forked from csuhan/s2anet
-
Notifications
You must be signed in to change notification settings - Fork 0
/
cascade_s2anet_2s_r50_fpn_1x_dota_iouloss.py
194 lines (192 loc) · 6.29 KB
/
cascade_s2anet_2s_r50_fpn_1x_dota_iouloss.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
# model settings
model = dict(
type='CascadeS2ANetDetector',
pretrained='torchvision://resnet50',
num_stages=2,
backbone=dict(
type='ResNet',
depth=50,
num_stages=4,
out_indices=(0, 1, 2, 3),
frozen_stages=1,
style='pytorch'),
neck=dict(
type='FPN',
in_channels=[256, 512, 1024, 2048],
out_channels=256,
start_level=1,
add_extra_convs=True,
num_outs=5),
bbox_head=[
dict(
type='CascadeS2ANetHead',
num_classes=16,
in_channels=256,
feat_channels=256,
stacked_convs=2,
with_align=True,
anchor_scales=[4],
anchor_ratios=[1.0],
anchor_strides=[8, 16, 32, 64, 128],
anchor_base_sizes=None,
target_means=(.0, .0, .0, .0, .0),
target_stds=(1.0, 1.0, 1.0, 1.0, 1.0),
loss_cls=dict(
type='FocalLoss',
use_sigmoid=True,
gamma=2.0,
alpha=0.25,
loss_weight=1.0),
loss_bbox=dict(
type='RotatedIoULoss', loss_weight=1.0)),
dict(
type='CascadeS2ANetHead',
num_classes=16,
in_channels=256,
feat_channels=256,
stacked_convs=2,
with_align=True,
anchor_scales=[4],
anchor_ratios=[1.0],
anchor_strides=[8, 16, 32, 64, 128],
anchor_base_sizes=None,
target_means=(.0, .0, .0, .0, .0),
target_stds=(1.0, 1.0, 1.0, 1.0, 1.0),
loss_cls=dict(
type='FocalLoss',
use_sigmoid=True,
gamma=2.0,
alpha=0.25,
loss_weight=1.0),
loss_bbox=dict(
type='RotatedIoULoss', loss_weight=1.0)),
]
)
# training and testing settings
train_cfg = dict(
loss_weight=[1.0, 1.0],
stage_cfg=[
dict(
assigner=dict(
type='MaxIoUAssigner',
pos_iou_thr=0.5,
neg_iou_thr=0.4,
min_pos_iou=0,
ignore_iof_thr=-1,
iou_calculator=dict(type='BboxOverlaps2D_rotated')),
bbox_coder=dict(type='DeltaXYWHABBoxCoder',
target_means=(0., 0., 0., 0., 0.),
target_stds=(1., 1., 1., 1., 1.),
clip_border=True),
reg_decoded_bbox=True, # Set True to use IoULoss
allowed_border=-1,
pos_weight=-1,
debug=False),
dict(
assigner=dict(
type='MaxIoUAssigner',
pos_iou_thr=0.5,
neg_iou_thr=0.4,
min_pos_iou=0,
ignore_iof_thr=-1,
iou_calculator=dict(type='BboxOverlaps2D_rotated')),
bbox_coder=dict(type='DeltaXYWHABBoxCoder',
target_means=(0., 0., 0., 0., 0.),
target_stds=(1., 1., 1., 1., 1.),
clip_border=True),
reg_decoded_bbox=True, # Set True to use IoULoss
allowed_border=-1,
pos_weight=-1,
debug=False),
]
)
test_cfg = dict(
nms_pre=2000,
min_bbox_size=0,
score_thr=0.05,
nms=dict(type='nms_rotated', iou_thr=0.1),
max_per_img=2000)
# dataset settings
dataset_type = 'DotaDataset'
data_root = 'data/dota_1024/'
img_norm_cfg = dict(
mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
train_pipeline = [
dict(type='LoadImageFromFile'),
dict(type='LoadAnnotations', with_bbox=True),
dict(type='RotatedResize', img_scale=(1024, 1024), keep_ratio=True),
dict(type='RotatedRandomFlip', flip_ratio=0.5),
dict(type='Normalize', **img_norm_cfg),
dict(type='Pad', size_divisor=32),
dict(type='DefaultFormatBundle'),
dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']),
]
test_pipeline = [
dict(type='LoadImageFromFile'),
dict(
type='MultiScaleFlipAug',
img_scale=(1024, 1024),
flip=False,
transforms=[
dict(type='RotatedResize', img_scale=(1024, 1024), keep_ratio=True),
dict(type='RotatedRandomFlip'),
dict(type='Normalize', **img_norm_cfg),
dict(type='Pad', size_divisor=32),
dict(type='ImageToTensor', keys=['img']),
dict(type='Collect', keys=['img']),
])
]
data = dict(
imgs_per_gpu=2,
workers_per_gpu=2,
train=dict(
type=dataset_type,
ann_file=data_root + 'trainval_split/trainval_s2anet.pkl',
img_prefix=data_root + 'trainval_split/images/',
pipeline=train_pipeline),
val=dict(
type=dataset_type,
ann_file=data_root + 'trainval_split/trainval_s2anet.pkl',
img_prefix=data_root + 'trainval_split/images/',
pipeline=test_pipeline),
test=dict(
type=dataset_type,
ann_file=data_root + 'test_split/test_s2anet.pkl',
img_prefix=data_root + 'test_split/images/',
pipeline=test_pipeline))
evaluation = dict(
gt_dir='data/dota/test/labelTxt/', # change it to valset for offline validation
imagesetfile='data/dota/test/test.txt')
# optimizer
optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
# learning policy
lr_config = dict(
policy='step',
warmup='linear',
warmup_iters=500,
warmup_ratio=1.0 / 3,
step=[8, 11])
checkpoint_config = dict(interval=4)
log_config = dict(
interval=50,
hooks=[
dict(type='TextLoggerHook'),
])
# runtime settings
total_epochs = 12
dist_params = dict(backend='nccl')
log_level = 'INFO'
load_from = None
resume_from = None
workflow = [('train', 1)]
# Smooth L1 Loss (baseline)
# map: 0.7383014297396017
# classaps: [89.03541803 80.24197191 50.6001401 71.36392112 78.21320025 78.39095839
# 87.33035768 90.87532082 85.61042113 85.08971767 59.48388398 62.39758068
# 66.94123242 67.90225536 53.97576506]
# IoU Loss
# map: 0.7457866189214475
# classaps: [89.10383024 79.07287493 52.13029794 71.75779494 78.03327998 78.43329951
# 87.70381405 90.84512074 84.8341351 85.58334633 62.42832233 64.17414811
# 67.60661276 69.13690926 57.83614217]