-
Notifications
You must be signed in to change notification settings - Fork 180
/
cfgs_res152_dota_bcd_v5.py
112 lines (93 loc) · 2.95 KB
/
cfgs_res152_dota_bcd_v5.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
# -*- coding: utf-8 -*-
from __future__ import division, print_function, absolute_import
import numpy as np
from alpharotate.utils.pretrain_zoo import PretrainModelZoo
from configs._base_.models.retinanet_r50_fpn import *
from configs._base_.datasets.dota_detection import *
from configs._base_.schedules.schedule_1x import *
# schedule
BATCH_SIZE = 1
GPU_GROUP = "0,1,2"
NUM_GPU = len(GPU_GROUP.strip().split(','))
LR = 1e-3
SAVE_WEIGHTS_INTE = 27000 * 2
DECAY_EPOCH = [36, 48, 60]
MAX_EPOCH = 51
DECAY_STEP = np.array(DECAY_EPOCH, np.int32) * SAVE_WEIGHTS_INTE
MAX_ITERATION = SAVE_WEIGHTS_INTE * MAX_EPOCH
WARM_EPOCH = 1. / 8.
WARM_SETP = int(WARM_EPOCH * SAVE_WEIGHTS_INTE)
# dataset
IMG_SHORT_SIDE_LEN = [800, 450, 500, 640, 700, 900, 1000, 1100, 1200]
IMG_MAX_LENGTH = 1200
# data augmentation
IMG_ROTATE = True
RGB2GRAY = True
VERTICAL_FLIP = True
HORIZONTAL_FLIP = True
IMAGE_PYRAMID = True
# model
NET_NAME = 'resnet152_v1d'
pretrain_zoo = PretrainModelZoo()
PRETRAINED_CKPT = pretrain_zoo.pretrain_weight_path(NET_NAME, ROOT_PATH)
TRAINED_CKPT = os.path.join(ROOT_PATH, 'output/trained_weights')
# bbox head
NUM_SUBNET_CONV = 5
# loss
CLS_WEIGHT = 1.0
REG_WEIGHT = 2.0
BCD_TAU = 2.0
BCD_FUNC = 0 # 0: sqrt 1: log
VERSION = 'RetinaNet_DOTA_BCD_6x_20210823'
"""
RetinaNet-H + 1-1/(sqrt(bcd)+2)
FLOPs: 1731833909; Trainable params: 68720548
This is your result for task 1:
mAP: 0.768573406585726
ap of each class:
plane:0.8880454724159974,
baseball-diamond:0.8440907408590022,
bridge:0.5373172162325457,
ground-track-field:0.702561153438066,
small-vehicle:0.778541230787683,
large-vehicle:0.7630502779604436,
ship:0.8518063491091422,
tennis-court:0.9082940506398122,
basketball-court:0.8590521594741499,
storage-tank:0.856084607613421,
soccer-ball-field:0.6477070835293367,
roundabout:0.6414710867102008,
harbor:0.765984221873397,
swimming-pool:0.7719071490415356,
helicopter:0.7126882991011564
The submitted information is :
Description: RetinaNet_DOTA_BCD_6x_20210823_275.4w_cpunms
Username: SJTU-Det
Institute: SJTU
Emailadress: yangxue-2019-sjtu@sjtu.edu.cn
TeamMembers: yangxue
This is your result for task 1:
mAP: 0.7851749645804825
ap of each class:
plane:0.8934950901089633,
baseball-diamond:0.8634613422890471,
bridge:0.547928015549363,
ground-track-field:0.8097701712658131,
small-vehicle:0.7778510118872703,
large-vehicle:0.7482730257889372,
ship:0.8418537972791365,
tennis-court:0.9078547385450373,
basketball-court:0.8626051059617449,
storage-tank:0.8574141297978217,
soccer-ball-field:0.7201907001872376,
roundabout:0.6918555163756489,
harbor:0.7720189968017138,
swimming-pool:0.7639497267610552,
helicopter:0.7191031001084477
The submitted information is :
Description: RetinaNet_DOTA_BCD_6x_20210823_275.4w_cpunms_mss
Username: yangxue
Institute: DetectionTeamUCAS
Emailadress: yangxue16@mails.ucas.ac.cn
TeamMembers: yangxue, yangjirui
"""