forked from open-mmlab/mmocr
/
crnn_academic_dataset.py
138 lines (120 loc) · 3.54 KB
/
crnn_academic_dataset.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
_base_ = []
checkpoint_config = dict(interval=1)
# yapf:disable
log_config = dict(
interval=1,
hooks=[
dict(type='TextLoggerHook')
])
# yapf:enable
dist_params = dict(backend='nccl')
log_level = 'INFO'
load_from = None
resume_from = None
workflow = [('train', 1)]
# model
label_convertor = dict(
type='CTCConvertor', dict_type='DICT36', with_unknown=False, lower=True)
model = dict(
type='CRNNNet',
preprocessor=None,
backbone=dict(type='VeryDeepVgg', leakyRelu=False, input_channels=1),
encoder=None,
decoder=dict(type='CRNNDecoder', in_channels=512, rnn_flag=True),
loss=dict(type='CTCLoss'),
label_convertor=label_convertor,
pretrained=None)
train_cfg = None
test_cfg = None
# optimizer
optimizer = dict(type='Adadelta', lr=1.0)
optimizer_config = dict(grad_clip=None)
# learning policy
lr_config = dict(policy='step', step=[])
total_epochs = 5
# data
img_norm_cfg = dict(mean=[0.5], std=[0.5])
train_pipeline = [
dict(type='LoadImageFromFile', color_type='grayscale'),
dict(
type='ResizeOCR',
height=32,
min_width=100,
max_width=100,
keep_aspect_ratio=False),
dict(type='ToTensorOCR'),
dict(type='NormalizeOCR', **img_norm_cfg),
dict(
type='Collect',
keys=['img'],
meta_keys=[
'filename', 'ori_shape', 'img_shape', 'text', 'valid_ratio'
]),
]
test_pipeline = [
dict(type='LoadImageFromFile', color_type='grayscale'),
dict(
type='ResizeOCR',
height=32,
min_width=4,
max_width=None,
keep_aspect_ratio=True),
dict(type='ToTensorOCR'),
dict(type='NormalizeOCR', **img_norm_cfg),
dict(
type='Collect',
keys=['img'],
meta_keys=['filename', 'ori_shape', 'img_shape', 'valid_ratio']),
]
dataset_type = 'OCRDataset'
train_img_prefix = 'data/mixture/Syn90k/mnt/ramdisk/max/90kDICT32px'
train_ann_file = 'data/mixture/Syn90k/label.lmdb'
train1 = dict(
type=dataset_type,
img_prefix=train_img_prefix,
ann_file=train_ann_file,
loader=dict(
type='LmdbLoader',
repeat=1,
parser=dict(
type='LineStrParser',
keys=['filename', 'text'],
keys_idx=[0, 1],
separator=' ')),
pipeline=train_pipeline,
test_mode=False)
test_prefix = 'data/mixture/'
test_img_prefix1 = test_prefix + 'icdar_2013/'
test_img_prefix2 = test_prefix + 'IIIT5K/'
test_img_prefix3 = test_prefix + 'svt/'
test_ann_file1 = test_prefix + 'icdar_2013/test_label_1015.txt'
test_ann_file2 = test_prefix + 'IIIT5K/test_label.txt'
test_ann_file3 = test_prefix + 'svt/test_label.txt'
test1 = dict(
type=dataset_type,
img_prefix=test_img_prefix1,
ann_file=test_ann_file1,
loader=dict(
type='HardDiskLoader',
repeat=1,
parser=dict(
type='LineStrParser',
keys=['filename', 'text'],
keys_idx=[0, 1],
separator=' ')),
pipeline=test_pipeline,
test_mode=True)
test2 = {key: value for key, value in test1.items()}
test2['img_prefix'] = test_img_prefix2
test2['ann_file'] = test_ann_file2
test3 = {key: value for key, value in test1.items()}
test3['img_prefix'] = test_img_prefix3
test3['ann_file'] = test_ann_file3
data = dict(
samples_per_gpu=64,
workers_per_gpu=4,
train=dict(type='ConcatDataset', datasets=[train1]),
val=dict(type='ConcatDataset', datasets=[test1, test2, test3]),
test=dict(type='ConcatDataset', datasets=[test1, test2, test3]))
evaluation = dict(interval=1, metric='acc')
cudnn_benchmark = True