mirrors
/
mmocr
mirror of https://github.com/open-mmlab/mmocr.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

[Config] Update base dataset configs (#1190)

pull/1193/head
Tong Gao 2022-07-23 21:25:15 +08:00 committed by GitHub
parent ca01ee5eb3
commit abb6c16095
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
6 changed files with 36 additions and 56 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
configs/_base_/recog_datasets/MJ_train.py
View File

@ -1,14 +1,12 @@
# Text Recognition Training set, including:
# Synthetic Datasets: Syn90k
data_root = 'data/rec'
train_img_prefix1 = 'Syn90k/mnt/ramdisk/max/90kDICT32px'
train_ann_file1 = 'Syn90k/label.json'
mj_rec_train = dict(
type='OCRDataset',
data_root=data_root,
data_prefix=dict(img_path=train_img_prefix1),
ann_file=train_ann_file1,
data_root='data/rec',
data_prefix=dict(img_path='Syn90k/mnt/ramdisk/max/90kDICT32px'),
ann_file='Syn90k/train_labels.json',
test_mode=False,
pipeline=None)
train_list = [mj_rec_train]

14
configs/_base_/recog_datasets/ST_MJ_alphanumeric_train.py
View File

@ -3,25 +3,21 @@
# Both annotations are filtered so that
# only alphanumeric terms are left
data_root = 'data/rec'
train_img_prefix1 = 'Syn90k/mnt/ramdisk/max/90kDICT32px'
train_ann_file1 = 'Syn90k/label.json'
mj_rec_train = dict(
type='OCRDataset',
data_root=data_root,
data_prefix=dict(img_path=train_img_prefix1),
ann_file=train_ann_file1,
data_prefix=dict(img_path='Syn90k/mnt/ramdisk/max/90kDICT32px'),
ann_file='Syn90k/train_labels.json',
test_mode=False,
pipeline=None)
train_img_prefix2 = 'SynthText/synthtext/SynthText_patch_horizontal'
train_ann_file2 = 'SynthText/alphanumeric_label.json'
st_rec_train = dict(
type='OCRDataset',
data_root=data_root,
data_prefix=dict(img_path=train_img_prefix2),
ann_file=train_ann_file2,
data_prefix=dict(
img_path='SynthText/synthtext/SynthText_patch_horizontal'),
ann_file='SynthText/alphanumeric_train_labels.json',
test_mode=False,
pipeline=None)

16
configs/_base_/recog_datasets/ST_MJ_train.py
View File

@ -3,26 +3,20 @@
data_root = 'data/rec'
train_img_prefix1 = 'Syn90k/mnt/ramdisk/max/90kDICT32px'
train_ann_file1 = 'Syn90k/label.json'
file_client_args = dict(backend='disk')
mj_rec_train = dict(
type='OCRDataset',
data_root=data_root,
data_prefix=dict(img_path=train_img_prefix1),
ann_file=train_ann_file1,
data_prefix=dict(img_path='Syn90k/mnt/ramdisk/max/90kDICT32px'),
ann_file='Syn90k/train_labels.json',
test_mode=False,
pipeline=None)
train_img_prefix2 = 'SynthText/synthtext/SynthText_patch_horizontal'
train_ann_file2 = 'SynthText/label.json'
st_rec_train = dict(
type='OCRDataset',
data_root=data_root,
data_prefix=dict(img_path=train_img_prefix2),
ann_file=train_ann_file2,
data_prefix=dict(
img_path='SynthText/synthtext/SynthText_patch_horizontal'),
ann_file='SynthText/train_labels.json',
test_mode=False,
pipeline=None)

18
configs/_base_/recog_datasets/ST_SA_MJ_real_train.py
View File

@ -8,18 +8,18 @@ train_img_prefix2 = 'icdar_2013'
train_img_prefix3 = 'icdar_2015'
train_img_prefix4 = 'coco_text'
train_img_prefix5 = 'IIIT5K'
train_img_prefix6 = 'SynthText_Add'
train_img_prefix6 = 'synthtext_add'
train_img_prefix7 = 'SynthText/synthtext/SynthText_patch_horizontal'
train_img_prefix8 = 'Syn90k/mnt/ramdisk/max/90kDICT32px'
train_ann_file1 = 'icdar_2011/train_label.json',
train_ann_file2 = 'icdar_2013/train_label.json',
train_ann_file3 = 'icdar_2015/train_label.json',
train_ann_file4 = 'coco_text/train_label.json',
train_ann_file5 = 'IIIT5K/train_label.json',
train_ann_file6 = 'SynthText_Add/train_label.json',
train_ann_file7 = 'SynthText/shuffle_label.json',
train_ann_file8 = 'Syn90k/mnt/ramdisk/max/90kDICT32px/shuffle_label.json'
train_ann_file1 = 'icdar_2011/train_labels.json',
train_ann_file2 = 'icdar_2013/train_labels.json',
train_ann_file3 = 'icdar_2015/train_labels.json',
train_ann_file4 = 'coco_text_v1/train_labels.json',
train_ann_file5 = 'IIIT5K/train_labels.json',
train_ann_file6 = 'synthtext_add/train_labels.json',
train_ann_file7 = 'SynthText/shuffle_train_labels.json',
train_ann_file8 = 'Syn90k/shuffle_train_labels.json'
ic11_rec_train = dict(
type='OCRDataset',

22
configs/_base_/recog_datasets/ST_SA_MJ_train.py
View File

@ -1,37 +1,29 @@
# Text Recognition Training set, including:
# Synthetic Datasets: SynthText, Syn90k
data_root = 'data/rec'
train_img_prefix1 = 'Syn90k/mnt/ramdisk/max/90kDICT32px'
train_ann_file1 = 'Syn90k/label.json'
mj_rec_train = dict(
type='OCRDataset',
data_root=data_root,
data_prefix=dict(img_path=train_img_prefix1),
ann_file=train_ann_file1,
data_prefix=dict(img_path='Syn90k/mnt/ramdisk/max/90kDICT32px'),
ann_file='Syn90k/train_labels.json',
test_mode=False,
pipeline=None)
train_img_prefix2 = 'SynthText/synthtext/' + \
'SynthText_patch_horizontal'
train_ann_file2 = 'SynthText/label.json',
st_rec_train = dict(
type='OCRDataset',
data_root=data_root,
data_prefix=dict(img_path=train_img_prefix2),
ann_file=train_ann_file2,
data_prefix=dict(
img_path='SynthText/synthtext/SynthText_patch_horizontal'),
ann_file='SynthText/train_labels.json',
test_mode=False,
pipeline=None)
train_img_prefix3 = 'SynthText_Add'
train_ann_file3 = 'SynthText_Add/label.json'
st_add_rec_train = dict(
type='OCRDataset',
data_root=data_root,
data_prefix=dict(img_path=train_img_prefix3),
ann_file=train_ann_file3,
data_prefix=dict(img_path='synthtext_add'),
ann_file='synthtext_add/train_labels.json',
test_mode=False,
pipeline=None)

12
configs/_base_/recog_datasets/academic_test.py
View File

@ -11,12 +11,12 @@ test_img_prefix4 = 'icdar_2015/ch4_test_word_images_gt'
test_img_prefix5 = 'svtp/'
test_img_prefix6 = 'ct80/'
test_ann_file1 = 'IIIT5K/test_label.json'
test_ann_file2 = 'svt/test_label.json'
test_ann_file3 = 'icdar_2013/test_label.json'
test_ann_file4 = 'icdar_2015/test_label.json'
test_ann_file5 = 'svtp/test_label.json'
test_ann_file6 = 'ct80/test_label.json'
test_ann_file1 = 'IIIT5K/test_labels.json'
test_ann_file2 = 'svt/test_labels.json'
test_ann_file3 = 'icdar_2013/test_labels.json'
test_ann_file4 = 'icdar_2015/test_labels.json'
test_ann_file5 = 'svtp/test_labels.json'
test_ann_file6 = 'ct80/test_labels.json'
iiit5k_rec_test = dict(
type='OCRDataset',