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

[Config] rename base dataset terms to {dataset-name}_task_train/test (#1541)

pull/1546/head
liukuikun 2022-11-17 10:15:33 +08:00 committed by GitHub
parent b8e395ed71
commit d8c0df4827
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
53 changed files with 291 additions and 275 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
configs/textdet/_base_/datasets/ctw1500.py
View File

@ -1,16 +1,16 @@
ctw_det_data_root = 'data/det/ctw1500'
ctw1500_textdet_data_root = 'data/det/ctw1500'
ctw_det_train = dict(
ctw1500_textdet_train = dict(
type='OCRDataset',
data_root=ctw_det_data_root,
data_root=ctw1500_textdet_data_root,
ann_file='instances_training.json',
data_prefix=dict(img_path='imgs/'),
filter_cfg=dict(filter_empty_gt=True, min_size=32),
pipeline=None)
ctw_det_test = dict(
ctw1500_textdet_test = dict(
type='OCRDataset',
data_root=ctw_det_data_root,
data_root=ctw1500_textdet_data_root,
ann_file='instances_test.json',
data_prefix=dict(img_path='imgs/'),
test_mode=True,

10
configs/textdet/_base_/datasets/icdar2015.py
View File

@ -1,16 +1,16 @@
ic15_det_data_root = 'data/det/icdar2015'
icdar2015_textdet_data_root = 'data/det/icdar2015'
ic15_det_train = dict(
icdar2015_textdet_train = dict(
type='OCRDataset',
data_root=ic15_det_data_root,
data_root=icdar2015_textdet_data_root,
ann_file='instances_training.json',
data_prefix=dict(img_path='imgs/'),
filter_cfg=dict(filter_empty_gt=True, min_size=32),
pipeline=None)
ic15_det_test = dict(
icdar2015_textdet_test = dict(
type='OCRDataset',
data_root=ic15_det_data_root,
data_root=icdar2015_textdet_data_root,
ann_file='instances_test.json',
data_prefix=dict(img_path='imgs/'),
test_mode=True,

10
configs/textdet/_base_/datasets/icdar2017.py
View File

@ -1,16 +1,16 @@
ic17_det_data_root = 'data/det/icdar_2017'
icdar2017_textdet_data_root = 'data/det/icdar_2017'
ic17_det_train = dict(
icdar2017_textdet_train = dict(
type='OCRDataset',
data_root=ic17_det_data_root,
data_root=icdar2017_textdet_data_root,
ann_file='instances_training.json',
data_prefix=dict(img_path='imgs/'),
filter_cfg=dict(filter_empty_gt=True, min_size=32),
pipeline=None)
ic17_det_test = dict(
icdar2017_textdet_test = dict(
type='OCRDataset',
data_root=ic17_det_data_root,
data_root=icdar2017_textdet_data_root,
ann_file='instances_test.json',
data_prefix=dict(img_path='imgs/'),
test_mode=True,

10
configs/textdet/_base_/datasets/synthtext.py
View File

@ -1,16 +1,16 @@
st_det_data_root = 'data/det/synthtext'
synthtext_textdet_data_root = 'data/det/synthtext'
st_det_train = dict(
synthtext_textdet_train = dict(
type='OCRDataset',
data_root=st_det_data_root,
data_root=synthtext_textdet_data_root,
ann_file='instances_training.json',
data_prefix=dict(img_path='imgs/'),
filter_cfg=dict(filter_empty_gt=True, min_size=32),
pipeline=None)
st_det_test = dict(
synthtext_textdet_test = dict(
type='OCRDataset',
data_root=st_det_data_root,
data_root=synthtext_textdet_data_root,
ann_file='instances_test.json',
data_prefix=dict(img_path='imgs/'),
test_mode=True,

10
configs/textdet/_base_/datasets/totaltext.py
View File

@ -1,15 +1,15 @@
tt_det_data_root = 'data/totaltext'
totaltext_textdet_data_root = 'data/totaltext'
tt_det_train = dict(
totaltext_textdet_train = dict(
type='OCRDataset',
data_root=tt_det_data_root,
data_root=totaltext_textdet_data_root,
ann_file='textdet_train.json',
filter_cfg=dict(filter_empty_gt=True, min_size=32),
pipeline=None)
tt_det_test = dict(
totaltext_textdet_test = dict(
type='OCRDataset',
data_root=tt_det_data_root,
data_root=totaltext_textdet_data_root,
ann_file='textdet_test.json',
test_mode=True,
pipeline=None)

12
configs/textdet/dbnet/dbnet_resnet18_fpnc_100k_synthtext.py
View File

@ -6,24 +6,24 @@ _base_ = [
]
# dataset settings
st_det_train = _base_.st_det_train
st_det_train.pipeline = _base_.train_pipeline
st_det_test = _base_.st_det_test
st_det_test.pipeline = _base_.test_pipeline
synthtext_textdet_train = _base_.synthtext_textdet_train
synthtext_textdet_train.pipeline = _base_.train_pipeline
synthtext_textdet_test = _base_.synthtext_textdet_test
synthtext_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=16,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=st_det_train)
dataset=synthtext_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=st_det_test)
dataset=synthtext_textdet_test)
test_dataloader = val_dataloader

12
configs/textdet/dbnet/dbnet_resnet18_fpnc_1200e_icdar2015.py
View File

@ -6,24 +6,24 @@ _base_ = [
]
# dataset settings
ic15_det_train = _base_.ic15_det_train
ic15_det_train.pipeline = _base_.train_pipeline
ic15_det_test = _base_.ic15_det_test
ic15_det_test.pipeline = _base_.test_pipeline
icdar2015_textdet_train = _base_.icdar2015_textdet_train
icdar2015_textdet_train.pipeline = _base_.train_pipeline
icdar2015_textdet_test = _base_.icdar2015_textdet_test
icdar2015_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=16,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ic15_det_train)
dataset=icdar2015_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ic15_det_test)
dataset=icdar2015_textdet_test)
test_dataloader = val_dataloader

12
configs/textdet/dbnet/dbnet_resnet50-dcnv2_fpnc_100k_synthtext.py
View File

@ -6,24 +6,24 @@ _base_ = [
]
# dataset settings
st_det_train = _base_.st_det_train
st_det_train.pipeline = _base_.train_pipeline
st_det_test = _base_.st_det_test
st_det_test.pipeline = _base_.test_pipeline
synthtext_textdet_train = _base_.synthtext_textdet_train
synthtext_textdet_train.pipeline = _base_.train_pipeline
synthtext_textdet_test = _base_.synthtext_textdet_test
synthtext_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=16,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=st_det_train)
dataset=synthtext_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=st_det_test)
dataset=synthtext_textdet_test)
test_dataloader = val_dataloader

12
configs/textdet/dbnet/dbnet_resnet50-dcnv2_fpnc_1200e_icdar2015.py
View File

@ -9,24 +9,24 @@ _base_ = [
load_from = 'https://download.openmmlab.com/mmocr/textdet/dbnet/tmp_1.0_pretrain/dbnet_r50dcnv2_fpnc_sbn_2e_synthtext_20210325-ed322016.pth' # noqa
# dataset settings
ic15_det_train = _base_.ic15_det_train
ic15_det_train.pipeline = _base_.train_pipeline
ic15_det_test = _base_.ic15_det_test
ic15_det_test.pipeline = _base_.test_pipeline
icdar2015_textdet_train = _base_.icdar2015_textdet_train
icdar2015_textdet_train.pipeline = _base_.train_pipeline
icdar2015_textdet_test = _base_.icdar2015_textdet_test
icdar2015_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=16,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ic15_det_train)
dataset=icdar2015_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ic15_det_test)
dataset=icdar2015_textdet_test)
test_dataloader = val_dataloader

4
configs/textdet/dbnetpp/dbnetpp_resnet50-dcnv2_fpnc_100k_synthtext.py
View File

@ -6,8 +6,8 @@ _base_ = [
]
# dataset settings
train_list = [_base_.st_det_train]
test_list = [_base_.st_det_test]
train_list = [_base_.synthtext_textdet_train]
test_list = [_base_.synthtext_textdet_test]
train_dataloader = dict(
batch_size=16,

4
configs/textdet/dbnetpp/dbnetpp_resnet50-dcnv2_fpnc_1200e_icdar2015.py
View File

@ -8,8 +8,8 @@ _base_ = [
load_from = 'https://download.openmmlab.com/mmocr/textdet/dbnetpp/tmp_1.0_pretrain/dbnetpp_r50dcnv2_fpnc_100k_iter_synthtext-20220502-352fec8a.pth' # noqa
# dataset settings
train_list = [_base_.ic15_det_train]
test_list = [_base_.ic15_det_test]
train_list = [_base_.icdar2015_textdet_train]
test_list = [_base_.icdar2015_textdet_test]
train_dataloader = dict(
batch_size=16,

12
configs/textdet/drrg/drrg_resnet50_fpn-unet_1200e_ctw1500.py
View File

@ -6,24 +6,24 @@ _base_ = [
]
# dataset settings
ctw_det_train = _base_.ctw_det_train
ctw_det_train.pipeline = _base_.train_pipeline
ctw_det_test = _base_.ctw_det_test
ctw_det_test.pipeline = _base_.test_pipeline
ctw1500_textdet_train = _base_.ctw1500_textdet_train
ctw1500_textdet_train.pipeline = _base_.train_pipeline
ctw1500_textdet_test = _base_.ctw1500_textdet_test
ctw1500_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=4,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ctw_det_train)
dataset=ctw1500_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=1,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ctw_det_test)
dataset=ctw1500_textdet_test)
test_dataloader = val_dataloader

12
configs/textdet/fcenet/fcenet_resnet50-dcnv2_fpn_1500e_ctw1500.py
View File

@ -14,8 +14,8 @@ param_scheduler = [
file_client_args = dict(backend='disk')
# dataset settings
ctw_det_train = _base_.ctw_det_train
ctw_det_test = _base_.ctw_det_test
ctw1500_textdet_train = _base_.ctw1500_textdet_train
ctw1500_textdet_test = _base_.ctw1500_textdet_test
# test pipeline for CTW1500
ctw_test_pipeline = [
@ -36,22 +36,22 @@ ctw_test_pipeline = [
meta_keys=('img_path', 'ori_shape', 'img_shape', 'scale_factor'))
]
ctw_det_train.pipeline = _base_.train_pipeline
ctw_det_test.pipeline = ctw_test_pipeline
ctw1500_textdet_train.pipeline = _base_.train_pipeline
ctw1500_textdet_test.pipeline = ctw_test_pipeline
train_dataloader = dict(
batch_size=8,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ctw_det_train)
dataset=ctw1500_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=1,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ctw_det_test)
dataset=ctw1500_textdet_test)
test_dataloader = val_dataloader

12
configs/textdet/fcenet/fcenet_resnet50_fpn_1500e_icdar2015.py
View File

@ -13,24 +13,24 @@ param_scheduler = [
]
# dataset settings
ic15_det_train = _base_.ic15_det_train
ic15_det_test = _base_.ic15_det_test
ic15_det_train.pipeline = _base_.train_pipeline
ic15_det_test.pipeline = _base_.test_pipeline
icdar2015_textdet_train = _base_.icdar2015_textdet_train
icdar2015_textdet_test = _base_.icdar2015_textdet_test
icdar2015_textdet_train.pipeline = _base_.train_pipeline
icdar2015_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=8,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ic15_det_train)
dataset=icdar2015_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=1,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ic15_det_test)
dataset=icdar2015_textdet_test)
test_dataloader = val_dataloader

12
configs/textdet/maskrcnn/mask-rcnn_resnet50_fpn_160e_ctw1500.py
View File

@ -15,8 +15,8 @@ param_scheduler = [
]
# dataset settings
ctw_det_train = _base_.ctw_det_train
ctw_det_test = _base_.ctw_det_test
ctw1500_textdet_train = _base_.ctw1500_textdet_train
ctw1500_textdet_test = _base_.ctw1500_textdet_test
# test pipeline for CTW1500
ctw_test_pipeline = [
@ -37,22 +37,22 @@ ctw_test_pipeline = [
meta_keys=('img_path', 'ori_shape', 'img_shape', 'scale_factor'))
]
ctw_det_train.pipeline = _base_.train_pipeline
ctw_det_test.pipeline = ctw_test_pipeline
ctw1500_textdet_train.pipeline = _base_.train_pipeline
ctw1500_textdet_test.pipeline = ctw_test_pipeline
train_dataloader = dict(
batch_size=8,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ctw_det_train)
dataset=ctw1500_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=1,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ctw_det_test)
dataset=ctw1500_textdet_test)
test_dataloader = val_dataloader

12
configs/textdet/maskrcnn/mask-rcnn_resnet50_fpn_160e_icdar2015.py
View File

@ -15,24 +15,24 @@ param_scheduler = [
]
# dataset settings
ic15_det_train = _base_.ic15_det_train
ic15_det_test = _base_.ic15_det_test
ic15_det_train.pipeline = _base_.train_pipeline
ic15_det_test.pipeline = _base_.test_pipeline
icdar2015_textdet_train = _base_.icdar2015_textdet_train
icdar2015_textdet_test = _base_.icdar2015_textdet_test
icdar2015_textdet_train.pipeline = _base_.train_pipeline
icdar2015_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=8,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ic15_det_train)
dataset=icdar2015_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=1,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ic15_det_test)
dataset=icdar2015_textdet_test)
test_dataloader = val_dataloader

12
configs/textdet/maskrcnn/mask-rcnn_resnet50_fpn_160e_icdar2017.py
View File

@ -3,12 +3,12 @@ _base_ = [
'../_base_/datasets/icdar2017.py',
]
ic17_det_train = _base_.ic17_det_train
ic17_det_test = _base_.ic17_det_test
icdar2017_textdet_train = _base_.icdar2017_textdet_train
icdar2017_textdet_test = _base_.icdar2017_textdet_test
# use the same pipeline as icdar2015
ic17_det_train.pipeline = _base_.train_pipeline
ic17_det_test.pipeline = _base_.test_pipeline
icdar2017_textdet_train.pipeline = _base_.train_pipeline
icdar2017_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(dataset=ic17_det_train)
val_dataloader = dict(dataset=ic17_det_test)
train_dataloader = dict(dataset=icdar2017_textdet_train)
val_dataloader = dict(dataset=icdar2017_textdet_test)
test_dataloader = val_dataloader

12
configs/textdet/panet/panet_resnet18_fpem-ffm_600e_ctw1500.py
View File

@ -59,24 +59,24 @@ test_pipeline = [
]
# dataset settings
ctw_det_train = _base_.ctw_det_train
ctw_det_test = _base_.ctw_det_test
ctw1500_textdet_train = _base_.ctw1500_textdet_train
ctw1500_textdet_test = _base_.ctw1500_textdet_test
# pipeline settings
ctw_det_train.pipeline = train_pipeline
ctw_det_test.pipeline = test_pipeline
ctw1500_textdet_train.pipeline = train_pipeline
ctw1500_textdet_test.pipeline = test_pipeline
train_dataloader = dict(
batch_size=16,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ctw_det_train)
dataset=ctw1500_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ctw_det_test)
dataset=ctw1500_textdet_test)
test_dataloader = val_dataloader
val_evaluator = dict(

12
configs/textdet/panet/panet_resnet18_fpem-ffm_600e_icdar2015.py
View File

@ -8,24 +8,24 @@ _base_ = [
default_hooks = dict(checkpoint=dict(type='CheckpointHook', interval=20), )
# dataset settings
ic15_det_train = _base_.ic15_det_train
ic15_det_test = _base_.ic15_det_test
icdar2015_textdet_train = _base_.icdar2015_textdet_train
icdar2015_textdet_test = _base_.icdar2015_textdet_test
# pipeline settings
ic15_det_train.pipeline = _base_.train_pipeline
ic15_det_test.pipeline = _base_.test_pipeline
icdar2015_textdet_train.pipeline = _base_.train_pipeline
icdar2015_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=64,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ic15_det_train)
dataset=icdar2015_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ic15_det_test)
dataset=icdar2015_textdet_test)
test_dataloader = val_dataloader
val_evaluator = dict(

12
configs/textdet/panet/panet_resnet50_fpem-ffm_600e_icdar2017.py
View File

@ -55,23 +55,23 @@ test_pipeline = [
type='PackTextDetInputs',
meta_keys=('img_path', 'ori_shape', 'img_shape', 'scale_factor'))
]
ic17_det_train = _base_.ic17_det_train
ic17_det_test = _base_.ic17_det_test
icdar2017_textdet_train = _base_.icdar2017_textdet_train
icdar2017_textdet_test = _base_.icdar2017_textdet_test
# pipeline settings
ic17_det_train.pipeline = train_pipeline
ic17_det_test.pipeline = test_pipeline
icdar2017_textdet_train.pipeline = train_pipeline
icdar2017_textdet_test.pipeline = test_pipeline
train_dataloader = dict(
batch_size=64,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ic17_det_train)
dataset=icdar2017_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ic17_det_test)
dataset=icdar2017_textdet_test)
test_dataloader = val_dataloader
val_evaluator = dict(

12
configs/textdet/psenet/psenet_resnet50_fpnf_600e_ctw1500.py
View File

@ -13,8 +13,8 @@ param_scheduler = [
]
# dataset settings
ctw_det_train = _base_.ctw_det_train
ctw_det_test = _base_.ctw_det_test
ctw1500_textdet_train = _base_.ctw1500_textdet_train
ctw1500_textdet_test = _base_.ctw1500_textdet_test
test_pipeline_ctw = [
dict(
@ -33,22 +33,22 @@ test_pipeline_ctw = [
]
# pipeline settings
ctw_det_train.pipeline = _base_.train_pipeline
ctw_det_test.pipeline = test_pipeline_ctw
ctw1500_textdet_train.pipeline = _base_.train_pipeline
ctw1500_textdet_test.pipeline = test_pipeline_ctw
train_dataloader = dict(
batch_size=16,
num_workers=8,
persistent_workers=False,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ctw_det_train)
dataset=ctw1500_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=1,
persistent_workers=False,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ctw_det_test)
dataset=ctw1500_textdet_test)
test_dataloader = val_dataloader

12
configs/textdet/psenet/psenet_resnet50_fpnf_600e_icdar2015.py
View File

@ -13,8 +13,8 @@ param_scheduler = [
]
# dataset settings
ic15_det_train = _base_.ic15_det_train
ic15_det_test = _base_.ic15_det_test
icdar2015_textdet_train = _base_.icdar2015_textdet_train
icdar2015_textdet_test = _base_.icdar2015_textdet_test
# use quadrilaterals for icdar2015
model = dict(
@ -22,22 +22,22 @@ model = dict(
det_head=dict(postprocessor=dict(text_repr_type='quad')))
# pipeline settings
ic15_det_train.pipeline = _base_.train_pipeline
ic15_det_test.pipeline = _base_.test_pipeline
icdar2015_textdet_train.pipeline = _base_.train_pipeline
icdar2015_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=16,
num_workers=8,
persistent_workers=False,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ic15_det_train)
dataset=icdar2015_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=1,
persistent_workers=False,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ic15_det_test)
dataset=icdar2015_textdet_test)
test_dataloader = val_dataloader

12
configs/textdet/psenet/psenet_resnet50_fpnf_600e_icdar2017.py
View File

@ -3,14 +3,14 @@ _base_ = [
'../_base_/datasets/icdar2017.py',
]
ic17_det_train = _base_.ic17_det_train
ic17_det_test = _base_.ic17_det_test
icdar2017_textdet_train = _base_.icdar2017_textdet_train
icdar2017_textdet_test = _base_.icdar2017_textdet_test
# use the same pipeline as icdar2015
ic17_det_train.pipeline = _base_.train_pipeline
ic17_det_test.pipeline = _base_.test_pipeline
icdar2017_textdet_train.pipeline = _base_.train_pipeline
icdar2017_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(dataset=ic17_det_train)
val_dataloader = dict(dataset=ic17_det_test)
train_dataloader = dict(dataset=icdar2017_textdet_train)
val_dataloader = dict(dataset=icdar2017_textdet_test)
test_dataloader = val_dataloader
auto_scale_lr = dict(base_batch_size=64 * 4)

12
configs/textdet/textsnake/textsnake_resnet50_fpn-unet_1200e_ctw1500.py
View File

@ -6,24 +6,24 @@ _base_ = [
]
# dataset settings
ctw_det_train = _base_.ctw_det_train
ctw_det_train.pipeline = _base_.train_pipeline
ctw_det_test = _base_.ctw_det_test
ctw_det_test.pipeline = _base_.test_pipeline
ctw1500_textdet_train = _base_.ctw1500_textdet_train
ctw1500_textdet_train.pipeline = _base_.train_pipeline
ctw1500_textdet_test = _base_.ctw1500_textdet_test
ctw1500_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=4,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ctw_det_train)
dataset=ctw1500_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=1,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ctw_det_test)
dataset=ctw1500_textdet_test)
test_dataloader = val_dataloader

6
configs/textrecog/_base_/datasets/coco_text_v1.py
View File

@ -1,8 +1,8 @@
cocov1_rec_train_data_root = 'data/rec/coco_text_v1'
cocotextv1_textrecog_data_root = 'data/rec/coco_text_v1'
cocov1_rec_train = dict(
cocotextv1_textrecog_train = dict(
type='OCRDataset',
data_root=cocov1_rec_train_data_root,
data_root=cocotextv1_textrecog_data_root,
ann_file='train_labels.json',
test_mode=False,
pipeline=None)

6
configs/textrecog/_base_/datasets/cute80.py
View File

@ -1,8 +1,8 @@
cute80_rec_data_root = 'data/rec/ct80/'
cute80_textrecog_data_root = 'data/rec/ct80/'
cute80_rec_test = dict(
cute80_textrecog_test = dict(
type='OCRDataset',
data_root=cute80_rec_data_root,
data_root=cute80_textrecog_data_root,
ann_file='test_labels.json',
test_mode=True,
pipeline=None)

6
configs/textrecog/_base_/datasets/icdar2011.py
View File

@ -1,8 +1,8 @@
ic11_rec_data_root = 'data/rec/icdar_2011/'
icdar2011_textrecog_data_root = 'data/rec/icdar_2011/'
ic11_rec_train = dict(
icdar2011_textrecog_train = dict(
type='OCRDataset',
data_root=ic11_rec_data_root,
data_root=icdar2011_textrecog_data_root,
ann_file='train_labels.json',
test_mode=False,
pipeline=None)

10
configs/textrecog/_base_/datasets/icdar2013.py
View File

@ -1,15 +1,15 @@
ic13_rec_data_root = 'data/rec/icdar_2013/'
icdar2013_textrecog_data_root = 'data/rec/icdar_2013/'
ic13_rec_train = dict(
icdar2013_textrecog_train = dict(
type='OCRDataset',
data_root=ic13_rec_data_root,
data_root=icdar2013_textrecog_data_root,
ann_file='train_labels.json',
test_mode=False,
pipeline=None)
ic13_rec_test = dict(
icdar2013_textrecog_test = dict(
type='OCRDataset',
data_root=ic13_rec_data_root,
data_root=icdar2013_textrecog_data_root,
ann_file='test_labels.json',
test_mode=True,
pipeline=None)

10
configs/textrecog/_base_/datasets/icdar2015.py
View File

@ -1,15 +1,15 @@
ic15_rec_data_root = 'data/rec/icdar_2015/'
icdar2015_textrecog_data_root = 'data/rec/icdar_2015/'
ic15_rec_train = dict(
icdar2015_textrecog_train = dict(
type='OCRDataset',
data_root=ic15_rec_data_root,
data_root=icdar2015_textrecog_data_root,
ann_file='train_labels.json',
test_mode=False,
pipeline=None)
ic15_rec_test = dict(
icdar2015_textrecog_test = dict(
type='OCRDataset',
data_root=ic15_rec_data_root,
data_root=icdar2015_textrecog_data_root,
ann_file='test_labels.json',
test_mode=True,
pipeline=None)

10
configs/textrecog/_base_/datasets/iiit5k.py
View File

@ -1,15 +1,15 @@
iiit5k_rec_data_root = 'data/rec/IIIT5K/'
iiit5k_textrecog_data_root = 'data/rec/IIIT5K/'
iiit5k_rec_train = dict(
iiit5k_textrecog_train = dict(
type='OCRDataset',
data_root=iiit5k_rec_data_root,
data_root=iiit5k_textrecog_data_root,
ann_file='train_labels.json',
test_mode=False,
pipeline=None)
iiit5k_rec_test = dict(
iiit5k_textrecog_test = dict(
type='OCRDataset',
data_root=iiit5k_rec_data_root,
data_root=iiit5k_textrecog_data_root,
ann_file='test_labels.json',
test_mode=True,
pipeline=None)

10
configs/textrecog/_base_/datasets/mjsynth.py
View File

@ -1,16 +1,16 @@
mj_rec_data_root = 'data/rec/Syn90k/'
mjsynth_textrecog_data_root = 'data/rec/Syn90k/'
mj_rec_train = dict(
mjsynth_textrecog_test = dict(
type='OCRDataset',
data_root=mj_rec_data_root,
data_root=mjsynth_textrecog_data_root,
data_prefix=dict(img_path='mnt/ramdisk/max/90kDICT32px'),
ann_file='train_labels.json',
test_mode=False,
pipeline=None)
mj_sub_rec_train = dict(
mjsynth_sub_textrecog_train = dict(
type='OCRDataset',
data_root=mj_rec_data_root,
data_root=mjsynth_textrecog_data_root,
data_prefix=dict(img_path='mnt/ramdisk/max/90kDICT32px'),
ann_file='subset_train_labels.json',
test_mode=False,

6
configs/textrecog/_base_/datasets/svt.py
View File

@ -1,8 +1,8 @@
svt_rec_data_root = 'data/rec/svt/'
svt_textrecog_data_root = 'data/rec/svt/'
svt_rec_test = dict(
svt_textrecog_test = dict(
type='OCRDataset',
data_root=svt_rec_data_root,
data_root=svt_textrecog_data_root,
ann_file='test_labels.json',
test_mode=True,
pipeline=None)

6
configs/textrecog/_base_/datasets/svtp.py
View File

@ -1,8 +1,8 @@
svtp_rec_data_root = 'data/rec/svtp/'
svtp_textrecog_data_root = 'data/rec/svtp/'
svtp_rec_test = dict(
svtp_textrecog_test = dict(
type='OCRDataset',
data_root=svtp_rec_data_root,
data_root=svtp_textrecog_data_root,
ann_file='test_labels.json',
test_mode=True,
pipeline=None)

14
configs/textrecog/_base_/datasets/synthtext.py
View File

@ -1,24 +1,24 @@
st_data_root = 'data/rec/SynthText/'
synthtext_textrecog_data_root = 'data/rec/SynthText/'
st_rec_train = dict(
synthtext_textrecog_train = dict(
type='OCRDataset',
data_root=st_data_root,
data_root=synthtext_textrecog_data_root,
data_prefix=dict(img_path='synthtext/SynthText_patch_horizontal'),
ann_file='train_labels.json',
test_mode=False,
pipeline=None)
st_an_rec_train = dict(
synthtext_an_textrecog_train = dict(
type='OCRDataset',
data_root=st_data_root,
data_root=synthtext_textrecog_data_root,
data_prefix=dict(img_path='synthtext/SynthText_patch_horizontal'),
ann_file='alphanumeric_train_labels.json',
test_mode=False,
pipeline=None)
st_sub_rec_train = dict(
synthtext_sub_textrecog_train = dict(
type='OCRDataset',
data_root=st_data_root,
data_root=synthtext_textrecog_data_root,
data_prefix=dict(img_path='synthtext/SynthText_patch_horizontal'),
ann_file='subset_train_labels.json',
test_mode=False,

6
configs/textrecog/_base_/datasets/synthtext_add.py
View File

@ -1,8 +1,8 @@
st_add_rec_data_root = 'data/rec/synthtext_add/'
synthtext_add_textrecog_data_root = 'data/rec/synthtext_add/'
st_add_rec_train = dict(
synthtext_add_textrecog_train = dict(
type='OCRDataset',
data_root=st_add_rec_data_root,
data_root=synthtext_add_textrecog_data_root,
ann_file='train_labels.json',
test_mode=False,
pipeline=None)

10
configs/textrecog/_base_/datasets/totaltext.py
View File

@ -1,15 +1,15 @@
tt_rec_data_root = 'data/totaltext/'
totaltext_textrecog_data_root = 'data/totaltext/'
tt_rec_train = dict(
totaltext_textrecog_train = dict(
type='OCRDataset',
data_root=tt_rec_data_root,
data_root=totaltext_textrecog_data_root,
ann_file='textrecog_train.json',
test_mode=False,
pipeline=None)
tt_rec_test = dict(
totaltext_textrecog_test = dict(
type='OCRDataset',
data_root=tt_rec_data_root,
data_root=totaltext_textrecog_data_root,
ann_file='textrecog_test.json',
test_mode=True,
pipeline=None)

9
configs/textrecog/abinet/abinet-vision_20e_st-an_mj.py
View File

@ -23,10 +23,13 @@ param_scheduler = [
]
# dataset settings
train_list = [_base_.mj_rec_train, _base_.st_an_rec_train]
train_list = [
_base_.mjsynth_textrecog_test, _base_.synthtext_an_textrecog_train
]
test_list = [
_base_.cute80_rec_test, _base_.iiit5k_rec_test, _base_.svt_rec_test,
_base_.svtp_rec_test, _base_.ic13_rec_test, _base_.ic15_rec_test
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test,
_base_.svt_textrecog_test, _base_.svtp_textrecog_test,
_base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
train_dataset = dict(

9
configs/textrecog/abinet/abinet_20e_st-an_mj.py
View File

@ -25,10 +25,13 @@ param_scheduler = [
]
# dataset settings
train_list = [_base_.mj_rec_train, _base_.st_an_rec_train]
train_list = [
_base_.mjsynth_textrecog_test, _base_.synthtext_an_textrecog_train
]
test_list = [
_base_.cute80_rec_test, _base_.iiit5k_rec_test, _base_.svt_rec_test,
_base_.svtp_rec_test, _base_.ic13_rec_test, _base_.ic15_rec_test
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test,
_base_.svt_textrecog_test, _base_.svtp_textrecog_test,
_base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
train_dataset = dict(

7
configs/textrecog/crnn/crnn_mini-vgg_5e_mj.py
View File

@ -13,10 +13,11 @@ _base_ = [
]
# dataset settings
train_list = [_base_.mj_rec_train]
train_list = [_base_.mjsynth_textrecog_test]
test_list = [
_base_.cute80_rec_test, _base_.iiit5k_rec_test, _base_.svt_rec_test,
_base_.svtp_rec_test, _base_.ic13_rec_test, _base_.ic15_rec_test
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test,
_base_.svt_textrecog_test, _base_.svtp_textrecog_test,
_base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
default_hooks = dict(logger=dict(type='LoggerHook', interval=50), )

8
configs/textrecog/master/master_resnet31_12e_st_mj_sa.py
View File

@ -23,11 +23,13 @@ param_scheduler = [
# dataset settings
train_list = [
_base_.mj_rec_train, _base_.st_rec_train, _base_.st_add_rec_train
_base_.mjsynth_textrecog_test, _base_.synthtext_textrecog_train,
_base_.synthtext_add_textrecog_train
]
test_list = [
_base_.cute80_rec_test, _base_.iiit5k_rec_test, _base_.svt_rec_test,
_base_.svtp_rec_test, _base_.ic13_rec_test, _base_.ic15_rec_test
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test,
_base_.svt_textrecog_test, _base_.svtp_textrecog_test,
_base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
train_dataset = dict(

7
configs/textrecog/nrtr/nrtr_modality-transform_6e_st_mj.py
View File

@ -20,10 +20,11 @@ param_scheduler = [
]
# dataset settings
train_list = [_base_.mj_rec_train, _base_.st_rec_train]
train_list = [_base_.mjsynth_textrecog_test, _base_.synthtext_textrecog_train]
test_list = [
_base_.cute80_rec_test, _base_.iiit5k_rec_test, _base_.svt_rec_test,
_base_.svtp_rec_test, _base_.ic13_rec_test, _base_.ic15_rec_test
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test,
_base_.svt_textrecog_test, _base_.svtp_textrecog_test,
_base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
train_dataset = dict(

7
configs/textrecog/nrtr/nrtr_resnet31-1by16-1by8_6e_st_mj.py
View File

@ -20,10 +20,11 @@ param_scheduler = [
]
# dataset settings
train_list = [_base_.mj_rec_train, _base_.st_rec_train]
train_list = [_base_.mjsynth_textrecog_test, _base_.synthtext_textrecog_train]
test_list = [
_base_.cute80_rec_test, _base_.iiit5k_rec_test, _base_.svt_rec_test,
_base_.svtp_rec_test, _base_.ic13_rec_test, _base_.ic15_rec_test
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test,
_base_.svt_textrecog_test, _base_.svtp_textrecog_test,
_base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
train_dataset = dict(

12
configs/textrecog/robust_scanner/robustscanner_resnet31_5e_st-sub_mj-sub_sa_real.py
View File

@ -19,13 +19,15 @@ default_hooks = dict(logger=dict(type='LoggerHook', interval=100))
# dataset settings
train_list = [
_base_.ic11_rec_train, _base_.ic13_rec_train, _base_.ic15_rec_train,
_base_.cocov1_rec_train, _base_.iiit5k_rec_train, _base_.mj_sub_rec_train,
_base_.st_sub_rec_train, _base_.st_add_rec_train
_base_.icdar2011_textrecog_train, _base_.icdar2013_textrecog_train,
_base_.icdar2015_textrecog_train, _base_.cocotextv1_textrecog_train,
_base_.iiit5k_textrecog_train, _base_.mjsynth_sub_textrecog_train,
_base_.synthtext_sub_textrecog_train, _base_.synthtext_add_textrecog_train
]
test_list = [
_base_.cute80_rec_test, _base_.iiit5k_rec_test, _base_.svt_rec_test,
_base_.svtp_rec_test, _base_.ic13_rec_test, _base_.ic15_rec_test
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test,
_base_.svt_textrecog_test, _base_.svtp_textrecog_test,
_base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
train_list = [

12
configs/textrecog/sar/sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real.py
View File

@ -19,13 +19,15 @@ default_hooks = dict(logger=dict(type='LoggerHook', interval=100))
# dataset settings
train_list = [
_base_.ic11_rec_train, _base_.ic13_rec_train, _base_.ic15_rec_train,
_base_.cocov1_rec_train, _base_.iiit5k_rec_train, _base_.mj_sub_rec_train,
_base_.st_sub_rec_train, _base_.st_add_rec_train
_base_.icdar2011_textrecog_train, _base_.icdar2013_textrecog_train,
_base_.icdar2015_textrecog_train, _base_.cocotextv1_textrecog_train,
_base_.iiit5k_textrecog_train, _base_.mjsynth_sub_textrecog_train,
_base_.synthtext_sub_textrecog_train, _base_.synthtext_add_textrecog_train
]
test_list = [
_base_.cute80_rec_test, _base_.iiit5k_rec_test, _base_.svt_rec_test,
_base_.svtp_rec_test, _base_.ic13_rec_test, _base_.ic15_rec_test
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test,
_base_.svt_textrecog_test, _base_.svtp_textrecog_test,
_base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
train_list = [

7
configs/textrecog/satrn/satrn_shallow_5e_st_mj.py
View File

@ -13,10 +13,11 @@ _base_ = [
]
# dataset settings
train_list = [_base_.mj_rec_train, _base_.st_rec_train]
train_list = [_base_.mjsynth_textrecog_test, _base_.synthtext_textrecog_train]
test_list = [
_base_.cute80_rec_test, _base_.iiit5k_rec_test, _base_.svt_rec_test,
_base_.svtp_rec_test, _base_.ic13_rec_test, _base_.ic15_rec_test
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test,
_base_.svt_textrecog_test, _base_.svtp_textrecog_test,
_base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
train_dataset = dict(

4
docs/en/get_started/quick_run.md
View File

@ -116,11 +116,11 @@ It may not have been trained to be optimal, but it is sufficient for a demo.
However, this value only reflects the performance of DBNet on the mini ICDAR 2015 dataset. For a comprehensive evaluation, we also need to see how it performs on out-of-distribution datasets. For example, `tests/data/det_toy_dataset` is a very small real dataset that we can use to verify the actual performance of DBNet.
Before testing, we also need to make some changes to the location of the dataset. Open `configs/_base_/det_datasets/icdar2015.py` and change `data_root` of `ic15_det_test` to `tests/data/det_toy_dataset`:
Before testing, we also need to make some changes to the location of the dataset. Open `configs/_base_/det_datasets/icdar2015.py` and change `data_root` of `icdar2015_textdet_test` to `tests/data/det_toy_dataset`:
```Python
# ...
ic15_det_test = dict(
icdar2015_textdet_test = dict(
type='OCRDataset',
data_root='tests/data/det_toy_dataset',
# ...

34
docs/en/user_guides/config.md
View File

@ -21,24 +21,24 @@ _base_ = [
]
# dataset settings
ic15_det_train = _base_.ic15_det_train
ic15_det_train.pipeline = _base_.train_pipeline
ic15_det_test = _base_.ic15_det_test
ic15_det_test.pipeline = _base_.test_pipeline
icdar2015_textdet_train = _base_.icdar2015_textdet_train
icdar2015_textdet_train.pipeline = _base_.train_pipeline
icdar2015_textdet_test = _base_.icdar2015_textdet_test
icdar2015_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=16,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ic15_det_train)
dataset=icdar2015_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ic15_det_test)
dataset=icdar2015_textdet_test)
```
### Configuration Inheritance
@ -77,11 +77,11 @@ Sometimes we may need to reference some fields in the `_base_` configuration dir
This syntax has been used extensively in the configuration of MMOCR, and the dataset and pipeline configurations for each model in MMOCR are referenced in the *_base_* configuration. For example,
```Python
ic15_det_train = _base_.ic15_det_train
icdar2015_textdet_train = _base_.icdar2015_textdet_train
# ...
train_dataloader = dict(
# ...
dataset=ic15_det_train)
dataset=icdar2015_textdet_train)
```
<div id="base_variable_modification"></div>
@ -98,18 +98,18 @@ In MMOCR, different algorithms usually have different pipelines in different dat
```Python
# Get the dataset in _base_
ic15_det_train = _base_.ic15_det_train
icdar2015_textdet_train = _base_.icdar2015_textdet_train
# You can modify the variables directly with Python's update
ic15_det_train.update(pipeline=_base_.train_pipeline)
icdar2015_textdet_train.update(pipeline=_base_.train_pipeline)
```
It can also be modified in the same way as changing Python class attributes.
```Python
# Get the dataset in _base_
ic15_det_train = _base_.ic15_det_train
icdar2015_textdet_train = _base_.icdar2015_textdet_train
# The class property method is modified
ic15_det_train.pipeline = _base_.train_pipeline
icdar2015_textdet_train.pipeline = _base_.train_pipeline
```
2. List
@ -291,7 +291,7 @@ For example, for text recognition tasks, Syn90k is used as the training set, whi
```Python
# text recognition dataset configuration
mj_rec_train = dict(
mjsynth_textrecog_test = dict(
type='OCRDataset',
data_root='data/rec/Syn90k/',
data_prefix=dict(img_path='mnt/ramdisk/max/90kDICT32px'),
@ -299,7 +299,7 @@ mj_rec_train = dict(
test_mode=False,
pipeline=None)
ic13_rec_test = dict(
icdar2013_textrecog_test = dict(
type='OCRDataset',
data_root='data/rec/icdar_2013/',
data_prefix=dict(img_path='Challenge2_Test_Task3_Images/'),
@ -307,7 +307,7 @@ ic13_rec_test = dict(
test_mode=True,
pipeline=None)
ic15_rec_test = dict(
icdar2015_textrecog_test = dict(
type='OCRDataset',
data_root='data/rec/icdar_2015/',
data_prefix=dict(img_path='ch4_test_word_images_gt/'),
@ -377,7 +377,7 @@ train_dataloader = dict(
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type='ConcatDataset',
datasets=[mj_rec_train],
datasets=[mjsynth_textrecog_test],
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
@ -387,7 +387,7 @@ val_dataloader = dict(
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type='ConcatDataset',
datasets=[ic13_rec_test, ic15_rec_test],
datasets=[icdar2013_textrecog_test, icdar2015_textrecog_test],
pipeline=test_pipeline))
test_dataloader = val_dataloader
```

22
docs/en/user_guides/dataset_prepare.md
View File

@ -90,7 +90,7 @@ When training or evaluating a model on new datasets, we need to write the datase
ic15_det_data_root = 'data/icdar2015' # dataset root path
# Train set config
ic15_det_train = dict(
icdar2015_textdet_train = dict(
type='OCRDataset',
data_root=ic15_det_data_root, # dataset root path
ann_file='instances_training.json', # name of annotation
@ -98,7 +98,7 @@ ic15_det_train = dict(
filter_cfg=dict(filter_empty_gt=True, min_size=32), # filtering empty images
pipeline=None)
# Test set config
ic15_det_test = dict(
icdar2015_textdet_test = dict(
type='OCRDataset',
data_root=ic15_det_data_root,
ann_file='instances_test.json',
@ -117,24 +117,24 @@ _base_ = [
'../_base_/schedules/schedule_sgd_1200e.py',
]
ic15_det_train = _base_.ic15_det_train # specify the training set
ic15_det_train.pipeline = _base_.train_pipeline # specify the training pipeline
ic15_det_test = _base_.ic15_det_test # specify the testing set
ic15_det_test.pipeline = _base_.test_pipeline # specify the testing pipeline
icdar2015_textdet_train = _base_.icdar2015_textdet_train # specify the training set
icdar2015_textdet_train.pipeline = _base_.train_pipeline # specify the training pipeline
icdar2015_textdet_test = _base_.icdar2015_textdet_test # specify the testing set
icdar2015_textdet_test.pipeline = _base_.test_pipeline # specify the testing pipeline
train_dataloader = dict(
batch_size=16,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ic15_det_train) # specify the dataset in train_dataloader
dataset=icdar2015_textdet_train) # specify the dataset in train_dataloader
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ic15_det_test) # specify the dataset in val_dataloader
dataset=icdar2015_textdet_test) # specify the dataset in val_dataloader
test_dataloader = val_dataloader
```
@ -167,11 +167,11 @@ _base_ = [ # Import all dataset configurations you want to use
]
# List of training datasets
train_list = [_base_.mj_rec_train]
train_list = [_base_.mjsynth_textrecog_test]
# List of testing datasets
test_list = [
_base_.cute80_rec_test, _base_.iiit5k_rec_test, _base_.svt_rec_test,
_base_.svtp_rec_test, _base_.ic13_rec_test, _base_.ic15_rec_test
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test, _base_.svt_textrecog_test,
_base_.svtp_textrecog_test, _base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
# Use ConcatDataset to combine the datasets in the list

4
docs/zh_cn/get_started/quick_run.md
View File

@ -118,11 +118,11 @@ python tools/train.py configs/textdet/dbnet/dbnet_resnet18_fpnc_1200e_icdar2015.
然而，这个数值只反映了 DBNet 在迷你 ICDAR 2015 数据集上的性能。要想更加客观地评判它的检测能力，我们还要看看它在分布外数据集上的表现。例如，`tests/data/det_toy_dataset` 就是一个很小的真实数据集，我们可以用它来验证一下 DBNet 的实际性能。
在测试前，我们同样需要对数据集的位置做一下修改。打开 `configs/_base_/det_datasets/icdar2015.py`，修改 `ic15_det_test` 的 `data_root` 为 `tests/data/det_toy_dataset`:
在测试前，我们同样需要对数据集的位置做一下修改。打开 `configs/_base_/det_datasets/icdar2015.py`，修改 `icdar2015_textdet_test` 的 `data_root` 为 `tests/data/det_toy_dataset`:
```Python
# ...
ic15_det_test = dict(
icdar2015_textdet_test = dict(
type='OCRDataset',
data_root='tests/data/det_toy_dataset',
# ...

34
docs/zh_cn/user_guides/config.md
View File

@ -21,24 +21,24 @@ _base_ = [
]
# dataset settings
ic15_det_train = _base_.ic15_det_train
ic15_det_train.pipeline = _base_.train_pipeline
ic15_det_test = _base_.ic15_det_test
ic15_det_test.pipeline = _base_.test_pipeline
icdar2015_textdet_train = _base_.icdar2015_textdet_train
icdar2015_textdet_train.pipeline = _base_.train_pipeline
icdar2015_textdet_test = _base_.icdar2015_textdet_test
icdar2015_textdet_test.pipeline = _base_.test_pipeline
train_dataloader = dict(
batch_size=16,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ic15_det_train)
dataset=icdar2015_textdet_train)
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ic15_det_test)
dataset=icdar2015_textdet_test)
```
### 配置文件的继承
@ -77,11 +77,11 @@ print(db_config)
该语法已广泛用于 MMOCR 的配置中。MMOCR 中各个模型的数据集和管道（pipeline）配置都引用于*基本*配置。如在
```Python
ic15_det_train = _base_.ic15_det_train
icdar2015_textdet_train = _base_.icdar2015_textdet_train
# ...
train_dataloader = dict(
# ...
dataset=ic15_det_train)
dataset=icdar2015_textdet_train)
```
### `_base_` 变量的修改
@ -96,18 +96,18 @@ train_dataloader = dict(
```python
# 获取 _base_ 中的数据集
ic15_det_train = _base_.ic15_det_train
icdar2015_textdet_train = _base_.icdar2015_textdet_train
# 可以直接利用 Python 的 update 修改变量
ic15_det_train.update(pipeline=_base_.train_pipeline)
icdar2015_textdet_train.update(pipeline=_base_.train_pipeline)
```
也可以使用类属性的方法进行修改：
```Python
# 获取 _base_ 中的数据集
ic15_det_train = _base_.ic15_det_train
icdar2015_textdet_train = _base_.icdar2015_textdet_train
# 类属性方法修改
ic15_det_train.pipeline = _base_.train_pipeline
icdar2015_textdet_train.pipeline = _base_.train_pipeline
```
2. 列表
@ -288,7 +288,7 @@ test_cfg = dict(type='TestLoop')
```Python
# 识别数据集配置
mj_rec_train = dict(
mjsynth_textrecog_test = dict(
type='OCRDataset',
data_root='data/rec/Syn90k/',
data_prefix=dict(img_path='mnt/ramdisk/max/90kDICT32px'),
@ -296,7 +296,7 @@ mj_rec_train = dict(
test_mode=False,
pipeline=None)
ic13_rec_test = dict(
icdar2013_textrecog_test = dict(
type='OCRDataset',
data_root='data/rec/icdar_2013/',
data_prefix=dict(img_path='Challenge2_Test_Task3_Images/'),
@ -304,7 +304,7 @@ ic13_rec_test = dict(
test_mode=True,
pipeline=None)
ic15_rec_test = dict(
icdar2015_textrecog_test = dict(
type='OCRDataset',
data_root='data/rec/icdar_2015/',
data_prefix=dict(img_path='ch4_test_word_images_gt/'),
@ -376,7 +376,7 @@ train_dataloader = dict(
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type='ConcatDataset',
datasets=[mj_rec_train],
datasets=[mjsynth_textrecog_test],
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
@ -386,7 +386,7 @@ val_dataloader = dict(
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type='ConcatDataset',
datasets=[ic13_rec_test, ic15_rec_test],
datasets=[icdar2013_textrecog_test, icdar2015_textrecog_test],
pipeline=test_pipeline))
test_dataloader = val_dataloader
```

22
docs/zh_cn/user_guides/dataset_prepare.md
View File

@ -90,7 +90,7 @@ data/icdar2015
ic15_det_data_root = 'data/icdar2015' # 数据集根目录
# 训练集配置
ic15_det_train = dict(
icdar2015_textdet_train = dict(
type='OCRDataset',
data_root=ic15_det_data_root, # 数据根目录
ann_file='instances_training.json', # 标注文件名称
@ -98,7 +98,7 @@ ic15_det_train = dict(
filter_cfg=dict(filter_empty_gt=True, min_size=32), # 数据过滤
pipeline=None)
# 测试集配置
ic15_det_test = dict(
icdar2015_textdet_test = dict(
type='OCRDataset',
data_root=ic15_det_data_root,
ann_file='instances_test.json',
@ -117,24 +117,24 @@ _base_ = [
'../_base_/schedules/schedule_sgd_1200e.py',
]
ic15_det_train = _base_.ic15_det_train # 指定训练集
ic15_det_train.pipeline = _base_.train_pipeline # 指定训练集使用的数据流水线
ic15_det_test = _base_.ic15_det_test # 指定测试集
ic15_det_test.pipeline = _base_.test_pipeline # 指定测试集使用的数据流水线
icdar2015_textdet_train = _base_.icdar2015_textdet_train # 指定训练集
icdar2015_textdet_train.pipeline = _base_.train_pipeline # 指定训练集使用的数据流水线
icdar2015_textdet_test = _base_.icdar2015_textdet_test # 指定测试集
icdar2015_textdet_test.pipeline = _base_.test_pipeline # 指定测试集使用的数据流水线
train_dataloader = dict(
batch_size=16,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=ic15_det_train) # 在 train_dataloader 中指定使用的训练数据集
dataset=icdar2015_textdet_train) # 在 train_dataloader 中指定使用的训练数据集
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=ic15_det_test) # 在 val_dataloader 中指定使用的验证数据集
dataset=icdar2015_textdet_test) # 在 val_dataloader 中指定使用的验证数据集
test_dataloader = val_dataloader
```
@ -167,11 +167,11 @@ _base_ = [ # 导入所有需要使用的数据集配置
]
# 训练集列表
train_list = [_base_.mj_rec_train]
train_list = [_base_.mjsynth_textrecog_test]
# 测试集列表
test_list = [
_base_.cute80_rec_test, _base_.iiit5k_rec_test, _base_.svt_rec_test,
_base_.svtp_rec_test, _base_.ic13_rec_test, _base_.ic15_rec_test
_base_.cute80_textrecog_test, _base_.iiit5k_textrecog_test, _base_.svt_textrecog_test,
_base_.svtp_textrecog_test, _base_.icdar2013_textrecog_test, _base_.icdar2015_textrecog_test
]
# 使用 ConcatDataset 来级联列表中的多个数据集

4
mmocr/datasets/preparers/data_converter.py
View File

@ -97,13 +97,13 @@ class BaseDataConverter:
Examples:
Generated dataset config
>>> ic15_rec_data_root = 'data/icdar2015/'
>>> ic15_rec_train = dict(
>>> icdar2015_textrecog_train = dict(
>>> type='OCRDataset',
>>> data_root=ic15_rec_data_root,
>>> ann_file='textrecog_train.json',
>>> test_mode=False,
>>> pipeline=None)
>>> ic15_rec_test = dict(
>>> icdar2015_textrecog_test = dict(
>>> type='OCRDataset',
>>> data_root=ic15_rec_data_root,
>>> ann_file='textrecog_test.json',

2
mmocr/datasets/preparers/dumpers/dumpers.py
View File

@ -29,7 +29,7 @@ class JsonDumper:
Examples:
The returned dataset config
>>> ic15_rec_train = dict(
>>> icdar2015_textrecog_train = dict(
>>> type='OCRDataset',
>>> data_root=ic15_rec_data_root,
>>> ann_file='textrecog_train.json',