File size: 6,259 Bytes
61d092a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
<--Spectrogram-->
Type   : <class 'dict'>
Length : 11138
Length : 319287046
Shape one element : (752, 400)

-----------------------------

<--EEG-->
Type   : <class 'dict'>
Length : 20183
Length : 127492639
Shape one element : (128, 256, 4)
Creadting 5 folds of the training data
========== fold: 0 training ==========
Epoch 1 - avg_train_loss: 0.4490  avg_val_loss: 0.4243  time: 125s
Epoch 1 - Save Best valid loss: 0.4243 Model
Epoch 2 - avg_train_loss: 0.3431  avg_val_loss: 0.3847  time: 120s
Epoch 2 - Save Best valid loss: 0.3847 Model
Epoch 3 - avg_train_loss: 0.3024  avg_val_loss: 0.3708  time: 120s
Epoch 3 - Save Best valid loss: 0.3708 Model
Epoch 4 - avg_train_loss: 0.2752  avg_val_loss: 0.3599  time: 120s
Epoch 4 - Save Best valid loss: 0.3599 Model
Epoch 5 - avg_train_loss: 0.2533  avg_val_loss: 0.3708  time: 120s
========== fold: 0 result ==========
Score with best loss weights stage1: 0.3598788937469328
========== fold: 1 training ==========
Epoch 1 - avg_train_loss: 0.4582  avg_val_loss: 0.3886  time: 121s
Epoch 1 - Save Best valid loss: 0.3886 Model
Epoch 2 - avg_train_loss: 0.3445  avg_val_loss: 0.3738  time: 120s
Epoch 2 - Save Best valid loss: 0.3738 Model
Epoch 3 - avg_train_loss: 0.3068  avg_val_loss: 0.3503  time: 120s
Epoch 3 - Save Best valid loss: 0.3503 Model
Epoch 4 - avg_train_loss: 0.2764  avg_val_loss: 0.3595  time: 120s
Epoch 5 - avg_train_loss: 0.2570  avg_val_loss: 0.3453  time: 120s
Epoch 5 - Save Best valid loss: 0.3453 Model
========== fold: 1 result ==========
Score with best loss weights stage1: 0.3452556534977812
========== fold: 2 training ==========
Epoch 1 - avg_train_loss: 0.4569  avg_val_loss: 0.3837  time: 121s
Epoch 1 - Save Best valid loss: 0.3837 Model
Epoch 2 - avg_train_loss: 0.3405  avg_val_loss: 0.3605  time: 120s
Epoch 2 - Save Best valid loss: 0.3605 Model
Epoch 3 - avg_train_loss: 0.2994  avg_val_loss: 0.3407  time: 121s
Epoch 3 - Save Best valid loss: 0.3407 Model
Epoch 4 - avg_train_loss: 0.2729  avg_val_loss: 0.3402  time: 120s
Epoch 4 - Save Best valid loss: 0.3402 Model
Epoch 5 - avg_train_loss: 0.2489  avg_val_loss: 0.3374  time: 121s
Epoch 5 - Save Best valid loss: 0.3374 Model
========== fold: 2 result ==========
Score with best loss weights stage1: 0.3373754414364581
========== fold: 3 training ==========
Epoch 1 - avg_train_loss: 0.4586  avg_val_loss: 0.3846  time: 121s
Epoch 1 - Save Best valid loss: 0.3846 Model
Epoch 2 - avg_train_loss: 0.3424  avg_val_loss: 0.3545  time: 120s
Epoch 2 - Save Best valid loss: 0.3545 Model
Epoch 3 - avg_train_loss: 0.3046  avg_val_loss: 0.3353  time: 120s
Epoch 3 - Save Best valid loss: 0.3353 Model
Epoch 4 - avg_train_loss: 0.2746  avg_val_loss: 0.3511  time: 120s
Epoch 5 - avg_train_loss: 0.2525  avg_val_loss: 0.3423  time: 120s
========== fold: 3 result ==========
Score with best loss weights stage1: 0.3353293751642418
========== fold: 4 training ==========
Epoch 1 - avg_train_loss: 0.4668  avg_val_loss: 0.3785  time: 121s
Epoch 1 - Save Best valid loss: 0.3785 Model
Epoch 2 - avg_train_loss: 0.3545  avg_val_loss: 0.3445  time: 119s
Epoch 2 - Save Best valid loss: 0.3445 Model
Epoch 3 - avg_train_loss: 0.3113  avg_val_loss: 0.3239  time: 120s
Epoch 3 - Save Best valid loss: 0.3239 Model
Epoch 4 - avg_train_loss: 0.2863  avg_val_loss: 0.3295  time: 120s
Epoch 5 - avg_train_loss: 0.2586  avg_val_loss: 0.3275  time: 120s
========== fold: 4 result ==========
Score with best loss weights stage1: 0.3239006738005828
========== CV ==========
Score with best loss weights stage1: 0.34034800752919936
========== fold: 0 training ==========
Epoch 1 - avg_train_loss: 0.2589  avg_val_loss: 0.4813  time: 48s
Epoch 1 - Save Best valid loss: 0.4813 Model
Epoch 2 - avg_train_loss: 0.2118  avg_val_loss: 0.5373  time: 48s
Epoch 3 - avg_train_loss: 0.1913  avg_val_loss: 0.4766  time: 47s
Epoch 3 - Save Best valid loss: 0.4766 Model
Epoch 4 - avg_train_loss: 0.1719  avg_val_loss: 0.4972  time: 47s
Epoch 5 - avg_train_loss: 0.1586  avg_val_loss: 0.5009  time: 47s
========== fold: 0 result ==========
Score with best loss weights stage2: 0.47662054733501263
========== fold: 1 training ==========
Epoch 1 - avg_train_loss: 0.2335  avg_val_loss: 0.4232  time: 50s
Epoch 1 - Save Best valid loss: 0.4232 Model
Epoch 2 - avg_train_loss: 0.1913  avg_val_loss: 0.4603  time: 48s
Epoch 3 - avg_train_loss: 0.1755  avg_val_loss: 0.4866  time: 49s
Epoch 4 - avg_train_loss: 0.1574  avg_val_loss: 0.4770  time: 49s
Epoch 5 - avg_train_loss: 0.1445  avg_val_loss: 0.4739  time: 48s
========== fold: 1 result ==========
Score with best loss weights stage2: 0.4232155225685631
========== fold: 2 training ==========
Epoch 1 - avg_train_loss: 0.2305  avg_val_loss: 0.4416  time: 50s
Epoch 1 - Save Best valid loss: 0.4416 Model
Epoch 2 - avg_train_loss: 0.1891  avg_val_loss: 0.4643  time: 49s
Epoch 3 - avg_train_loss: 0.1665  avg_val_loss: 0.4959  time: 49s
Epoch 4 - avg_train_loss: 0.1552  avg_val_loss: 0.4505  time: 49s
Epoch 5 - avg_train_loss: 0.1433  avg_val_loss: 0.4583  time: 49s
========== fold: 2 result ==========
Score with best loss weights stage2: 0.4416244997448457
========== fold: 3 training ==========
Epoch 1 - avg_train_loss: 0.2756  avg_val_loss: 0.4538  time: 50s
Epoch 1 - Save Best valid loss: 0.4538 Model
Epoch 2 - avg_train_loss: 0.2203  avg_val_loss: 0.4531  time: 49s
Epoch 2 - Save Best valid loss: 0.4531 Model
Epoch 3 - avg_train_loss: 0.1964  avg_val_loss: 0.4876  time: 50s
Epoch 4 - avg_train_loss: 0.1778  avg_val_loss: 0.4753  time: 49s
Epoch 5 - avg_train_loss: 0.1682  avg_val_loss: 0.4917  time: 49s
========== fold: 3 result ==========
Score with best loss weights stage2: 0.453083156121616
========== fold: 4 training ==========
Epoch 1 - avg_train_loss: 0.2741  avg_val_loss: 0.4381  time: 50s
Epoch 1 - Save Best valid loss: 0.4381 Model
Epoch 2 - avg_train_loss: 0.2243  avg_val_loss: 0.4424  time: 49s
Epoch 3 - avg_train_loss: 0.2020  avg_val_loss: 0.4471  time: 49s
Epoch 4 - avg_train_loss: 0.1878  avg_val_loss: 0.4826  time: 50s
Epoch 5 - avg_train_loss: 0.1698  avg_val_loss: 0.4790  time: 49s
========== fold: 4 result ==========
Score with best loss weights stage2: 0.4381120401343222
========== CV ==========
Score with best loss weights stage2: 0.44653115318087194