openocr-python 0.0.2__py3-none-any.whl

openocr/configs/rec/busnet/vit_busnet.yml

@@ -0,0 +1,104 @@
+Global:
+  device: gpu
+  epoch_num: 10
+  log_smooth_window: 20
+  print_batch_step: 10
+  output_dir: ./output/rec/u14m_filter/vit_busnet/
+  eval_epoch_step: [0, 1]
+  eval_batch_step: [0, 500]
+  cal_metric_during_train: True
+  pretrained_model:
+  checkpoints:
+  use_tensorboard: false
+  infer_img:
+  # for data or label process
+  character_dict_path: ./tools/utils/EN_symbol_dict.txt
+  max_text_length: 25
+  use_space_char: False
+  save_res_path: ./output/rec/u14m_filter/predicts_vit_busnet.txt
+  grad_clip_val: 20
+  use_amp: True
+
+Optimizer:
+  name: Adam
+  lr: 0.00053 # 4gpus bs256/gpu
+  weight_decay: 0.0
+  filter_bias_and_bn: False
+
+LRScheduler:
+  name: MultiStepLR
+  milestones: [6]
+  gamma: 0.1
+
+Architecture:
+  model_type: rec
+  algorithm: BUSBet
+  Transform:
+  Encoder:
+    name: ViT
+    img_size: [32,128]
+    patch_size: [4, 8]
+    embed_dim: 384
+    depth: 12
+    num_heads: 6
+    mlp_ratio: 4
+    qkv_bias: True
+  Decoder:
+    name: BUSDecoder
+    nhead: 6
+    num_layers: 6
+    dim_feedforward: 1536
+    ignore_index: &ignore_index 100
+    pretraining: False
+Loss:
+  name: ABINetLoss
+  ignore_index: *ignore_index
+
+PostProcess:
+  name: ABINetLabelDecode
+
+Metric:
+  name: RecMetric
+  main_indicator: acc
+  is_filter: True
+
+Train:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../Union14M-L-LMDB-Filtered
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - PARSeqAugPIL:
+      - ABINetLabelEncode:
+          ignore_index: *ignore_index
+      - RecTVResize:
+          image_shape: [32, 128]
+          padding: False
+      - 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: 4
+
+Eval:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../evaluation
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - ABINetLabelEncode:
+          ignore_index: *ignore_index
+      - RecTVResize:
+          image_shape: [32, 128]
+          padding: False
+      - 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: 2

openocr/configs/rec/busnet/vit_busnet_pretraining.yml

@@ -0,0 +1,104 @@
+Global:
+  device: gpu
+  epoch_num: 10
+  log_smooth_window: 20
+  print_batch_step: 10
+  output_dir: ./output/rec/u14m_filter/vit_busnet_pretraining/
+  eval_epoch_step: [0, 1]
+  eval_batch_step: [0, 500]
+  cal_metric_during_train: True
+  pretrained_model:
+  checkpoints:
+  use_tensorboard: false
+  infer_img:
+  # for data or label process
+  character_dict_path: ./tools/utils/EN_symbol_dict.txt
+  max_text_length: 25
+  use_space_char: False
+  save_res_path: ./output/rec/u14m_filter/predicts_vit_busnet_pretraining.txt
+  grad_clip_val: 20
+  use_amp: True
+
+Optimizer:
+  name: Adam
+  lr: 0.00053 # 4gpus bs256/gpu
+  weight_decay: 0.0
+  filter_bias_and_bn: False
+
+LRScheduler:
+  name: MultiStepLR
+  milestones: [6]
+  gamma: 0.1
+
+Architecture:
+  model_type: rec
+  algorithm: BUSBet
+  Transform:
+  Encoder:
+    name: ViT
+    img_size: [32,128]
+    patch_size: [4, 8]
+    embed_dim: 384
+    depth: 12
+    num_heads: 6
+    mlp_ratio: 4
+    qkv_bias: True
+  Decoder:
+    name: BUSDecoder
+    nhead: 6
+    num_layers: 6
+    dim_feedforward: 1536
+    ignore_index: &ignore_index 100
+    pretraining: True
+Loss:
+  name: ABINetLoss
+  ignore_index: *ignore_index
+
+PostProcess:
+  name: ABINetLabelDecode
+
+Metric:
+  name: RecMetric
+  main_indicator: acc
+  is_filter: True
+
+Train:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../Union14M-L-LMDB-Filtered
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - PARSeqAugPIL:
+      - ABINetLabelEncode:
+          ignore_index: *ignore_index
+      - RecTVResize:
+          image_shape: [32, 128]
+          padding: False
+      - 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: 4
+
+Eval:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../evaluation
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - ABINetLabelEncode:
+          ignore_index: *ignore_index
+      - RecTVResize:
+          image_shape: [32, 128]
+          padding: False
+      - 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: 2

openocr/configs/rec/cam/convnextv2_cam_tps_on.yml

@@ -0,0 +1,118 @@
+Global:
+  device: gpu
+  epoch_num: 20
+  log_smooth_window: 20
+  print_batch_step: 10
+  output_dir: ./output/rec/u14m_filter/convnextv2_cam_tps_on
+  eval_epoch_step: [0, 1]
+  eval_batch_step: [0, 500]
+  cal_metric_during_train: False
+  pretrained_model:
+  checkpoints:
+  use_tensorboard: false
+  infer_img:
+  # for data or label process
+  character_dict_path: ./tools/utils/EN_symbol_dict.txt
+  max_text_length: &max_text_length 25
+  use_space_char: False
+  save_res_path: ./output/rec/u14m_filter/predicts_convnextv2_cam_tps_on.txt
+  use_amp: True
+
+Optimizer:
+  name: AdamW
+  lr: 0.0008 # for 4gpus bs256/gpu
+  weight_decay: 0.05
+  filter_bias_and_bn: True
+  eps: 1.e-8
+
+LRScheduler:
+  name: OneCycleLR
+  warmup_epoch: 1.5 # pct_start 0.075*20 :  1.5ep
+  cycle_momentum: False
+
+Architecture:
+  model_type: rec
+  algorithm: CAM
+  Transform:
+    name: Aster_TPS
+    tps_inputsize: [32, 64]
+    tps_outputsize: &img_shape [32, 128]
+  Encoder:
+    name: CAMEncoder
+    encoder_config:
+      name: ConvNeXtV2
+      depths: [2, 2, 8, 2]
+      dims: [80, 160, 320, 640]
+      strides: [[4,4], [2,1], [2,1], [1,1]]
+      drop_path_rate: 0.2
+      feat2d: True
+    nb_classes: 97
+    strides: [[4,4], [2,1], [2,1], [1,1]]
+    deform_stride: 2
+    stage_idx: 2
+    use_depthwise_unet: True
+    use_more_unet: False
+    binary_loss_type: BanlanceMultiClassCrossEntropyLoss
+    mid_size: True
+    d_embedding: 384
+  Decoder:
+    name: CAMDecoder
+    num_encoder_layers: -1
+    beam_size: 0
+    num_decoder_layers: 2
+    nhead: 8
+    max_len: *max_text_length
+
+Loss:
+  name: CAMLoss
+  loss_weight_binary: 1.5
+  label_smoothing: 0.
+
+Metric:
+  name: RecMetric
+  main_indicator: acc
+  is_filter: True
+
+PostProcess:
+  name: ARLabelDecode
+
+Train:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../Union14M-L-LMDB-Filtered
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - PARSeqAugPIL:
+      - CAMLabelEncode: # Class handling label
+          font_path: ./arial.ttf
+          image_shape: *img_shape
+      - RecTVResize:
+          image_shape: [64, 256]
+          padding: False
+      - KeepKeys:
+          keep_keys: ['image', 'label', 'length', 'binary_mask'] # dataloader will return list in this order
+  loader:
+    shuffle: True
+    batch_size_per_card: 256
+    drop_last: True
+    num_workers: 4
+
+Eval:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../evaluation
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - ARLabelEncode: # Class handling label
+      - RecTVResize:
+          image_shape: [64, 256]
+          padding: False
+      - 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: 2

openocr/configs/rec/cam/convnextv2_tiny_cam_tps_on.yml

@@ -0,0 +1,118 @@
+Global:
+  device: gpu
+  epoch_num: 20
+  log_smooth_window: 20
+  print_batch_step: 10
+  output_dir: ./output/rec/u14m_filter/convnextv2_tiny_cam_tps_on
+  eval_epoch_step: [0, 1]
+  eval_batch_step: [0, 500]
+  cal_metric_during_train: False
+  pretrained_model:
+  checkpoints:
+  use_tensorboard: false
+  infer_img:
+  # for data or label process
+  character_dict_path: ./tools/utils/EN_symbol_dict.txt
+  max_text_length: &max_text_length 25
+  use_space_char: False
+  save_res_path: ./output/rec/u14m_filter/predicts_convnextv2_cam_tps_on.txt
+  use_amp: True
+
+Optimizer:
+  name: AdamW
+  lr: 0.0008 # for 4gpus bs256/gpu
+  weight_decay: 0.05
+  filter_bias_and_bn: True
+  eps: 1.e-8
+
+LRScheduler:
+  name: OneCycleLR
+  warmup_epoch: 1.5 # pct_start 0.075*20 :  1.5ep
+  cycle_momentum: False
+
+Architecture:
+  model_type: rec
+  algorithm: CAM
+  Transform:
+    name: Aster_TPS
+    tps_inputsize: [32, 64]
+    tps_outputsize: &img_shape [32, 128]
+  Encoder:
+    name: CAMEncoder
+    encoder_config:
+      name: ConvNeXtV2
+      depths: [3, 3, 9, 3]
+      dims: [96, 192, 384, 768]
+      strides: [[4,4], [2,1], [2,1], [1,1]]
+      drop_path_rate: 0.2
+      feat2d: True
+    nb_classes: 97
+    strides: [[4,4], [2,1], [2,1], [1,1]]
+    deform_stride: 2
+    stage_idx: 2
+    use_depthwise_unet: True
+    use_more_unet: False
+    binary_loss_type: BanlanceMultiClassCrossEntropyLoss
+    mid_size: False
+    d_embedding: 512
+  Decoder:
+    name: CAMDecoder
+    num_encoder_layers: -1
+    beam_size: 0
+    num_decoder_layers: 2
+    nhead: 8
+    max_len: *max_text_length
+
+Loss:
+  name: CAMLoss
+  loss_weight_binary: 1.5
+  label_smoothing: 0.
+
+Metric:
+  name: RecMetric
+  main_indicator: acc
+  is_filter: True
+
+PostProcess:
+  name: ARLabelDecode
+
+Train:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../Union14M-L-LMDB-Filtered
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - PARSeqAugPIL:
+      - CAMLabelEncode: # Class handling label
+          font_path: ./arial.ttf
+          image_shape: *img_shape
+      - RecTVResize:
+          image_shape: [64, 256]
+          padding: False
+      - KeepKeys:
+          keep_keys: ['image', 'label', 'length', 'binary_mask'] # dataloader will return list in this order
+  loader:
+    shuffle: True
+    batch_size_per_card: 256
+    drop_last: True
+    num_workers: 4
+
+Eval:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../evaluation
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - ARLabelEncode: # Class handling label
+      - RecTVResize:
+          image_shape: [64, 256]
+          padding: False
+      - 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: 2

openocr/configs/rec/cam/svtrv2_cam_tps_on.yml

@@ -0,0 +1,123 @@
+Global:
+  device: gpu
+  epoch_num: 20
+  log_smooth_window: 20
+  print_batch_step: 10
+  output_dir: ./output/rec/u14m_filter/svtrv2_cam_tps_on
+  eval_epoch_step: [0, 1]
+  eval_batch_step: [0, 500]
+  cal_metric_during_train: False
+  pretrained_model:
+  checkpoints:
+  use_tensorboard: false
+  infer_img:
+  # for data or label process
+  character_dict_path: ./tools/utils/EN_symbol_dict.txt
+  max_text_length: &max_text_length 25
+  use_space_char: False
+  save_res_path: ./output/rec/u14m_filter/predicts_svtrv2_cam_tps_on.txt
+  use_amp: True
+
+Optimizer:
+  name: AdamW
+  lr: 0.00065 # for 4gpus bs256/gpu
+  weight_decay: 0.05
+  filter_bias_and_bn: True
+
+LRScheduler:
+  name: OneCycleLR
+  warmup_epoch: 1.5 # pct_start 0.075*20 :  1.5ep
+  cycle_momentum: False
+
+Architecture:
+  model_type: rec
+  algorithm: CAM
+  Transform:
+    name: Aster_TPS
+    tps_inputsize: [32, 64]
+    tps_outputsize: &img_shape [32, 128]
+  Encoder:
+    name: CAMEncoder
+    encoder_config:
+      name: SVTRv2LNConvTwo33
+      use_pos_embed: False
+      dims: [128, 256, 384]
+      depths: [6, 6, 6]
+      num_heads: [4, 8, 12]
+      mixer: [['Conv','Conv','Conv','Conv','Conv','Conv'],['Conv','Conv','FGlobal','Global','Global','Global'],['Global','Global','Global','Global','Global','Global']]
+      local_k: [[5, 5], [5, 5], [-1, -1]]
+      sub_k: [[1, 1], [2, 1], [-1, -1]]
+      last_stage: false
+      feat2d: True
+    nb_classes: 97
+    strides: [[4, 4], [1, 1], [2, 1], [1, 1]]
+    k_size: [[2, 2], [1, 1], [2, 1], [1, 1]]
+    q_size: [4, 32]
+    deform_stride: 2
+    stage_idx: 2
+    use_depthwise_unet: True
+    use_more_unet: False
+    binary_loss_type: BanlanceMultiClassCrossEntropyLoss
+    mid_size: True
+    d_embedding: 384
+  Decoder:
+    name: CAMDecoder
+    num_encoder_layers: -1
+    beam_size: 0
+    num_decoder_layers: 2
+    nhead: 8
+    max_len: *max_text_length
+
+Loss:
+  name: CAMLoss
+  loss_weight_binary: 1.5
+  label_smoothing: 0.
+
+Metric:
+  name: RecMetric
+  main_indicator: acc
+  is_filter: True
+
+PostProcess:
+  name: ARLabelDecode
+
+Train:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../Union14M-L-LMDB-Filtered
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - PARSeqAugPIL:
+      - CAMLabelEncode: # Class handling label
+          font_path: ./arial.ttf
+          image_shape: *img_shape
+      - RecTVResize:
+          image_shape: [64, 256]
+          padding: False
+      - KeepKeys:
+          keep_keys: ['image', 'label', 'length', 'binary_mask'] # dataloader will return list in this order
+  loader:
+    shuffle: True
+    batch_size_per_card: 256
+    drop_last: True
+    num_workers: 4
+
+Eval:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../evaluation
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - ARLabelEncode: # Class handling label
+      - RecTVResize:
+          image_shape: [64, 256]
+          padding: False
+      - 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: 2

openocr/configs/rec/cdistnet/resnet45_trans_cdistnet.yml

@@ -0,0 +1,93 @@
+Global:
+  device: gpu
+  epoch_num: 20
+  log_smooth_window: 20
+  print_batch_step: 10
+  output_dir: ./output/rec/u14m_filter/resnet45_trans_cdistnet
+  eval_epoch_step: [0, 1]
+  eval_batch_step: [0, 500]
+  cal_metric_during_train: True
+  pretrained_model:
+  checkpoints:
+  use_tensorboard: false
+  infer_img:
+  # for data or label process
+  character_dict_path: ./tools/utils/EN_symbol_dict.txt
+  max_text_length: 25
+  use_space_char: False
+  save_res_path: ./output/rec/u14m_filter/predicts_resnet45_trans_cdistnet.txt
+  use_amp: True
+  grad_clip_val: 5
+
+Optimizer:
+  name: Adam
+  lr: 0.002 # for 4gpus bs256/gpu
+  weight_decay: 0.0
+  filter_bias_and_bn: False
+
+LRScheduler:
+  name: OneCycleLR
+  warmup_epoch: 1.5 # pct_start 0.075*20 = 1.5ep
+  cycle_momentum: False
+
+Architecture:
+  model_type: rec
+  algorithm: CDistNet
+  Transform:
+  Encoder:
+    name: ResNet45
+    in_channels: 3
+    strides: [2, 1, 2, 1, 1]
+  Decoder:
+    name: CDistNetDecoder
+    add_conv: True
+
+Loss:
+  name: ARLoss
+
+PostProcess:
+  name: ARLabelDecode
+
+Metric:
+  name: RecMetric
+  main_indicator: acc
+  is_filter: True
+
+Train:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../Union14M-L-LMDB-Filtered
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - PARSeqAugPIL:
+      - ARLabelEncode: # Class handling label
+      - RecTVResize:
+          image_shape: [32, 128]
+          padding: False
+      - 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: 4
+
+Eval:
+  dataset:
+    name: LMDBDataSet
+    data_dir: ../evaluation
+    transforms:
+      - DecodeImagePIL: # load image
+          img_mode: RGB
+      - ARLabelEncode: # Class handling label
+      - RecTVResize:
+          image_shape: [32, 128]
+          padding: False
+      - 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: 2