-
Notifications
You must be signed in to change notification settings - Fork 181
/
cfgs_res50_dota1.5_v11.py
105 lines (86 loc) · 2.79 KB
/
cfgs_res50_dota1.5_v11.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
# -*- 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 = 32000 * 2
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
DATASET_NAME = 'DOTA1.5'
CLASS_NUM = 16
# model
pretrain_zoo = PretrainModelZoo()
PRETRAINED_CKPT = pretrain_zoo.pretrain_weight_path(NET_NAME, ROOT_PATH)
TRAINED_CKPT = os.path.join(ROOT_PATH, 'output/trained_weights')
# loss
CLS_WEIGHT = 1.0
REG_WEIGHT = 2.0
REG_LOSS_MODE = 2 # GWD loss
GWD_TAU = 2.0
GWD_FUNC = tf.sqrt
VERSION = 'RetinaNet_DOTA1.5_GWD_2x_20210320'
"""
RetinaNet-H + gwd fix bug + sqrt + tau=2
FLOPs: 862193665; Trainable params: 33051321
This is your evaluation result for task 1:
mAP: 0.600346226078195
ap of each class:
plane:0.7939431922975534,
baseball-diamond:0.7446174663646932,
bridge:0.4195239451373103,
ground-track-field:0.6015244842574132,
small-vehicle:0.4997726370365391,
large-vehicle:0.6002761372691583,
ship:0.7616626473950818,
tennis-court:0.9031303299213864,
basketball-court:0.7175429727485252,
storage-tank:0.5848017376755156,
soccer-ball-field:0.48043525038685025,
roundabout:0.6779911622252887,
harbor:0.5579857368665273,
swimming-pool:0.6579145411010502,
helicopter:0.5135082856591359,
container-crane:0.09090909090909091
The submitted information is :
Description: RetinaNet_DOTA1.5_GWD_2x_20210320_83.2w
Username: SJTU-Det
Institute: SJTU
Emailadress: [email protected]
TeamMembers: yangxue
SWA
This is your evaluation result for task 1:
mAP: 0.6060358478025953
ap of each class:
plane:0.7967255029000768,
baseball-diamond:0.7429314386766407,
bridge:0.4386606848687486,
ground-track-field:0.61637498980047,
small-vehicle:0.4954853728442736,
large-vehicle:0.5942575317263504,
ship:0.7612793329714422,
tennis-court:0.8943097462172913,
basketball-court:0.7477277785488338,
storage-tank:0.591234002230469,
soccer-ball-field:0.4684316925081105,
roundabout:0.6707910130275366,
harbor:0.5665038559497408,
swimming-pool:0.6698656284482921,
helicopter:0.5484466070264743,
container-crane:0.0935483870967742
The submitted information is :
Description: RetinaNet_DOTA1.5_GWD_2x_20210320_swa12
Username: SJTU-Det
Institute: SJTU
Emailadress: [email protected]
TeamMembers: yangxue
"""