File size: 7,573 Bytes
0b422fb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
[i] AffinityNet@ResNet-50@Puzzle

[i] mean values is [0.485, 0.456, 0.406]
[i] std values is [0.229, 0.224, 0.225]
[i] The number of class is 20
[i] train_transform is Compose(
    <tools.ai.augment_utils.RandomResize_For_Segmentation object at 0x79be15df3430>
    <tools.ai.augment_utils.RandomHorizontalFlip_For_Segmentation object at 0x79be15df3400>
    <tools.ai.augment_utils.Normalize_For_Segmentation object at 0x79be15df3490>
    <tools.ai.augment_utils.RandomCrop_For_Segmentation object at 0x79be15df3550>
    <tools.ai.augment_utils.Transpose_For_Segmentation object at 0x79be15df35b0>
    <tools.ai.augment_utils.Resize_For_Mask object at 0x79be15df35e0>
)

[i] log_iteration : 33
[i] val_iteration : 330
[i] max_iteration : 990
[i] Architecture is resnet50
[i] Total Params: 23.63M

[i]                 iteration=33,                 learning_rate=0.0971,                 loss=0.5866,                 bg_loss=0.5234,                 fg_loss=0.7140,                 neg_loss=0.5546,                 time=21sec
[i]                 iteration=66,                 learning_rate=0.0941,                 loss=0.4096,                 bg_loss=0.3311,                 fg_loss=0.5322,                 neg_loss=0.3876,                 time=18sec
[i]                 iteration=99,                 learning_rate=0.0910,                 loss=0.3690,                 bg_loss=0.2907,                 fg_loss=0.4917,                 neg_loss=0.3468,                 time=18sec
[i]                 iteration=132,                 learning_rate=0.0880,                 loss=0.3705,                 bg_loss=0.2945,                 fg_loss=0.4954,                 neg_loss=0.3460,                 time=18sec
[i]                 iteration=165,                 learning_rate=0.0850,                 loss=0.3655,                 bg_loss=0.2897,                 fg_loss=0.4852,                 neg_loss=0.3436,                 time=19sec
[i]                 iteration=198,                 learning_rate=0.0819,                 loss=0.3529,                 bg_loss=0.2799,                 fg_loss=0.4650,                 neg_loss=0.3334,                 time=19sec
[i]                 iteration=231,                 learning_rate=0.0788,                 loss=0.3469,                 bg_loss=0.2809,                 fg_loss=0.4627,                 neg_loss=0.3220,                 time=19sec
[i]                 iteration=264,                 learning_rate=0.0757,                 loss=0.3695,                 bg_loss=0.2976,                 fg_loss=0.4803,                 neg_loss=0.3500,                 time=19sec
[i]                 iteration=297,                 learning_rate=0.0726,                 loss=0.3496,                 bg_loss=0.2698,                 fg_loss=0.4781,                 neg_loss=0.3252,                 time=19sec
[i]                 iteration=330,                 learning_rate=0.0695,                 loss=0.3384,                 bg_loss=0.2712,                 fg_loss=0.4460,                 neg_loss=0.3183,                 time=19sec
[i]                 iteration=363,                 learning_rate=0.0664,                 loss=0.3259,                 bg_loss=0.2599,                 fg_loss=0.4418,                 neg_loss=0.3010,                 time=21sec
[i]                 iteration=396,                 learning_rate=0.0632,                 loss=0.3375,                 bg_loss=0.2621,                 fg_loss=0.4632,                 neg_loss=0.3123,                 time=18sec
[i]                 iteration=429,                 learning_rate=0.0601,                 loss=0.3277,                 bg_loss=0.2583,                 fg_loss=0.4373,                 neg_loss=0.3076,                 time=18sec
[i]                 iteration=462,                 learning_rate=0.0569,                 loss=0.3313,                 bg_loss=0.2533,                 fg_loss=0.4549,                 neg_loss=0.3084,                 time=18sec
[i]                 iteration=495,                 learning_rate=0.0537,                 loss=0.3301,                 bg_loss=0.2494,                 fg_loss=0.4540,                 neg_loss=0.3085,                 time=19sec
[i]                 iteration=528,                 learning_rate=0.0505,                 loss=0.3229,                 bg_loss=0.2521,                 fg_loss=0.4341,                 neg_loss=0.3028,                 time=19sec
[i]                 iteration=561,                 learning_rate=0.0472,                 loss=0.3174,                 bg_loss=0.2464,                 fg_loss=0.4381,                 neg_loss=0.2925,                 time=19sec
[i]                 iteration=594,                 learning_rate=0.0439,                 loss=0.3270,                 bg_loss=0.2472,                 fg_loss=0.4452,                 neg_loss=0.3079,                 time=19sec
[i]                 iteration=627,                 learning_rate=0.0406,                 loss=0.3237,                 bg_loss=0.2511,                 fg_loss=0.4465,                 neg_loss=0.2987,                 time=19sec
[i]                 iteration=660,                 learning_rate=0.0373,                 loss=0.3258,                 bg_loss=0.2443,                 fg_loss=0.4472,                 neg_loss=0.3058,                 time=19sec
[i]                 iteration=693,                 learning_rate=0.0339,                 loss=0.3254,                 bg_loss=0.2507,                 fg_loss=0.4396,                 neg_loss=0.3056,                 time=21sec
[i]                 iteration=726,                 learning_rate=0.0305,                 loss=0.3242,                 bg_loss=0.2441,                 fg_loss=0.4472,                 neg_loss=0.3027,                 time=18sec
[i]                 iteration=759,                 learning_rate=0.0271,                 loss=0.3185,                 bg_loss=0.2413,                 fg_loss=0.4289,                 neg_loss=0.3019,                 time=19sec
[i]                 iteration=792,                 learning_rate=0.0236,                 loss=0.3287,                 bg_loss=0.2424,                 fg_loss=0.4540,                 neg_loss=0.3091,                 time=19sec
[i]                 iteration=825,                 learning_rate=0.0200,                 loss=0.3113,                 bg_loss=0.2352,                 fg_loss=0.4350,                 neg_loss=0.2875,                 time=19sec
[i]                 iteration=858,                 learning_rate=0.0164,                 loss=0.3110,                 bg_loss=0.2417,                 fg_loss=0.4271,                 neg_loss=0.2876,                 time=19sec
[i]                 iteration=891,                 learning_rate=0.0127,                 loss=0.3182,                 bg_loss=0.2422,                 fg_loss=0.4330,                 neg_loss=0.2987,                 time=19sec
[i]                 iteration=924,                 learning_rate=0.0089,                 loss=0.3107,                 bg_loss=0.2489,                 fg_loss=0.4280,                 neg_loss=0.2829,                 time=19sec
[i]                 iteration=957,                 learning_rate=0.0048,                 loss=0.3141,                 bg_loss=0.2364,                 fg_loss=0.4319,                 neg_loss=0.2940,                 time=19sec
[i]                 iteration=990,                 learning_rate=0.0002,                 loss=0.3115,                 bg_loss=0.2404,                 fg_loss=0.4173,                 neg_loss=0.2942,                 time=19sec