OpenOCR-Demo / configs /rec /parseq /vit_parseq.yml
topdu's picture
openocr demo
29f689c
raw
history blame
2.44 kB
Global:
device: gpu
epoch_num: 20
log_smooth_window: 20
print_batch_step: 10
output_dir: ./output/rec/u14m_filter/vit_parseq/
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: &character_dict_path ./tools/utils/EN_symbol_dict.txt
max_text_length: &max_text_length 25
use_space_char: &use_space_char False
use_amp: True
save_res_path: ./output/rec/u14m_filter/predicts_vit_parseq.txt
grad_clip_val: 20
Optimizer:
name: AdamW
lr: 0.001485 # 2gpus 384bs/gpu
weight_decay: 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: PARSeq
Transform:
Encoder:
name: ViT
Decoder:
name: PARSeqDecoder
decode_ar: True
refine_iters: 1
Loss:
name: PARSeqLoss
PostProcess:
name: ARLabelDecode
character_dict_path: *character_dict_path
use_space_char: *use_space_char
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
- PARSeqAug:
- ARLabelEncode: # Class handling label
character_dict_path: *character_dict_path
use_space_char: *use_space_char
max_text_length: *max_text_length
- 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: 384
drop_last: True
num_workers: 4
Eval:
dataset:
name: LMDBDataSet
data_dir: ../evaluation/
transforms:
- DecodeImagePIL: # load image
img_mode: RGB
- ARLabelEncode: # Class handling label
character_dict_path: *character_dict_path
use_space_char: *use_space_char
max_text_length: *max_text_length
- 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