-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathexp03.py
134 lines (134 loc) · 7.63 KB
/
exp03.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
# model settings, FasterRCNN
base = 'https://github.com/SwinTransformer/storage/releases/download/v1.0.0'
ckpt = '{}/swin_tiny_patch4_window7_224.pth'.format(base)
model = dict(type='FasterRCNN',
backbone=dict(type='SwinTransformer',
embed_dims=96,
depths=[2, 2, 6, 2],
num_heads=[3, 6, 12, 24],
drop_path_rate=0.2,
convert_weights=True,
init_cfg=dict(type='Pretrained', checkpoint=ckpt)),
neck=dict(type='FPN',
in_channels=[96, 192, 384, 768],
out_channels=256, num_outs=5),
rpn_head=dict(type='RPNHead',
in_channels=256,
anchor_generator=dict(type='AnchorGenerator',
scales=[8],
ratios=[0.5, 1.0, 2.0],
strides=[4, 8, 16, 32, 64]),
loss_bbox=dict(type='L1Loss', loss_weight=1.0)),
roi_head=dict(type='StandardRoIHead',
bbox_roi_extractor=dict(type='SingleRoIExtractor',
roi_layer=dict(type='RoIAlign',
output_size=7,
sampling_ratio=0),
out_channels=256,
featmap_strides=[4, 8, 16, 32]),
bbox_head=dict(type='Shared2FCBBoxHead',
in_channels=256,
num_classes=80,
loss_bbox=dict(type='L1Loss', loss_weight=1.0))),
# model training and testing settings
train_cfg=dict(rpn=dict(assigner=dict(type='MaxIoUAssigner',
pos_iou_thr=0.7,
neg_iou_thr=0.3,
min_pos_iou=0.3,
ignore_iof_thr=-1,
match_low_quality=True),
sampler=dict(type='RandomSampler',
num=256,
neg_pos_ub=-1,
pos_fraction=0.5,
add_gt_as_proposals=False),
allowed_border=-1, pos_weight=-1, debug=False),
rpn_proposal=dict(nms_pre=2000,
min_bbox_size=0,
max_per_img=1000,
nms=dict(type='nms', iou_threshold=0.7), ),
rcnn=dict(assigner=dict(type='MaxIoUAssigner',
pos_iou_thr=0.5,
neg_iou_thr=0.5,
min_pos_iou=0.5,
ignore_iof_thr=-1,
match_low_quality=False),
sampler=dict(type='RandomSampler',
num=512,
neg_pos_ub=-1,
pos_fraction=0.25,
add_gt_as_proposals=True),
pos_weight=-1, debug=False)),
test_cfg=dict(rpn=dict(nms_pre=1000,
min_bbox_size=0,
max_per_img=1000,
nms=dict(type='nms', iou_threshold=0.7)),
rcnn=dict(score_thr=0.05,
max_per_img=100,
nms=dict(type='nms', iou_threshold=0.5))))
# dataset settings
dataset_type = 'CocoDataset'
data_root = '../Dataset/COCO/'
samples_per_gpu = 4
workers_per_gpu = 8
image_norm = dict(mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
train_pipeline = [dict(type='LoadImageFromFile', to_float32=False, color_type='color'),
dict(type='LoadAnnotations', with_bbox=True, with_mask=False),
dict(type='Resize',
img_scale=[(1333, 640), (1333, 800)],
multiscale_mode='range', keep_ratio=True),
dict(type='RandomFlip', flip_ratio=0.5),
dict(type='Normalize', **image_norm),
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=(1333, 800),
flip=False,
transforms=[dict(type='Resize', keep_ratio=True),
dict(type='RandomFlip'),
dict(type='Normalize', **image_norm),
dict(type='Pad', size_divisor=32),
dict(type='ImageToTensor', keys=['img']),
dict(type='Collect', keys=['img'])])]
data = dict(samples_per_gpu=samples_per_gpu,
workers_per_gpu=workers_per_gpu,
train=dict(type='RepeatDataset', times=3,
dataset=dict(type=dataset_type,
ann_file=data_root + 'annotation/train2017.json',
img_prefix=data_root + 'images/train2017/',
pipeline=train_pipeline)),
val=dict(type=dataset_type,
ann_file=data_root + 'annotation/val2017.json',
img_prefix=data_root + 'images/val2017/',
pipeline=test_pipeline),
test=dict(type=dataset_type,
ann_file=data_root + 'annotation/val2017.json',
img_prefix=data_root + 'images/val2017/',
pipeline=test_pipeline))
optimizer = dict(type='AdamW', lr=0.0001,
betas=(0.9, 0.999), weight_decay=0.05,
paramwise_cfg=dict(custom_keys={'norm': dict(decay_mult=0.),
'absolute_pos_embed': dict(decay_mult=0.),
'relative_position_bias_table': dict(decay_mult=0.)}))
optimizer_config = dict(grad_clip=None)
fp16 = dict(loss_scale='dynamic')
lr_config = dict(step=[8, 11],
policy='step',
warmup='linear',
warmup_iters=1000,
warmup_ratio=0.001)
runner = dict(type='EpochBasedRunner', max_epochs=12)
checkpoint_config = dict(interval=12)
evaluation = dict(interval=12, metric='bbox')
log_config = dict(interval=50, hooks=[dict(type='TextLoggerHook')])
custom_hooks = [dict(type='NumClassCheckHook')]
dist_params = dict(backend='nccl')
log_level = 'INFO'
load_from = None
resume_from = None
workflow = [('train', 1)]
opencv_num_threads = 0
mp_start_method = 'fork'
auto_scale_lr = dict(enable=True, base_batch_size=16)