centernet_resnet18_dcnv2_140e_coco.py
3.95 KB
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
_base_ = [
'../_base_/datasets/coco_detection.py',
'../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py'
]
model = dict(
type='CenterNet',
backbone=dict(
type='ResNet',
depth=18,
norm_eval=False,
norm_cfg=dict(type='BN'),
init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet18')),
neck=dict(
type='CTResNetNeck',
in_channel=512,
num_deconv_filters=(256, 128, 64),
num_deconv_kernels=(4, 4, 4),
use_dcn=True),
bbox_head=dict(
type='CenterNetHead',
num_classes=80,
in_channel=64,
feat_channel=64,
loss_center_heatmap=dict(type='GaussianFocalLoss', loss_weight=1.0),
loss_wh=dict(type='L1Loss', loss_weight=0.1),
loss_offset=dict(type='L1Loss', loss_weight=1.0)),
train_cfg=None,
test_cfg=dict(topk=100, local_maximum_kernel=3, max_per_img=100))
# We fixed the incorrect img_norm_cfg problem in the source code.
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', to_float32=True, color_type='color'),
dict(type='LoadAnnotations', with_bbox=True),
dict(
type='PhotoMetricDistortion',
brightness_delta=32,
contrast_range=(0.5, 1.5),
saturation_range=(0.5, 1.5),
hue_delta=18),
dict(
type='RandomCenterCropPad',
crop_size=(512, 512),
ratios=(0.6, 0.7, 0.8, 0.9, 1.0, 1.1, 1.2, 1.3),
mean=[0, 0, 0],
std=[1, 1, 1],
to_rgb=True,
test_pad_mode=None),
dict(type='Resize', img_scale=(512, 512), keep_ratio=True),
dict(type='RandomFlip', flip_ratio=0.5),
dict(type='Normalize', **img_norm_cfg),
dict(type='DefaultFormatBundle'),
dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels'])
]
test_pipeline = [
dict(type='LoadImageFromFile', to_float32=True),
dict(
type='MultiScaleFlipAug',
scale_factor=1.0,
flip=False,
transforms=[
dict(type='Resize', keep_ratio=True),
dict(
type='RandomCenterCropPad',
ratios=None,
border=None,
mean=[0, 0, 0],
std=[1, 1, 1],
to_rgb=True,
test_mode=True,
test_pad_mode=['logical_or', 31],
test_pad_add_pix=1),
dict(type='RandomFlip'),
dict(type='Normalize', **img_norm_cfg),
dict(type='DefaultFormatBundle'),
dict(
type='Collect',
meta_keys=('filename', 'ori_shape', 'img_shape', 'pad_shape',
'scale_factor', 'flip', 'flip_direction',
'img_norm_cfg', 'border'),
keys=['img'])
])
]
dataset_type = 'CocoDataset'
data_root = 'data/coco/'
# Use RepeatDataset to speed up training
data = dict(
samples_per_gpu=16,
workers_per_gpu=4,
train=dict(
_delete_=True,
type='RepeatDataset',
times=5,
dataset=dict(
type=dataset_type,
ann_file=data_root + 'annotations/instances_train2017.json',
img_prefix=data_root + 'train2017/',
pipeline=train_pipeline)),
val=dict(pipeline=test_pipeline),
test=dict(pipeline=test_pipeline))
# optimizer
# Based on the default settings of modern detectors, the SGD effect is better
# than the Adam in the source code, so we use SGD default settings and
# if you use adam+lr5e-4, the map is 29.1.
optimizer_config = dict(
_delete_=True, grad_clip=dict(max_norm=35, norm_type=2))
# learning policy
# Based on the default settings of modern detectors, we added warmup settings.
lr_config = dict(
policy='step',
warmup='linear',
warmup_iters=1000,
warmup_ratio=1.0 / 1000,
step=[18, 24]) # the real step is [18*5, 24*5]
runner = dict(max_epochs=28) # the real epoch is 28*5=140