PaddleOCR_training / config.yml
fohra's picture
Upload model weights for training
edf9459 verified
Global:
debug: false
use_gpu: true
epoch_num: 25
log_smooth_window: 20
print_batch_step: 50
save_model_dir: ./output/final_e25
save_epoch_step: 14
eval_batch_step:
- 0
- 697
cal_metric_during_train: true
pretrained_model: ./models/latin_PP-OCRv3_rec_train/best_accuracy
checkpoints: null
save_inference_dir: ./inference/final
use_visualdl: false
infer_img: doc/imgs_words/ch/word_1.jpg
character_dict_path: ppocr/utils/dict/latin_dict.txt
max_text_length: 25
infer_mode: false
use_space_char: true
distributed: true
save_res_path: ./output/rec/predicts_ppocrv3.txt
seed: 42
Optimizer:
name: Adam
beta1: 0.9
beta2: 0.999
lr:
name: Cosine
learning_rate: 0.001
warmup_epoch: 5
regularizer:
name: L2
factor: 3.0e-05
Architecture:
model_type: rec
algorithm: SVTR
Transform: null
Backbone:
name: MobileNetV1Enhance
scale: 0.5
last_conv_stride:
- 1
- 2
last_pool_type: avg
Head:
name: MultiHead
head_list:
- CTCHead:
Neck:
name: svtr
dims: 64
depth: 2
hidden_dims: 120
use_guide: true
Head:
fc_decay: 1.0e-05
- SARHead:
enc_dim: 512
max_text_length: 25
Loss:
name: MultiLoss
loss_config_list:
- CTCLoss: null
- SARLoss: null
PostProcess:
name: CTCLabelDecode
Metric:
name: RecMetric
main_indicator: acc
keep_space: false
Train:
dataset:
name: SimpleDataSet
data_dir: /data/AIDA/iter4/rivikuvat
label_file_list:
- /data/AIDA/iter4/annot/train_kone_best_iter2_14_02_24.txt
- /data/AIDA/iter4/annot/train_kone_best_iter3_04_03_24.txt
- /data/AIDA/iter4/annot/train_sini3_2_19_03_kone_best.txt
- /data/AIDA/iter4/annot/train_kasi_best_iter2_14_2_24.txt
- /data/AIDA/iter4/annot/train_kasi_best_iter3_04_03_24.txt
- /data/AIDA/iter4/annot/train_sini3_2_19_03_kasi_best.txt
- /data/AIDA/iter4/annot/normaali_16k_train.txt
- /data/AIDA/iter4/annot/shaken_train.txt
- /data/AIDA/iter4/annot/noise_k5_train.txt
- /data/AIDA/iter4/annot/noise_k5_sin_train.txt
transforms:
- DecodeImage:
img_mode: BGR
channel_first: false
- RecAug:
reverse_prob: 0.25
noise_prob: 0.25
blur_prob: 0.25
hsv_aug_prob: 0.25
- MultiLabelEncode: null
- SARRecResizeImg:
image_shape:
- 3
- 48
- 48
- 320
width_downsample_ratio: 0.25
mode: v3
- KeepKeys:
keep_keys:
- image
- label_ctc
- label_sar
- length
- valid_ratio
loader:
shuffle: true
batch_size_per_card: 128
drop_last: true
num_workers: 22
Eval:
dataset:
name: SimpleDataSet
data_dir: /data/AIDA/iter4/rivikuvat
label_file_list:
- /data/AIDA/iter4/annot/valid_kone_best_iter2_filttered_14_2_24.txt
- /data/AIDA/iter4/annot/valid_kone_best_iter3_04_03_24.txt
- /data/AIDA/iter4/annot/valid_sini3_2_19_03_kone_best.txt
transforms:
- DecodeImage:
img_mode: BGR
channel_first: false
- MultiLabelEncode: null
- SARRecResizeImg:
image_shape:
- 3
- 48
- 48
- 320
width_downsample_ratio: 0.25
mode: v3
- KeepKeys:
keep_keys:
- image
- label_ctc
- label_sar
- length
- valid_ratio
loader:
shuffle: false
drop_last: false
batch_size_per_card: 128
num_workers: 16
profiler_options: null