-
Notifications
You must be signed in to change notification settings - Fork 8
/
abinet_10e_union14m.py
118 lines (109 loc) · 3.52 KB
/
abinet_10e_union14m.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
_base_ = [
'../_base_/datasets/union14m_train.py',
'../_base_/datasets/union14m_benchmark.py',
'../_base_/datasets/cute80.py',
'../_base_/datasets/iiit5k.py',
'../_base_/datasets/svt.py',
'../_base_/datasets/svtp.py',
'../_base_/datasets/icdar2013.py',
'../_base_/datasets/icdar2015.py',
'../_base_/default_runtime.py',
'../_base_/schedules/schedule_adamw_cos_10e.py',
'_base_abinet.py',
]
load_from = 'https://download.openmmlab.com/mmocr/textrecog/abinet/abinet_pretrain-45deac15.pth' # noqa
_base_.pop('model')
dictionary = dict(
type='Dictionary',
dict_file= # noqa
'{{ fileDirname }}/../../../dicts/english_digits_symbols_space.txt',
with_padding=True,
with_unknown=True,
same_start_end=True,
with_start=True,
with_end=True)
model = dict(
type='ABINet',
backbone=dict(type='ResNetABI'),
encoder=dict(
type='ABIEncoder',
n_layers=3,
n_head=8,
d_model=512,
d_inner=2048,
dropout=0.1,
max_len=8 * 32,
),
decoder=dict(
type='ABIFuser',
vision_decoder=dict(
type='ABIVisionDecoder',
in_channels=512,
num_channels=64,
attn_height=8,
attn_width=32,
attn_mode='nearest',
init_cfg=dict(type='Xavier', layer='Conv2d')),
module_loss=dict(type='ABIModuleLoss'),
postprocessor=dict(type='AttentionPostprocessor'),
dictionary=dictionary,
max_seq_len=26,
),
data_preprocessor=dict(
type='TextRecogDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375]))
# dataset settings
train_list = [
_base_.union14m_challenging, _base_.union14m_hard, _base_.union14m_medium,
_base_.union14m_normal, _base_.union14m_easy
]
val_list = [
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test,
_base_.svt_textrecog_test, _base_.svtp_textrecog_test,
_base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
test_list = [
_base_.union14m_benchmark_artistic,
_base_.union14m_benchmark_multi_oriented,
_base_.union14m_benchmark_contextless,
_base_.union14m_benchmark_curve,
_base_.union14m_benchmark_incomplete,
_base_.union14m_benchmark_incomplete_ori,
_base_.union14m_benchmark_multi_words,
_base_.union14m_benchmark_salient,
_base_.union14m_benchmark_general,
]
train_dataset = dict(
type='ConcatDataset', datasets=train_list, pipeline=_base_.train_pipeline)
test_dataset = dict(
type='ConcatDataset', datasets=test_list, pipeline=_base_.test_pipeline)
val_dataset = dict(
type='ConcatDataset', datasets=val_list, pipeline=_base_.test_pipeline)
train_dataloader = dict(
batch_size=128,
num_workers=24,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=train_dataset)
test_dataloader = dict(
batch_size=128,
num_workers=4,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=test_dataset)
val_dataloader = dict(
batch_size=128,
num_workers=4,
persistent_workers=True,
pin_memory=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=val_dataset)
val_evaluator = dict(
dataset_prefixes=['CUTE80', 'IIIT5K', 'SVT', 'SVTP', 'IC13', 'IC15'])
test_evaluator = dict(dataset_prefixes=[
'artistic', 'multi-oriented', 'contextless', 'curve', 'incomplete',
'incomplete-ori', 'multi-words', 'salient', 'general'
])