-
Notifications
You must be signed in to change notification settings - Fork 7.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add rec_sar #3798
Add rec_sar #3798
Changes from 3 commits
ffa9441
8a95b33
69f9bdd
d43688a
ea52f47
ea01960
37cd4ed
dc3bce7
073fad3
8a7d3a2
df4a2f6
f57a6d1
8123688
ae09ef6
d611515
b68c895
e10f71d
38fb2ba
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,99 @@ | ||
Global: | ||
use_gpu: true | ||
epoch_num: 5 | ||
log_smooth_window: 20 | ||
print_batch_step: 20 | ||
save_model_dir: ./sar_rec | ||
save_epoch_step: 1 | ||
# evaluation is run every 2000 iterations | ||
eval_batch_step: [0, 2000] | ||
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/dict90.txt | ||
character_type: ch | ||
max_text_length: 30 | ||
infer_mode: False | ||
use_space_char: False | ||
save_res_path: ./output/rec/predicts_sar.txt | ||
|
||
Optimizer: | ||
name: Adam | ||
beta1: 0.9 | ||
beta2: 0.999 | ||
lr: | ||
name: Piecewise | ||
decay_epochs: [3, 4] | ||
values: [0.001, 0.0001, 0.00001] | ||
regularizer: | ||
name: 'L2' | ||
factor: 0 | ||
|
||
Architecture: | ||
model_type: rec | ||
algorithm: SAR | ||
Transform: | ||
Backbone: | ||
name: ResNet31 | ||
Head: | ||
name: SARHead | ||
|
||
Loss: | ||
name: SARLoss | ||
|
||
PostProcess: | ||
name: SARLabelDecode | ||
|
||
Metric: | ||
name: RecMetric | ||
|
||
|
||
Train: | ||
dataset: | ||
name: SimpleDataSet | ||
delimiter: ' ' | ||
label_file_list: ['/paddle/data/concat_data/icdar_2013_train20.txt', '/paddle/data/concat_data/icdar_2015_train20.txt', '/paddle/data/concat_data/coco_text_train20.txt', '/paddle/data/concat_data/IIIt5k_train20.txt', '/paddle/data/concat_data/SynthAdd_train.txt', '/paddle/data/concat_data/SynthText_train.txt', '/paddle/data/concat_data/Syn90k_train.txt'] | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. 建议把数据路径替换成 train_data/train_list.txt 在文档里说明训练需要用到哪些数据,有什么不同 There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. 那就是把这几个txt合成一个吗? |
||
data_dir: /paddle/data/concat_data/ | ||
ratio_list: 1.0 | ||
transforms: | ||
- DecodeImage: # load image | ||
img_mode: BGR | ||
channel_first: False | ||
- SARLabelEncode: # Class handling label | ||
- SARRecResizeImg: | ||
image_shape: [3, 48, 48, 160] # h:48 w:[48,160] | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. 4维的shape? There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. 后两维是宽度的范围,宽度是变长的 |
||
width_downsample_ratio: 0.25 | ||
- KeepKeys: | ||
keep_keys: ['image', 'label', 'valid_ratio'] # dataloader will return list in this order | ||
loader: | ||
shuffle: True | ||
batch_size_per_card: 64 # 32 | ||
drop_last: True | ||
num_workers: 8 | ||
use_shared_memory: False | ||
|
||
Eval: | ||
dataset: | ||
name: LMDBDataSet | ||
data_dir: /paddle/data/ocr_data/evaluation/ | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. 不要用绝对路径,指向相对路径,让用户可以很方便跑通,参考其他算法的配置文件。上面train同理 There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. 好的,我改一下 |
||
transforms: | ||
- DecodeImage: # load image | ||
img_mode: BGR | ||
channel_first: False | ||
- SARLabelEncode: # Class handling label | ||
- SARRecResizeImg: | ||
image_shape: [3, 48, 48, 160] | ||
width_downsample_ratio: 0.25 | ||
- KeepKeys: | ||
keep_keys: ['image', 'label', 'valid_ratio'] # dataloader will return list in this order | ||
loader: | ||
shuffle: False | ||
drop_last: False | ||
batch_size_per_card: 64 | ||
num_workers: 4 | ||
use_shared_memory: False | ||
|
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,25 @@ | ||
from __future__ import absolute_import | ||
from __future__ import division | ||
from __future__ import print_function | ||
|
||
import paddle | ||
from paddle import nn | ||
|
||
|
||
class SARLoss(nn.Layer): | ||
def __init__(self, **kwargs): | ||
super(SARLoss, self).__init__() | ||
self.loss_func = paddle.nn.loss.CrossEntropyLoss(reduction="mean", ignore_index=92) | ||
|
||
def forward(self, predicts, batch): | ||
predict = predicts[:, :-1, :] # ignore last index of outputs to be in same seq_len with targets | ||
label = batch[1].astype("int64")[:, 1:] # ignore first index of target in loss calculation | ||
batch_size, num_steps, num_classes = predict.shape[0], predict.shape[ | ||
1], predict.shape[2] | ||
assert len(label.shape) == len(list(predict.shape)) - 1, \ | ||
"The target's shape and inputs's shape is [N, d] and [N, num_steps]" | ||
|
||
inputs = paddle.reshape(predict, [-1, num_classes]) | ||
targets = paddle.reshape(label, [-1]) | ||
loss = self.loss_func(inputs, targets) | ||
return {'loss': loss} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
已经有resnet,复用即可
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
resnet31是作者新改的一个网络结构,和常用的不太一样