PaddleOCR/configs/rec/rec_mv3_tps_bilstm_att.yml

103 lines
2.2 KiB
YAML
Raw Normal View History

2021-01-29 11:15:03 +08:00
Global:
2021-02-02 19:28:29 +08:00
use_gpu: True
2021-01-29 11:15:03 +08:00
epoch_num: 72
log_smooth_window: 20
print_batch_step: 10
save_model_dir: ./output/rec/rec_mv3_tps_bilstm_att/
save_epoch_step: 3
# evaluation is run every 5000 iterations after the 4000th iteration
eval_batch_step: [0, 2000]
cal_metric_during_train: True
pretrained_model:
checkpoints:
save_inference_dir:
use_visualdl: False
infer_img: doc/imgs_words/ch/word_1.jpg
# for data or label process
character_dict_path:
character_type: en
max_text_length: 25
infer_mode: False
use_space_char: False
2021-04-25 20:49:45 +08:00
save_res_path: ./output/rec/predicts_mv3_tps_bilstm_att.txt
2021-01-29 11:15:03 +08:00
Optimizer:
name: Adam
beta1: 0.9
beta2: 0.999
lr:
learning_rate: 0.0005
regularizer:
name: 'L2'
factor: 0.00001
Architecture:
model_type: rec
algorithm: RARE
Transform:
name: TPS
num_fiducial: 20
loc_lr: 0.1
model_name: small
Backbone:
name: MobileNetV3
scale: 0.5
model_name: large
Neck:
name: SequenceEncoder
encoder_type: rnn
hidden_size: 96
Head:
name: AttentionHead
hidden_size: 96
Loss:
name: AttentionLoss
PostProcess:
name: AttnLabelDecode
Metric:
name: RecMetric
main_indicator: acc
Train:
dataset:
2021-02-02 19:28:29 +08:00
name: LMDBDataSet
2021-02-05 16:22:32 +08:00
data_dir: ./train_data/data_lmdb_release/training/
2021-01-29 11:15:03 +08:00
transforms:
- DecodeImage: # load image
img_mode: BGR
channel_first: False
- AttnLabelEncode: # Class handling label
- RecResizeImg:
image_shape: [3, 32, 100]
- KeepKeys:
keep_keys: ['image', 'label', 'length'] # dataloader will return list in this order
loader:
shuffle: True
batch_size_per_card: 256
drop_last: True
num_workers: 8
Eval:
dataset:
2021-02-02 19:28:29 +08:00
name: LMDBDataSet
2021-02-05 16:22:32 +08:00
data_dir: ./train_data/data_lmdb_release/validation/
2021-01-29 11:15:03 +08:00
transforms:
- DecodeImage: # load image
img_mode: BGR
channel_first: False
- AttnLabelEncode: # Class handling label
- RecResizeImg:
image_shape: [3, 32, 100]
- KeepKeys:
keep_keys: ['image', 'label', 'length'] # dataloader will return list in this order
loader:
shuffle: False
drop_last: False
batch_size_per_card: 256
num_workers: 1