Skip to content

scooter fastscnn trainning #3808

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.

Sign up for GitHub

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

Open
wants to merge 4 commits into
base: main
Choose a base branch
from
+921 −17
Open

scooter fastscnn trainning #3808

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
9a87949
scooter fastscnn trainning
li3coLich Nov 6, 2024
ff52772
scooter seg
Nov 7, 2024
1714923
scooter central
Nov 7, 2024
c4ceea5
add 540x960
Nov 8, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
68 changes: 68 additions & 0 deletions configs/_base_/datasets/ucsd_half-256x512.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
# dataset settings
dataset_type = 'MapillaryDataset_v1'
data_root = '/lich-central-vol/dataset/campus_dataset'
crop_size = (256, 512)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations'),
dict(
type='RandomResize',
scale=(1280, 720),
ratio_range=(0.5, 2.0),
keep_ratio=True),
dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
dict(type='RandomFlip', prob=0.5),
dict(type='PhotoMetricDistortion'),
dict(type='PackSegInputs')
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='Resize', scale=(1280, 720), keep_ratio=True),
# add loading annotation after ``Resize`` because ground truth
# does not need to do resize data transform
dict(type='LoadAnnotations'),
dict(type='PackSegInputs')
]
img_ratios = [0.5, 0.75, 1.0, 1.25, 1.5, 1.75]
tta_pipeline = [
dict(type='LoadImageFromFile', file_client_args=dict(backend='disk')),
dict(
type='TestTimeAug',
transforms=[
[
dict(type='Resize', scale_factor=r, keep_ratio=True)
for r in img_ratios
],
[
dict(type='RandomFlip', prob=0., direction='horizontal'),
dict(type='RandomFlip', prob=1., direction='horizontal')
], [dict(type='LoadAnnotations')], [dict(type='PackSegInputs')]
])
]
train_dataloader = dict(
batch_size=2,
num_workers=2,
persistent_workers=True,
sampler=dict(type='InfiniteSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='training/images', seg_map_path='training/labels'),
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='validation/images',
seg_map_path='validation/labels'),
pipeline=test_pipeline))
test_dataloader = val_dataloader

val_evaluator = dict(type='IoUMetric', iou_metrics=['mIoU'])
test_evaluator = val_evaluator
68 changes: 68 additions & 0 deletions configs/_base_/datasets/ucsd_half-360x640.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
# dataset settings
dataset_type = 'MapillaryDataset_v1'
data_root = '/lich-central-vol/dataset/campus_dataset'
crop_size = (360, 640)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations'),
dict(
type='RandomResize',
scale=(1280, 720),
ratio_range=(0.5, 2.0),
keep_ratio=True),
dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
dict(type='RandomFlip', prob=0.5),
dict(type='PhotoMetricDistortion'),
dict(type='PackSegInputs')
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='Resize', scale=(1280, 720), keep_ratio=True),
# add loading annotation after ``Resize`` because ground truth
# does not need to do resize data transform
dict(type='LoadAnnotations'),
dict(type='PackSegInputs')
]
img_ratios = [0.5, 0.75, 1.0, 1.25, 1.5, 1.75]
tta_pipeline = [
dict(type='LoadImageFromFile', file_client_args=dict(backend='disk')),
dict(
type='TestTimeAug',
transforms=[
[
dict(type='Resize', scale_factor=r, keep_ratio=True)
for r in img_ratios
],
[
dict(type='RandomFlip', prob=0., direction='horizontal'),
dict(type='RandomFlip', prob=1., direction='horizontal')
], [dict(type='LoadAnnotations')], [dict(type='PackSegInputs')]
])
]
train_dataloader = dict(
batch_size=2,
num_workers=4,
persistent_workers=True,
sampler=dict(type='InfiniteSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='training/images', seg_map_path='training/labels'),
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='validation/images',
seg_map_path='validation/labels'),
pipeline=test_pipeline))
test_dataloader = val_dataloader

val_evaluator = dict(type='IoUMetric', iou_metrics=['mIoU'])
test_evaluator = val_evaluator
68 changes: 68 additions & 0 deletions configs/_base_/datasets/ucsd_half-512x1024.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
# dataset settings
dataset_type = 'MapillaryDataset_v1'
data_root = '/lich-central-vol/dataset/ucsd_half'
crop_size = (512, 1024)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations'),
dict(
type='RandomResize',
scale=(1280, 720),
ratio_range=(0.5, 2.0),
keep_ratio=True),
dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
dict(type='RandomFlip', prob=0.5),
dict(type='PhotoMetricDistortion'),
dict(type='PackSegInputs')
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='Resize', scale=(1280, 720), keep_ratio=True),
# add loading annotation after ``Resize`` because ground truth
# does not need to do resize data transform
dict(type='LoadAnnotations'),
dict(type='PackSegInputs')
]
img_ratios = [0.5, 0.75, 1.0, 1.25, 1.5, 1.75]
tta_pipeline = [
dict(type='LoadImageFromFile', file_client_args=dict(backend='disk')),
dict(
type='TestTimeAug',
transforms=[
[
dict(type='Resize', scale_factor=r, keep_ratio=True)
for r in img_ratios
],
[
dict(type='RandomFlip', prob=0., direction='horizontal'),
dict(type='RandomFlip', prob=1., direction='horizontal')
], [dict(type='LoadAnnotations')], [dict(type='PackSegInputs')]
])
]
train_dataloader = dict(
batch_size=2,
num_workers=2,
persistent_workers=True,
sampler=dict(type='InfiniteSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='training/images', seg_map_path='training/labels'),
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='validation/images',
seg_map_path='validation/labels'),
pipeline=test_pipeline))
test_dataloader = val_dataloader

val_evaluator = dict(type='IoUMetric', iou_metrics=['mIoU'])
test_evaluator = val_evaluator
68 changes: 68 additions & 0 deletions configs/_base_/datasets/ucsd_half-540x960.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
# dataset settings
dataset_type = 'MapillaryDataset_v1'
data_root = '/lich-central-vol/dataset/campus_dataset'
crop_size = (540, 960)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations'),
dict(
type='RandomResize',
scale=(1280, 720),
ratio_range=(0.5, 2.0),
keep_ratio=True),
dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
dict(type='RandomFlip', prob=0.5),
dict(type='PhotoMetricDistortion'),
dict(type='PackSegInputs')
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='Resize', scale=(1280, 720), keep_ratio=True),
# add loading annotation after ``Resize`` because ground truth
# does not need to do resize data transform
dict(type='LoadAnnotations'),
dict(type='PackSegInputs')
]
img_ratios = [0.5, 0.75, 1.0, 1.25, 1.5, 1.75]
tta_pipeline = [
dict(type='LoadImageFromFile', file_client_args=dict(backend='disk')),
dict(
type='TestTimeAug',
transforms=[
[
dict(type='Resize', scale_factor=r, keep_ratio=True)
for r in img_ratios
],
[
dict(type='RandomFlip', prob=0., direction='horizontal'),
dict(type='RandomFlip', prob=1., direction='horizontal')
], [dict(type='LoadAnnotations')], [dict(type='PackSegInputs')]
])
]
train_dataloader = dict(
batch_size=2,
num_workers=2,
persistent_workers=True,
sampler=dict(type='InfiniteSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='training/images', seg_map_path='training/labels'),
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='validation/images',
seg_map_path='validation/labels'),
pipeline=test_pipeline))
test_dataloader = val_dataloader

val_evaluator = dict(type='IoUMetric', iou_metrics=['mIoU'])
test_evaluator = val_evaluator
68 changes: 68 additions & 0 deletions configs/_base_/datasets/ucsd_half-720x1280.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
# dataset settings
dataset_type = 'MapillaryDataset_v1'
data_root = '/lich-central-vol/dataset/campus_dataset'
crop_size = (720, 1280)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations'),
dict(
type='RandomResize',
scale=(1280, 720),
ratio_range=(0.5, 2.0),
keep_ratio=True),
dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
dict(type='RandomFlip', prob=0.5),
dict(type='PhotoMetricDistortion'),
dict(type='PackSegInputs')
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='Resize', scale=(1280, 720), keep_ratio=True),
# add loading annotation after ``Resize`` because ground truth
# does not need to do resize data transform
dict(type='LoadAnnotations'),
dict(type='PackSegInputs')
]
img_ratios = [0.5, 0.75, 1.0, 1.25, 1.5, 1.75]
tta_pipeline = [
dict(type='LoadImageFromFile', file_client_args=dict(backend='disk')),
dict(
type='TestTimeAug',
transforms=[
[
dict(type='Resize', scale_factor=r, keep_ratio=True)
for r in img_ratios
],
[
dict(type='RandomFlip', prob=0., direction='horizontal'),
dict(type='RandomFlip', prob=1., direction='horizontal')
], [dict(type='LoadAnnotations')], [dict(type='PackSegInputs')]
])
]
train_dataloader = dict(
batch_size=2,
num_workers=4,
persistent_workers=True,
sampler=dict(type='InfiniteSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='training/images', seg_map_path='training/labels'),
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='validation/images',
seg_map_path='validation/labels'),
pipeline=test_pipeline))
test_dataloader = val_dataloader

val_evaluator = dict(type='IoUMetric', iou_metrics=['mIoU'])
test_evaluator = val_evaluator
68 changes: 68 additions & 0 deletions configs/_base_/datasets/ucsd_half-720x720.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
# dataset settings
dataset_type = 'MapillaryDataset_v1'
data_root = '/lich-central-vol/dataset/ucsd_half'
crop_size = (720, 720)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations'),
dict(
type='RandomResize',
scale=(1280, 720),
ratio_range=(0.5, 2.0),
keep_ratio=True),
dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
dict(type='RandomFlip', prob=0.5),
dict(type='PhotoMetricDistortion'),
dict(type='PackSegInputs')
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='Resize', scale=(1280, 720), keep_ratio=True),
# add loading annotation after ``Resize`` because ground truth
# does not need to do resize data transform
dict(type='LoadAnnotations'),
dict(type='PackSegInputs')
]
img_ratios = [0.5, 0.75, 1.0, 1.25, 1.5, 1.75]
tta_pipeline = [
dict(type='LoadImageFromFile', file_client_args=dict(backend='disk')),
dict(
type='TestTimeAug',
transforms=[
[
dict(type='Resize', scale_factor=r, keep_ratio=True)
for r in img_ratios
],
[
dict(type='RandomFlip', prob=0., direction='horizontal'),
dict(type='RandomFlip', prob=1., direction='horizontal')
], [dict(type='LoadAnnotations')], [dict(type='PackSegInputs')]
])
]
train_dataloader = dict(
batch_size=2,
num_workers=2,
persistent_workers=True,
sampler=dict(type='InfiniteSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='training/images', seg_map_path='training/labels'),
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='validation/images',
seg_map_path='validation/labels'),
pipeline=test_pipeline))
test_dataloader = val_dataloader

val_evaluator = dict(type='IoUMetric', iou_metrics=['mIoU'])
test_evaluator = val_evaluator
68 changes: 68 additions & 0 deletions configs/_base_/datasets/ucsd_half-832x832.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
# dataset settings
dataset_type = 'MapillaryDataset_v1'
data_root = '/lich-central-vol/dataset/ucsd_half'
crop_size = (832, 832)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations'),
dict(
type='RandomResize',
scale=(1280, 720),
ratio_range=(0.5, 2.0),
keep_ratio=True),
dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
dict(type='RandomFlip', prob=0.5),
dict(type='PhotoMetricDistortion'),
dict(type='PackSegInputs')
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='Resize', scale=(1280, 720), keep_ratio=True),
# add loading annotation after ``Resize`` because ground truth
# does not need to do resize data transform
dict(type='LoadAnnotations'),
dict(type='PackSegInputs')
]
img_ratios = [0.5, 0.75, 1.0, 1.25, 1.5, 1.75]
tta_pipeline = [
dict(type='LoadImageFromFile', file_client_args=dict(backend='disk')),
dict(
type='TestTimeAug',
transforms=[
[
dict(type='Resize', scale_factor=r, keep_ratio=True)
for r in img_ratios
],
[
dict(type='RandomFlip', prob=0., direction='horizontal'),
dict(type='RandomFlip', prob=1., direction='horizontal')
], [dict(type='LoadAnnotations')], [dict(type='PackSegInputs')]
])
]
train_dataloader = dict(
batch_size=2,
num_workers=2,
persistent_workers=True,
sampler=dict(type='InfiniteSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='training/images', seg_map_path='training/labels'),
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='validation/images',
seg_map_path='validation/labels'),
pipeline=test_pipeline))
test_dataloader = val_dataloader

val_evaluator = dict(type='IoUMetric', iou_metrics=['mIoU'])
test_evaluator = val_evaluator
68 changes: 68 additions & 0 deletions configs/_base_/datasets/ucsd_half.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
# dataset settings
dataset_type = 'MapillaryDataset_v1'
data_root = '/lich-central-vol/dataset/campus_dataset'
crop_size = (720, 1280)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations'),
dict(
type='RandomResize',
scale=(1280, 720),
ratio_range=(0.5, 2.0),
keep_ratio=True),
dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
dict(type='RandomFlip', prob=0.5),
dict(type='PhotoMetricDistortion'),
dict(type='PackSegInputs')
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='Resize', scale=(1280, 720), keep_ratio=True),
# add loading annotation after ``Resize`` because ground truth
# does not need to do resize data transform
dict(type='LoadAnnotations'),
dict(type='PackSegInputs')
]
img_ratios = [0.5, 0.75, 1.0, 1.25, 1.5, 1.75]
tta_pipeline = [
dict(type='LoadImageFromFile', file_client_args=dict(backend='disk')),
dict(
type='TestTimeAug',
transforms=[
[
dict(type='Resize', scale_factor=r, keep_ratio=True)
for r in img_ratios
],
[
dict(type='RandomFlip', prob=0., direction='horizontal'),
dict(type='RandomFlip', prob=1., direction='horizontal')
], [dict(type='LoadAnnotations')], [dict(type='PackSegInputs')]
])
]
train_dataloader = dict(
batch_size=2,
num_workers=2,
persistent_workers=True,
sampler=dict(type='InfiniteSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='training/images', seg_map_path='training/labels'),
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='validation/images',
seg_map_path='validation/labels'),
pipeline=test_pipeline))
test_dataloader = val_dataloader

val_evaluator = dict(type='IoUMetric', iou_metrics=['mIoU'])
test_evaluator = val_evaluator
68 changes: 68 additions & 0 deletions configs/_base_/datasets/ucsd_half_4x4-512x1024.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
# dataset settings
dataset_type = 'MapillaryDataset_v1'
data_root = '/lich-central-vol/dataset/ucsd_half'
crop_size = (512, 1024)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations'),
dict(
type='RandomResize',
scale=(1280, 720),
ratio_range=(0.5, 2.0),
keep_ratio=True),
dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
dict(type='RandomFlip', prob=0.5),
dict(type='PhotoMetricDistortion'),
dict(type='PackSegInputs')
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='Resize', scale=(1280, 720), keep_ratio=True),
# add loading annotation after ``Resize`` because ground truth
# does not need to do resize data transform
dict(type='LoadAnnotations'),
dict(type='PackSegInputs')
]
img_ratios = [0.5, 0.75, 1.0, 1.25, 1.5, 1.75]
tta_pipeline = [
dict(type='LoadImageFromFile', file_client_args=dict(backend='disk')),
dict(
type='TestTimeAug',
transforms=[
[
dict(type='Resize', scale_factor=r, keep_ratio=True)
for r in img_ratios
],
[
dict(type='RandomFlip', prob=0., direction='horizontal'),
dict(type='RandomFlip', prob=1., direction='horizontal')
], [dict(type='LoadAnnotations')], [dict(type='PackSegInputs')]
])
]
train_dataloader = dict(
batch_size=2,
num_workers=4,
persistent_workers=True,
sampler=dict(type='InfiniteSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='training/images', seg_map_path='training/labels'),
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='validation/images',
seg_map_path='validation/labels'),
pipeline=test_pipeline))
test_dataloader = val_dataloader

val_evaluator = dict(type='IoUMetric', iou_metrics=['mIoU'])
test_evaluator = val_evaluator
68 changes: 68 additions & 0 deletions configs/_base_/datasets/ucsd_half_4x4-512x512.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
# dataset settings
dataset_type = 'MapillaryDataset_v1'
data_root = '/lich-central-vol/dataset/ucsd_half'
crop_size = (512, 512)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations'),
dict(
type='RandomResize',
scale=(1280, 720),
ratio_range=(0.5, 2.0),
keep_ratio=True),
dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
dict(type='RandomFlip', prob=0.5),
dict(type='PhotoMetricDistortion'),
dict(type='PackSegInputs')
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='Resize', scale=(1280, 720), keep_ratio=True),
# add loading annotation after ``Resize`` because ground truth
# does not need to do resize data transform
dict(type='LoadAnnotations'),
dict(type='PackSegInputs')
]
img_ratios = [0.5, 0.75, 1.0, 1.25, 1.5, 1.75]
tta_pipeline = [
dict(type='LoadImageFromFile', file_client_args=dict(backend='disk')),
dict(
type='TestTimeAug',
transforms=[
[
dict(type='Resize', scale_factor=r, keep_ratio=True)
for r in img_ratios
],
[
dict(type='RandomFlip', prob=0., direction='horizontal'),
dict(type='RandomFlip', prob=1., direction='horizontal')
], [dict(type='LoadAnnotations')], [dict(type='PackSegInputs')]
])
]
train_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='InfiniteSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='training/images', seg_map_path='training/labels'),
pipeline=train_pipeline))
val_dataloader = dict(
batch_size=1,
num_workers=4,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_prefix=dict(
img_path='validation/images',
seg_map_path='validation/labels'),
pipeline=test_pipeline))
test_dataloader = val_dataloader

val_evaluator = dict(type='IoUMetric', iou_metrics=['mIoU'])
test_evaluator = val_evaluator
14 changes: 9 additions & 5 deletions configs/_base_/models/fast_scnn.py
View file
Original file line number Diff line number Diff line change
@@ -2,8 +2,12 @@
norm_cfg = dict(type='SyncBN', requires_grad=True, momentum=0.01)
data_preprocessor = dict(
type='SegDataPreProcessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
# mean=[123.675, 116.28, 103.53],
# std=[58.395, 57.12, 57.375],
# mean=[97.85609871, 96.75819166, 101.97295342], # ucsd_half
# std=[83.26665057, 80.24579992, 79.9720934],
mean=[100.37816636, 103.69571653, 104.02350565], # campus data
std=[74.34289061, 76.53997023, 83.26519555],
bgr_to_rgb=True,
pad_val=0,
seg_pad_val=255)
@@ -28,7 +32,7 @@
in_channels=128,
channels=128,
concat_input=False,
num_classes=19,
num_classes=66,
in_index=-1,
norm_cfg=norm_cfg,
align_corners=False,
@@ -40,7 +44,7 @@
in_channels=128,
channels=32,
num_convs=1,
num_classes=19,
num_classes=66,
in_index=-2,
norm_cfg=norm_cfg,
concat_input=False,
@@ -52,7 +56,7 @@
in_channels=64,
channels=32,
num_convs=1,
num_classes=19,
num_classes=66,
in_index=-3,
norm_cfg=norm_cfg,
concat_input=False,
12 changes: 7 additions & 5 deletions configs/_base_/models/icnet_r50-d8.py
View file
Original file line number Diff line number Diff line change
@@ -2,8 +2,10 @@
norm_cfg = dict(type='SyncBN', requires_grad=True)
data_preprocessor = dict(
type='SegDataPreProcessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
# mean=[123.675, 116.28, 103.53],
# std=[58.395, 57.12, 57.375],
mean=[97.85609871, 96.75819166, 101.97295342],
std=[83.26665057, 80.24579992, 79.9720934],
bgr_to_rgb=True,
pad_val=0,
seg_pad_val=255)
@@ -45,7 +47,7 @@
num_convs=1,
in_index=2,
dropout_ratio=0,
num_classes=19,
num_classes=66,
norm_cfg=norm_cfg,
concat_input=False,
align_corners=False,
@@ -57,7 +59,7 @@
in_channels=128,
channels=128,
num_convs=1,
num_classes=19,
num_classes=66,
in_index=0,
norm_cfg=norm_cfg,
concat_input=False,
@@ -69,7 +71,7 @@
in_channels=128,
channels=128,
num_convs=1,
num_classes=19,
num_classes=66,
in_index=1,
norm_cfg=norm_cfg,
concat_input=False,
8 changes: 5 additions & 3 deletions configs/_base_/models/lraspp_m-v3-d8.py
View file
Original file line number Diff line number Diff line change
@@ -2,8 +2,10 @@
norm_cfg = dict(type='SyncBN', eps=0.001, requires_grad=True)
data_preprocessor = dict(
type='SegDataPreProcessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
# mean=[123.675, 116.28, 103.53],
# std=[58.395, 57.12, 57.375],
mean=[97.85609871, 96.75819166, 101.97295342],
std=[83.26665057, 80.24579992, 79.9720934],
bgr_to_rgb=True,
pad_val=0,
seg_pad_val=255)
@@ -22,7 +24,7 @@
channels=128,
input_transform='multiple_select',
dropout_ratio=0.1,
num_classes=19,
num_classes=66,
norm_cfg=norm_cfg,
act_cfg=dict(type='ReLU'),
align_corners=False,
15 changes: 15 additions & 0 deletions configs/fastscnn/fast_scnn_8xb4-160k_ucsd-256x512.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
_base_ = [
'../_base_/models/fast_scnn.py', '../_base_/datasets/ucsd_half-256x512.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_160k.py'
]
crop_size = (256, 512)
data_preprocessor = dict(size=crop_size)
model = dict(data_preprocessor=data_preprocessor)
# Re-config the data sampler.
train_dataloader = dict(batch_size=4, num_workers=4)
val_dataloader = dict(batch_size=1, num_workers=4)
test_dataloader = val_dataloader

# Re-config the optimizer.
optimizer = dict(type='SGD', lr=0.12, momentum=0.9, weight_decay=4e-5)
optim_wrapper = dict(type='OptimWrapper', optimizer=optimizer)
15 changes: 15 additions & 0 deletions configs/fastscnn/fast_scnn_8xb4-160k_ucsd-360x640.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
_base_ = [
'../_base_/models/fast_scnn.py', '../_base_/datasets/ucsd_half-360x640.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_160k.py'
]
crop_size = (360, 640)
data_preprocessor = dict(size=crop_size)
model = dict(data_preprocessor=data_preprocessor)
# Re-config the data sampler.
train_dataloader = dict(batch_size=4, num_workers=4)
val_dataloader = dict(batch_size=1, num_workers=4)
test_dataloader = val_dataloader

# Re-config the optimizer.
optimizer = dict(type='SGD', lr=0.12, momentum=0.9, weight_decay=4e-5)
optim_wrapper = dict(type='OptimWrapper', optimizer=optimizer)
15 changes: 15 additions & 0 deletions configs/fastscnn/fast_scnn_8xb4-160k_ucsd-540x960.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
_base_ = [
'../_base_/models/fast_scnn.py', '../_base_/datasets/ucsd_half-540x960.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_160k.py'
]
crop_size = (540, 960)
data_preprocessor = dict(size=crop_size)
model = dict(data_preprocessor=data_preprocessor)
# Re-config the data sampler.
train_dataloader = dict(batch_size=4, num_workers=4)
val_dataloader = dict(batch_size=1, num_workers=4)
test_dataloader = val_dataloader

# Re-config the optimizer.
optimizer = dict(type='SGD', lr=0.12, momentum=0.9, weight_decay=4e-5)
optim_wrapper = dict(type='OptimWrapper', optimizer=optimizer)
15 changes: 15 additions & 0 deletions configs/fastscnn/fast_scnn_8xb4-160k_ucsd-720x1280.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
_base_ = [
'../_base_/models/fast_scnn.py', '../_base_/datasets/ucsd_half.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_160k.py'
]
crop_size = (720, 1280)
data_preprocessor = dict(size=crop_size)
model = dict(data_preprocessor=data_preprocessor)
# Re-config the data sampler.
train_dataloader = dict(batch_size=4, num_workers=4)
val_dataloader = dict(batch_size=1, num_workers=4)
test_dataloader = val_dataloader

# Re-config the optimizer.
optimizer = dict(type='SGD', lr=0.12, momentum=0.9, weight_decay=4e-5)
optim_wrapper = dict(type='OptimWrapper', optimizer=optimizer)
8 changes: 8 additions & 0 deletions configs/icnet/icnet_r18-d8-in1k-pre_4xb2-160k_cityscapes-832x832_.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
_base_ = './icnet_r50-d8_4xb2-160k_cityscapes-832x832_.py'
model = dict(
backbone=dict(
layer_channels=(128, 512),
backbone_cfg=dict(
depth=18,
init_cfg=dict(
type='Pretrained', checkpoint='open-mmlab://resnet18_v1c'))))
4 changes: 2 additions & 2 deletions configs/icnet/icnet_r50-d8_4xb2-160k_cityscapes-832x832.py
View file
Original file line number Diff line number Diff line change
@@ -1,8 +1,8 @@
_base_ = [
'../_base_/models/icnet_r50-d8.py',
'../_base_/datasets/cityscapes_832x832.py', '../_base_/default_runtime.py',
'../_base_/datasets/ucsd_half.py', '../_base_/default_runtime.py',
'../_base_/schedules/schedule_160k.py'
]
crop_size = (832, 832)
crop_size = (720, 1280)
data_preprocessor = dict(size=crop_size)
model = dict(data_preprocessor=data_preprocessor)
8 changes: 8 additions & 0 deletions configs/icnet/icnet_r50-d8_4xb2-160k_cityscapes-832x832_.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
_base_ = [
'../_base_/models/icnet_r50-d8.py',
'../_base_/datasets/ucsd_half.py', '../_base_/default_runtime.py',
'../_base_/schedules/schedule_160k.py'
]
crop_size = (720, 1280)
data_preprocessor = dict(size=crop_size)
model = dict(data_preprocessor=data_preprocessor)
22 changes: 22 additions & 0 deletions configs/mobilenet_v3/mobilenet-v3-d8-scratch-s_lraspp_4xb4-320k_ucsd-256x512.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,22 @@
_base_ = './mobilenet-v3-d8-scratch_lraspp_4xb4-320k_ucsd-256x512.py'
norm_cfg = dict(type='SyncBN', eps=0.001, requires_grad=True)
model = dict(
type='EncoderDecoder',
backbone=dict(
type='MobileNetV3',
arch='small',
out_indices=(0, 1, 12),
norm_cfg=norm_cfg),
decode_head=dict(
type='LRASPPHead',
in_channels=(16, 16, 576),
in_index=(0, 1, 2),
channels=128,
input_transform='multiple_select',
dropout_ratio=0.1,
num_classes=19,
norm_cfg=norm_cfg,
act_cfg=dict(type='ReLU'),
align_corners=False,
loss_decode=dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0)))
22 changes: 22 additions & 0 deletions configs/mobilenet_v3/mobilenet-v3-d8-scratch-s_lraspp_4xb4-320k_ucsd-360x640.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,22 @@
_base_ = './mobilenet-v3-d8-scratch_lraspp_4xb4-320k_ucsd-360x640.py'
norm_cfg = dict(type='SyncBN', eps=0.001, requires_grad=True)
model = dict(
type='EncoderDecoder',
backbone=dict(
type='MobileNetV3',
arch='small',
out_indices=(0, 1, 12),
norm_cfg=norm_cfg),
decode_head=dict(
type='LRASPPHead',
in_channels=(16, 16, 576),
in_index=(0, 1, 2),
channels=128,
input_transform='multiple_select',
dropout_ratio=0.1,
num_classes=19,
norm_cfg=norm_cfg,
act_cfg=dict(type='ReLU'),
align_corners=False,
loss_decode=dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0)))
22 changes: 22 additions & 0 deletions configs/mobilenet_v3/mobilenet-v3-d8-scratch-s_lraspp_4xb4-320k_ucsd-720x1280.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,22 @@
_base_ = './mobilenet-v3-d8-scratch_lraspp_4xb4-320k_ucsd-720x1280.py'
norm_cfg = dict(type='SyncBN', eps=0.001, requires_grad=True)
model = dict(
type='EncoderDecoder',
backbone=dict(
type='MobileNetV3',
arch='small',
out_indices=(0, 1, 12),
norm_cfg=norm_cfg),
decode_head=dict(
type='LRASPPHead',
in_channels=(16, 16, 576),
in_index=(0, 1, 2),
channels=128,
input_transform='multiple_select',
dropout_ratio=0.1,
num_classes=19,
norm_cfg=norm_cfg,
act_cfg=dict(type='ReLU'),
align_corners=False,
loss_decode=dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0)))
22 changes: 22 additions & 0 deletions configs/mobilenet_v3/mobilenet-v3-d8-scratch-s_lraspp_4xb4-320k_ucsd-720x720.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,22 @@
_base_ = './mobilenet-v3-d8-scratch_lraspp_4xb4-320k_ucsd-720x720.py'
norm_cfg = dict(type='SyncBN', eps=0.001, requires_grad=True)
model = dict(
type='EncoderDecoder',
backbone=dict(
type='MobileNetV3',
arch='small',
out_indices=(0, 1, 12),
norm_cfg=norm_cfg),
decode_head=dict(
type='LRASPPHead',
in_channels=(16, 16, 576),
in_index=(0, 1, 2),
channels=128,
input_transform='multiple_select',
dropout_ratio=0.1,
num_classes=19,
norm_cfg=norm_cfg,
act_cfg=dict(type='ReLU'),
align_corners=False,
loss_decode=dict(
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0)))
13 changes: 13 additions & 0 deletions configs/mobilenet_v3/mobilenet-v3-d8-scratch_lraspp_4xb4-320k_ucsd-256x512.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,13 @@
_base_ = [
'../_base_/models/lraspp_m-v3-d8.py', '../_base_/datasets/ucsd_half-256x512.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_320k.py'
]
crop_size = (256, 512)
data_preprocessor = dict(size=crop_size)
# Re-config the data sampler.
model = dict(data_preprocessor=data_preprocessor)
train_dataloader = dict(batch_size=4, num_workers=4)
val_dataloader = dict(batch_size=1, num_workers=4)
test_dataloader = val_dataloader

runner = dict(type='IterBasedRunner', max_iters=320000)
13 changes: 13 additions & 0 deletions configs/mobilenet_v3/mobilenet-v3-d8-scratch_lraspp_4xb4-320k_ucsd-360x640.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,13 @@
_base_ = [
'../_base_/models/lraspp_m-v3-d8.py', '../_base_/datasets/ucsd_half-360x640.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_320k.py'
]
crop_size = (360, 640)
data_preprocessor = dict(size=crop_size)
# Re-config the data sampler.
model = dict(data_preprocessor=data_preprocessor)
train_dataloader = dict(batch_size=4, num_workers=4)
val_dataloader = dict(batch_size=1, num_workers=4)
test_dataloader = val_dataloader

runner = dict(type='IterBasedRunner', max_iters=320000)
13 changes: 13 additions & 0 deletions configs/mobilenet_v3/mobilenet-v3-d8-scratch_lraspp_4xb4-320k_ucsd-720x1280.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,13 @@
_base_ = [
'../_base_/models/lraspp_m-v3-d8.py', '../_base_/datasets/ucsd_half_4x4-512x1024.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_320k.py'
]
crop_size = (512, 1024)
data_preprocessor = dict(size=crop_size)
# Re-config the data sampler.
model = dict(data_preprocessor=data_preprocessor)
train_dataloader = dict(batch_size=4, num_workers=4)
val_dataloader = dict(batch_size=1, num_workers=4)
test_dataloader = val_dataloader

runner = dict(type='IterBasedRunner', max_iters=320000)
13 changes: 13 additions & 0 deletions configs/mobilenet_v3/mobilenet-v3-d8-scratch_lraspp_4xb4-320k_ucsd-720x720.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,13 @@
_base_ = [
'../_base_/models/lraspp_m-v3-d8.py', '../_base_/datasets/ucsd_half_4x4-512x512.py',
'../_base_/default_runtime.py', '../_base_/schedules/schedule_320k.py'
]
crop_size = (512, 512)
data_preprocessor = dict(size=crop_size)
# Re-config the data sampler.
model = dict(data_preprocessor=data_preprocessor)
train_dataloader = dict(batch_size=2, num_workers=4)
val_dataloader = dict(batch_size=1, num_workers=4)
test_dataloader = val_dataloader

runner = dict(type='IterBasedRunner', max_iters=320000)
4 changes: 2 additions & 2 deletions mmseg/datasets/mapillary.py
View file
Original file line number Diff line number Diff line change
@@ -58,7 +58,7 @@ class MapillaryDataset_v1(BaseSegDataset):
10], [0, 0, 0]])

def __init__(self,
img_suffix='.jpg',
img_suffix='.png',
seg_map_suffix='.png',
**kwargs) -> None:
super().__init__(
@@ -169,7 +169,7 @@ class MapillaryDataset_v2(BaseSegDataset):
[111, 111, 0], [0, 0, 0]])

def __init__(self,
img_suffix='.jpg',
img_suffix='.png',
seg_map_suffix='.png',
**kwargs) -> None:
super().__init__(