratio_h, ratio_w])
         return data
 
     def resize_image_type1(self, img):
         resize_h, resize_w = self.image_shape
         ori_h, ori_w = img.shape[:2]  # (h, w, c)
+        ratio_h = float(resize_h) / ori_h
+        ratio_w = float(resize_w) / ori_w
         img = cv2.resize(img, (int(resize_w), int(resize_h)))
-        return img, np.array([ori_h, ori_w])
+        # return img, np.array([ori_h, ori_w])
+        return img, [ratio_h, ratio_w]
 
     def resize_image_type0(self, img):
         """
@@ -182,4 +195,31 @@ class DetResizeForTest(object):
         except:
             print(img.shape, resize_w, resize_h)
             sys.exit(0)
-        return img, np.array([h, w])
+        ratio_h = resize_h / float(h)
+        ratio_w = resize_w / float(w)
+        # return img, np.array([h, w])
+        return img, [ratio_h, ratio_w]
+
+    def resize_image_type2(self, img):
+        h, w, _ = img.shape
+
+        resize_w = w
+        resize_h = h
+
+        # Fix the longer side
+        if resize_h > resize_w:
+            ratio = float(self.resize_long) / resize_h
+        else:
+            ratio = float(self.resize_long) / resize_w
+
+        resize_h = int(resize_h * ratio)
+        resize_w = int(resize_w * ratio)
+
+        max_stride = 128
+        resize_h = (resize_h + max_stride - 1) // max_stride * max_stride
+        resize_w = (resize_w + max_stride - 1) // max_stride * max_stride
+        img = cv2.resize(img, (int(resize_w), int(resize_h)))
+        ratio_h = resize_h / float(h)
+        ratio_w = resize_w / float(w)
+
+        return img, [ratio_h, ratio_w]

ppocr/data/simple_dataset.py

@@ -27,14 +27,13 @@ class SimpleDataSet(Dataset):
         global_config = config['Global']
         dataset_config = config[mode]['dataset']
         loader_config = config[mode]['loader']
-        batch_size = loader_config['batch_size_per_card']
 
         self.delimiter = dataset_config.get('delimiter', '\t')
         label_file_list = dataset_config.pop('label_file_list')
         data_source_num = len(label_file_list)
         ratio_list = dataset_config.get("ratio_list", [1.0])
         if isinstance(ratio_list, (float, int)):
-            ratio_list = [float(ratio_list)] * len(data_source_num)
+            ratio_list = [float(ratio_list)] * int(data_source_num)
 
         assert len(
             ratio_list
@@ -76,6 +75,8 @@ class SimpleDataSet(Dataset):
             label = substr[1]
             img_path = os.path.join(self.data_dir, file_name)
             data = {'img_path': img_path, 'label': label}
+            if not os.path.exists(img_path):
+                raise Exception("{} does not exist!".format(img_path))
             with open(data['img_path'], 'rb') as f:
                 img = f.read()
                 data['image'] = img

ppocr/losses/__init__.py

@@ -18,6 +18,8 @@ import copy
 def build_loss(config):
     # det loss
     from .det_db_loss import DBLoss
+    from .det_east_loss import EASTLoss
+    from .det_sast_loss import SASTLoss
 
     # rec loss
     from .rec_ctc_loss import CTCLoss
@@ -25,7 +27,7 @@ def build_loss(config):
     # cls loss
     from .cls_loss import ClsLoss
 
-    support_dict = ['DBLoss', 'CTCLoss', 'ClsLoss']
+    support_dict = ['DBLoss', 'EASTLoss', 'SASTLoss', 'CTCLoss', 'ClsLoss']
 
     config = copy.deepcopy(config)
     module_name = config.pop('name')

ppocr/losses/det_east_loss.py

+# copyright (c) 2019 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import paddle
+from paddle import nn
+from .det_basic_loss import DiceLoss
+
+
+class EASTLoss(nn.Layer):
+    """
+    """
+
+    def __init__(self,
+                 eps=1e-6,
+                 **kwargs):
+        super(EASTLoss, self).__init__()
+        self.dice_loss = DiceLoss(eps=eps)
+
+    def forward(self, predicts, labels):
+        l_score, l_geo, l_mask = labels[1:]
+        f_score = predicts['f_score']
+        f_geo = predicts['f_geo']
+
+        dice_loss = self.dice_loss(f_score, l_score, l_mask)
+
+        #smoooth_l1_loss
+        channels = 8
+        l_geo_split = paddle.split(
+            l_geo, num_or_sections=channels + 1, axis=1)
+        f_geo_split = paddle.split(f_geo, num_or_sections=channels, axis=1)
+        smooth_l1 = 0
+        for i in range(0, channels):
+            geo_diff = l_geo_split[i] - f_geo_split[i]
+            abs_geo_diff = paddle.abs(geo_diff)
+            smooth_l1_sign = paddle.less_than(abs_geo_diff, l_score)
+            smooth_l1_sign = paddle.cast(smooth_l1_sign, dtype='float32')
+            in_loss = abs_geo_diff * abs_geo_diff * smooth_l1_sign + \
+                (abs_geo_diff - 0.5) * (1.0 - smooth_l1_sign)
+            out_loss = l_geo_split[-1] / channels * in_loss * l_score
+            smooth_l1 += out_loss
+        smooth_l1_loss = paddle.mean(smooth_l1 * l_score)
+
+        dice_loss = dice_loss * 0.01
+        total_loss = dice_loss + smooth_l1_loss
+        losses = {"loss":total_loss, \
+                  "dice_loss":dice_loss,\
+                  "smooth_l1_loss":smooth_l1_loss}
+        return losses

ppocr/losses/det_sast_loss.py

+# copyright (c) 2019 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import paddle
+from paddle import nn
+from .det_basic_loss import DiceLoss
+import paddle.fluid as fluid
+import numpy as np
+
+
+class SASTLoss(nn.Layer):
+    """
+    """
+
+    def __init__(self,
+                 eps=1e-6,
+                 **kwargs):
+        super(SASTLoss, self).__init__()
+        self.dice_loss = DiceLoss(eps=eps)
+
+    def forward(self, predicts, labels):
+        """
+        tcl_pos: N x 128 x 3
+        tcl_mask: N x 128 x 1
+        tcl_label: N x X list or LoDTensor
+        """
+                
+        f_score = predicts['f_score']
+        f_border = predicts['f_border']
+        f_tvo = predicts['f_tvo']
+        f_tco = predicts['f_tco']
+
+        l_score, l_border, l_mask, l_tvo, l_tco = labels[1:]
+
+        #score_loss
+        intersection = paddle.sum(f_score * l_score * l_mask)
+        union = paddle.sum(f_score * l_mask) + paddle.sum(l_score * l_mask)
+        score_loss = 1.0 - 2 * intersection / (union + 1e-5)
+
+        #border loss
+        l_border_split, l_border_norm = paddle.split(l_border, num_or_sections=[4, 1], axis=1)
+        f_border_split = f_border
+        border_ex_shape = l_border_norm.shape * np.array([1, 4, 1, 1])
+        l_border_norm_split = paddle.expand(x=l_border_norm, shape=border_ex_shape)
+        l_border_score = paddle.expand(x=l_score, shape=border_ex_shape)   
+        l_border_mask = paddle.expand(x=l_mask, shape=border_ex_shape)  
+
+        border_diff = l_border_split - f_border_split
+        abs_border_diff = paddle.abs(border_diff) 
+        border_sign = abs_border_diff < 1.0
+        border_sign = paddle.cast(border_sign, dtype='float32')
+        border_sign.stop_gradient = True
+        border_in_loss = 0.5 * abs_border_diff * abs_border_diff * border_sign + \
+                    (abs_border_diff - 0.5) * (1.0 - border_sign)
+        border_out_loss = l_border_norm_split * border_in_loss
+        border_loss = paddle.sum(border_out_loss * l_border_score * l_border_mask) / \
+                    (paddle.sum(l_border_score * l_border_mask) + 1e-5)
+
+        #tvo_loss
+        l_tvo_split, l_tvo_norm = paddle.split(l_tvo, num_or_sections=[8, 1], axis=1)
+        f_tvo_split = f_tvo
+        tvo_ex_shape = l_tvo_norm.shape * np.array([1, 8, 1, 1])
+        l_tvo_norm_split = paddle.expand(x=l_tvo_norm, shape=tvo_ex_shape)
+        l_tvo_score = paddle.expand(x=l_score, shape=tvo_ex_shape)   
+        l_tvo_mask = paddle.expand(x=l_mask, shape=tvo_ex_shape)   
+        #
+        tvo_geo_diff = l_tvo_split - f_tvo_split
+        abs_tvo_geo_diff = paddle.abs(tvo_geo_diff) 
+        tvo_sign = abs_tvo_geo_diff < 1.0
+        tvo_sign = paddle.cast(tvo_sign, dtype='float32')
+        tvo_sign.stop_gradient = True
+        tvo_in_loss = 0.5 * abs_tvo_geo_diff * abs_tvo_geo_diff * tvo_sign + \
+                    (abs_tvo_geo_diff - 0.5) * (1.0 - tvo_sign)
+        tvo_out_loss = l_tvo_norm_split * tvo_in_loss
+        tvo_loss = paddle.sum(tvo_out_loss * l_tvo_score * l_tvo_mask) / \
+                    (paddle.sum(l_tvo_score * l_tvo_mask) + 1e-5)
+
+        #tco_loss
+        l_tco_split, l_tco_norm = paddle.split(l_tco, num_or_sections=[2, 1], axis=1)
+        f_tco_split = f_tco
+        tco_ex_shape = l_tco_norm.shape * np.array([1, 2, 1, 1])
+        l_tco_norm_split = paddle.expand(x=l_tco_norm, shape=tco_ex_shape)
+        l_tco_score = paddle.expand(x=l_score, shape=tco_ex_shape)   
+        l_tco_mask = paddle.expand(x=l_mask, shape=tco_ex_shape) 
+        
+        tco_geo_diff = l_tco_split - f_tco_split
+        abs_tco_geo_diff = paddle.abs(tco_geo_diff) 
+        tco_sign = abs_tco_geo_diff < 1.0
+        tco_sign = paddle.cast(tco_sign, dtype='float32')
+        tco_sign.stop_gradient = True
+        tco_in_loss = 0.5 * abs_tco_geo_diff * abs_tco_geo_diff * tco_sign + \
+                    (abs_tco_geo_diff - 0.5) * (1.0 - tco_sign)
+        tco_out_loss = l_tco_norm_split * tco_in_loss
+        tco_loss = paddle.sum(tco_out_loss * l_tco_score * l_tco_mask) / \
+                    (paddle.sum(l_tco_score * l_tco_mask) + 1e-5)
+
+
+        # total loss
+        tvo_lw, tco_lw = 1.5, 1.5
+        score_lw, border_lw = 1.0, 1.0
+        total_loss = score_loss * score_lw + border_loss * border_lw + \
+                    tvo_loss * tvo_lw + tco_loss * tco_lw
+                    
+        losses = {'loss':total_loss, "score_loss":score_loss,\
+            "border_loss":border_loss, 'tvo_loss':tvo_loss, 'tco_loss':tco_loss}
+        return losses
\ No newline at end of file

ppocr/modeling/architectures/base_model.py

@@ -16,7 +16,7 @@ from __future__ import division
 from __future__ import print_function
 
 from paddle import nn
-from ppocr.modeling.transform import build_transform
+from ppocr.modeling.transforms import build_transform
 from ppocr.modeling.backbones import build_backbone
 from ppocr.modeling.necks import build_neck
 from ppocr.modeling.heads import build_head

ppocr/modeling/backbones/__init__.py

@@ -19,6 +19,7 @@ def build_backbone(config, model_type):
     if model_type == 'det':
         from .det_mobilenet_v3 import MobileNetV3
         from .det_resnet_vd import ResNet
+        from .det_resnet_vd_sast import ResNet_SAST
         support_dict = ['MobileNetV3', 'ResNet', 'ResNet_SAST']
     elif model_type == 'rec' or model_type == 'cls':
         from .rec_mobilenet_v3 import MobileNetV3

ppocr/modeling/backbones/det_mobilenet_v3.py

@@ -34,13 +34,21 @@ def make_divisible(v, divisor=8, min_value=None):
 
 
 class MobileNetV3(nn.Layer):
-    def __init__(self, in_channels=3, model_name='large', scale=0.5, **kwargs):
+    def __init__(self,
+                 in_channels=3,
+                 model_name='large',
+                 scale=0.5,
+                 disable_se=False,
+                 **kwargs):
         """
         the MobilenetV3 backbone network for detection module.
         Args:
             params(dict): the super parameters for build network
         """
         super(MobileNetV3, self).__init__()
+
+        self.disable_se = disable_se
+
         if model_name == "large":
             cfg = [
                 # k, exp, c,  se,     nl,  s,
@@ -103,6 +111,7 @@ class MobileNetV3(nn.Layer):
         i = 0
         inplanes = make_divisible(inplanes * scale)
         for (k, exp, c, se, nl, s) in cfg:
+            se = se and not self.disable_se
             if s == 2 and i > 2:
                 self.out_channels.append(inplanes)
                 self.stages.append(nn.Sequential(*block_list))
@@ -273,4 +282,4 @@ class SEModule(nn.Layer):
         outputs = F.relu(outputs)
         outputs = self.conv2(outputs)
         outputs = F.activation.hard_sigmoid(outputs)
-        return inputs * outputs
\ No newline at end of file
+        return inputs * outputs

ppocr/modeling/backbones/det_resnet_vd_sast.py

+# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import paddle
+from paddle import ParamAttr
+import paddle.nn as nn
+import paddle.nn.functional as F
+
+__all__ = ["ResNet_SAST"]
+
+
+class ConvBNLayer(nn.Layer):
+    def __init__(
+            self,
+            in_channels,
+            out_channels,
+            kernel_size,
+            stride=1,
+            groups=1,
+            is_vd_mode=False,
+            act=None,
+            name=None, ):
+        super(ConvBNLayer, self).__init__()
+
+        self.is_vd_mode = is_vd_mode
+        self._pool2d_avg = nn.AvgPool2D(
+            kernel_size=2, stride=2, padding=0, ceil_mode=True)
+        self._conv = nn.Conv2D(
+            in_channels=in_channels,
+            out_channels=out_channels,
+            kernel_size=kernel_size,
+            stride=stride,
+            padding=(kernel_size - 1) // 2,
+            groups=groups,
+            weight_attr=ParamAttr(name=name + "_weights"),
+            bias_attr=False)
+        if name == "conv1":
+            bn_name = "bn_" + name
+        else:
+            bn_name = "bn" + name[3:]
+        self._batch_norm = nn.BatchNorm(
+            out_channels,
+            act=act,
+            param_attr=ParamAttr(name=bn_name + '_scale'),
+            bias_attr=ParamAttr(bn_name + '_offset'),
+            moving_mean_name=bn_name + '_mean',
+            moving_variance_name=bn_name + '_variance')
+
+    def forward(self, inputs):
+        if self.is_vd_mode:
+            inputs = self._pool2d_avg(inputs)
+        y = self._conv(inputs)
+        y = self._batch_norm(y)
+        return y
+
+
+class BottleneckBlock(nn.Layer):
+    def __init__(self,
+                 in_channels,
+                 out_channels,
+                 stride,
+                 shortcut=True,
+                 if_first=False,
+                 name=None):
+        super(BottleneckBlock, self).__init__()
+
+        self.conv0 = ConvBNLayer(
+            in_channels=in_channels,
+            out_channels=out_channels,
+            kernel_size=1,
+            act='relu',
+            name=name + "_branch2a")
+        self.conv1 = ConvBNLayer(
+            in_channels=out_channels,
+            out_channels=out_channels,
+            kernel_size=3,
+            stride=stride,
+            act='relu',
+            name=name + "_branch2b")
+        self.conv2 = ConvBNLayer(
+            in_channels=out_channels,
+            out_channels=out_channels * 4,
+            kernel_size=1,
+            act=None,
+            name=name + "_branch2c")
+
+        if not shortcut:
+            self.short = ConvBNLayer(
+                in_channels=in_channels,
+                out_channels=out_channels * 4,
+                kernel_size=1,
+                stride=1,
+                is_vd_mode=False if if_first else True,
+                name=name + "_branch1")
+
+        self.shortcut = shortcut
+
+    def forward(self, inputs):
+        y = self.conv0(inputs)
+        conv1 = self.conv1(y)
+        conv2 = self.conv2(conv1)
+
+        if self.shortcut:
+            short = inputs
+        else:
+            short = self.short(inputs)
+        y = paddle.add(x=short, y=conv2)
+        y = F.relu(y)
+        return y
+
+
+class BasicBlock(nn.Layer):
+    def __init__(self,
+                 in_channels,
+                 out_channels,
+                 stride,
+                 shortcut=True,
+                 if_first=False,
+                 name=None):
+        super(BasicBlock, self).__init__()
+        self.stride = stride
+        self.conv0 = ConvBNLayer(
+            in_channels=in_channels,
+            out_channels=out_channels,
+            kernel_size=3,
+            stride=stride,
+            act='relu',
+            name=name + "_branch2a")
+        self.conv1 = ConvBNLayer(
+            in_channels=out_channels,
+            out_channels=out_channels,
+            kernel_size=3,
+            act=None,
+            name=name + "_branch2b")
+
+        if not shortcut:
+            self.short = ConvBNLayer(
+                in_channels=in_channels,
+                out_channels=out_channels,
+                kernel_size=1,
+                stride=1,
+                is_vd_mode=False if if_first else True,
+                name=name + "_branch1")
+
+        self.shortcut = shortcut
+
+    def forward(self, inputs):
+        y = self.conv0(inputs)
+        conv1 = self.conv1(y)
+
+        if self.shortcut:
+            short = inputs
+        else:
+            short = self.short(inputs)
+        y = paddle.add(x=short, y=conv1)
+        y = F.relu(y)
+        return y
+
+
+class ResNet_SAST(nn.Layer):
+    def __init__(self, in_channels=3, layers=50, **kwargs):
+        super(ResNet_SAST, self).__init__()
+
+        self.layers = layers
+        supported_layers = [18, 34, 50, 101, 152, 200]
+        assert layers in supported_layers, \
+            "supported layers are {} but input layer is {}".format(
+                supported_layers, layers)
+
+        if layers == 18:
+            depth = [2, 2, 2, 2]
+        elif layers == 34 or layers == 50:
+            # depth = [3, 4, 6, 3]
+            depth = [3, 4, 6, 3, 3]
+        elif layers == 101:
+            depth = [3, 4, 23, 3]
+        elif layers == 152:
+            depth = [3, 8, 36, 3]
+        elif layers == 200:
+            depth = [3, 12, 48, 3]
+        # num_channels = [64, 256, 512,
+        #                 1024] if layers >= 50 else [64, 64, 128, 256]
+        # num_filters = [64, 128, 256, 512]
+        num_channels = [64, 256, 512,
+                        1024, 2048] if layers >= 50 else [64, 64, 128, 256]
+        num_filters = [64, 128, 256, 512, 512]
+
+        self.conv1_1 = ConvBNLayer(
+            in_channels=in_channels,
+            out_channels=32,
+            kernel_size=3,
+            stride=2,
+            act='relu',
+            name="conv1_1")
+        self.conv1_2 = ConvBNLayer(
+            in_channels=32,
+            out_channels=32,
+            kernel_size=3,
+            stride=1,
+            act='relu',
+            name="conv1_2")
+        self.conv1_3 = ConvBNLayer(
+            in_channels=32,
+            out_channels=64,
+            kernel_size=3,
+            stride=1,
+            act='relu',
+            name="conv1_3")
+        self.pool2d_max = nn.MaxPool2D(kernel_size=3, stride=2, padding=1)
+
+        self.stages = []
+        self.out_channels = [3, 64]
+        if layers >= 50:
+            for block in range(len(depth)):
+                block_list = []
+                shortcut = False
+                for i in range(depth[block]):
+                    if layers in [101, 152] and block == 2:
+                        if i == 0:
+                            conv_name = "res" + str(block + 2) + "a"
+                        else:
+                            conv_name = "res" + str(block + 2) + "b" + str(i)
+                    else:
+                        conv_name = "res" + str(block + 2) + chr(97 + i)
+                    bottleneck_block = self.add_sublayer(
+                        'bb_%d_%d' % (block, i),
+                        BottleneckBlock(
+                            in_channels=num_channels[block]
+                            if i == 0 else num_filters[block] * 4,
+                            out_channels=num_filters[block],
+                            stride=2 if i == 0 and block != 0 else 1,
+                            shortcut=shortcut,
+                            if_first=block == i == 0,
+                            name=conv_name))
+                    shortcut = True
+                    block_list.append(bottleneck_block)
+                self.out_channels.append(num_filters[block] * 4)
+                self.stages.append(nn.Sequential(*block_list))
+        else:
+            for block in range(len(depth)):
+                block_list = []
+                shortcut = False
+                for i in range(depth[block]):
+                    conv_name = "res" + str(block + 2) + chr(97 + i)
+                    basic_block = self.add_sublayer(
+                        'bb_%d_%d' % (block, i),
+                        BasicBlock(
+                            in_channels=num_channels[block]
+                            if i == 0 else num_filters[block],
+                            out_channels=num_filters[block],
+                            stride=2 if i == 0 and block != 0 else 1,
+                            shortcut=shortcut,
+                            if_first=block == i == 0,
+                            name=conv_name))
+                    shortcut = True
+                    block_list.append(basic_block)
+                self.out_channels.append(num_filters[block])
+                self.stages.append(nn.Sequential(*block_list))
+
+    def forward(self, inputs):
+        out = [inputs]
+        y = self.conv1_1(inputs)
+        y = self.conv1_2(y)
+        y = self.conv1_3(y)
+        out.append(y)
+        y = self.pool2d_max(y)
+        for block in self.stages:
+            y = block(y)
+            out.append(y)
+        return out
\ No newline at end of file

ppocr/modeling/heads/__init__.py

@@ -18,13 +18,15 @@ __all__ = ['build_head']
 def build_head(config):
     # det head
     from .det_db_head import DBHead
+    from .det_east_head import EASTHead
+    from .det_sast_head import SASTHead
 
     # rec head
     from .rec_ctc_head import CTCHead
 
     # cls head
     from .cls_head import ClsHead
-    support_dict = ['DBHead', 'CTCHead', 'ClsHead']
+    support_dict = ['DBHead', 'EASTHead', 'SASTHead', 'CTCHead', 'ClsHead']
 
     module_name = config.pop('name')
     assert module_name in support_dict, Exception('head only support {}'.format(

ppocr/modeling/necks/__init__.py

@@ -16,8 +16,10 @@ __all__ = ['build_neck']
 
 def build_neck(config):
     from .db_fpn import DBFPN
+    from .east_fpn import EASTFPN
+    from .sast_fpn import SASTFPN
     from .rnn import SequenceEncoder
-    support_dict = ['DBFPN', 'SequenceEncoder']
+    support_dict = ['DBFPN', 'EASTFPN', 'SASTFPN', 'SequenceEncoder']
 
     module_name = config.pop('name')
     assert module_name in support_dict, Exception('neck only support {}'.format(

ppocr/postprocess/__init__.py

@@ -24,11 +24,13 @@ __all__ = ['build_post_process']
 
 def build_post_process(config, global_config=None):
     from .db_postprocess import DBPostProcess
+    from .east_postprocess import EASTPostProcess
+    from .sast_postprocess import SASTPostProcess
     from .rec_postprocess import CTCLabelDecode, AttnLabelDecode
     from .cls_postprocess import ClsPostProcess
 
     support_dict = [
-        'DBPostProcess', 'CTCLabelDecode', 'AttnLabelDecode', 'ClsPostProcess'
+        'DBPostProcess', 'EASTPostProcess', 'SASTPostProcess', 'CTCLabelDecode', 'AttnLabelDecode', 'ClsPostProcess'
     ]
 
     config = copy.deepcopy(config)

ppocr/postprocess/db_postprocess.py

@@ -33,12 +33,14 @@ class DBPostProcess(object):
                  box_thresh=0.7,
                  max_candidates=1000,
                  unclip_ratio=2.0,
+                 use_dilation=False,
                  **kwargs):
         self.thresh = thresh
         self.box_thresh = box_thresh
         self.max_candidates = max_candidates
         self.unclip_ratio = unclip_ratio
         self.min_size = 3
+        self.dilation_kernel = None if not use_dilation else [[1, 1], [1, 1]]
 
     def boxes_from_bitmap(self, pred, _bitmap, dest_width, dest_height):
         '''
@@ -138,9 +140,15 @@ class DBPostProcess(object):
 
         boxes_batch = []
         for batch_index in range(pred.shape[0]):
-            height, width = shape_list[batch_index]
-            boxes, scores = self.boxes_from_bitmap(
-                pred[batch_index], segmentation[batch_index], width, height)
+            src_h, src_w, ratio_h, ratio_w = shape_list[batch_index]
+            if self.dilation_kernel is not None:
+                mask = cv2.dilate(
+                    np.array(segmentation[batch_index]).astype(np.uint8),
+                    self.dilation_kernel)
+            else:
+                mask = segmentation[batch_index]
+            boxes, scores = self.boxes_from_bitmap(pred[batch_index], mask,
+                                                   src_w, src_h)
 
             boxes_batch.append({'points': boxes})
-        return boxes_batch
\ No newline at end of file
+        return boxes_batch