-
Notifications
You must be signed in to change notification settings - Fork 4
/
laban_normal_atis.txt
135 lines (120 loc) · 3.94 KB
/
laban_normal_atis.txt
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
126
127
128
129
130
131
132
133
134
135
Dataset to use: data/MixATIS_clean/raw_data_multi_ma_train.pkl
Dictionary to use: data/MixATIS_clean/intent2id_multi_ma_with_tokens.pkl
Data Mode: multi
Surface encoder mode: normal
Label aware layer mode: zero-shot
Use weights from models pretrained with in-domain data: False
Train from scratch...
====== epoch 1 / 15: ======
Average train loss: 11.5289
P = 0.2519, R = 0.4642, F1 = 0.3266
Accuracy: 0.06799878437927366
Average val loss: 2.4929
P = 0.9016, R = 0.7190, F1 = 0.8000
Accuracy: 0.45694444444444443
saving with loss of 1794.8778076171875 improved over previous 100
====== epoch 2 / 15: ======
Average train loss: 1.6771
P = 0.9107, R = 0.8610, F1 = 0.8852
Accuracy: 0.6252849111077344
Average val loss: 0.8430
P = 0.9677, R = 0.9298, F1 = 0.9484
Accuracy: 0.8375
saving with loss of 606.9461669921875 improved over previous 1794.8778076171875
====== epoch 3 / 15: ======
Average train loss: 0.4374
P = 0.9824, R = 0.9683, F1 = 0.9753
Accuracy: 0.9026743655979335
Average val loss: 0.9124
P = 0.9683, R = 0.9477, F1 = 0.9579
Accuracy: 0.8805555555555555
saving with loss of 656.9547119140625 improved over previous 606.9461669921875
====== epoch 4 / 15: ======
Average train loss: 0.2095
P = 0.9925, R = 0.9869, F1 = 0.9897
Accuracy: 0.9581370612368941
Average val loss: 1.1153
P = 0.9676, R = 0.9463, F1 = 0.9568
Accuracy: 0.8805555555555555
====== epoch 5 / 15: ======
Average train loss: 0.1317
P = 0.9959, R = 0.9920, F1 = 0.9939
Accuracy: 0.9751557514055614
Average val loss: 1.2700
P = 0.9636, R = 0.9477, F1 = 0.9556
Accuracy: 0.8902777777777777
saving with loss of 914.427001953125 improved over previous 656.9547119140625
====== epoch 6 / 15: ======
Average train loss: 0.0913
P = 0.9970, R = 0.9944, F1 = 0.9957
Accuracy: 0.9822975231727701
Average val loss: 1.2864
P = 0.9682, R = 0.9442, F1 = 0.9561
Accuracy: 0.8777777777777778
====== epoch 7 / 15: ======
Average train loss: 0.0711
P = 0.9976, R = 0.9955, F1 = 0.9966
Accuracy: 0.9856404801701869
Average val loss: 1.4799
P = 0.9590, R = 0.9497, F1 = 0.9543
Accuracy: 0.8763888888888889
====== epoch 8 / 15: ======
Average train loss: 0.0568
P = 0.9982, R = 0.9966, F1 = 0.9974
Accuracy: 0.9892113660537912
Average val loss: 1.2743
P = 0.9671, R = 0.9518, F1 = 0.9594
Accuracy: 0.8916666666666667
saving with loss of 917.5289306640625 improved over previous 914.427001953125
====== epoch 9 / 15: ======
Average train loss: 0.0476
P = 0.9986, R = 0.9971, F1 = 0.9979
Accuracy: 0.9914906549156663
Average val loss: 1.7224
P = 0.9696, R = 0.9456, F1 = 0.9575
Accuracy: 0.8875
====== epoch 10 / 15: ======
Average train loss: 0.0414
P = 0.9989, R = 0.9977, F1 = 0.9983
Accuracy: 0.993010180823583
Average val loss: 1.4452
P = 0.9636, R = 0.9470, F1 = 0.9552
Accuracy: 0.8930555555555556
saving with loss of 1040.5784912109375 improved over previous 917.5289306640625
====== epoch 11 / 15: ======
Average train loss: 0.0444
P = 0.9984, R = 0.9971, F1 = 0.9977
Accuracy: 0.9909588208478954
Average val loss: 1.2266
P = 0.9597, R = 0.9504, F1 = 0.9550
Accuracy: 0.8916666666666667
====== epoch 12 / 15: ======
Average train loss: 0.0448
P = 0.9981, R = 0.9973, F1 = 0.9977
Accuracy: 0.9908068682571037
Average val loss: 1.6467
P = 0.9609, R = 0.9477, F1 = 0.9542
Accuracy: 0.8805555555555555
====== epoch 13 / 15: ======
Average train loss: 0.0459
P = 0.9985, R = 0.9975, F1 = 0.9980
Accuracy: 0.9917945600972496
Average val loss: 1.2801
P = 0.9473, R = 0.9532, F1 = 0.9502
Accuracy: 0.8819444444444444
====== epoch 14 / 15: ======
Average train loss: 0.0240
P = 0.9991, R = 0.9987, F1 = 0.9989
Accuracy: 0.9957453274578332
Average val loss: 1.5600
P = 0.9569, R = 0.9477, F1 = 0.9522
Accuracy: 0.8916666666666667
====== epoch 15 / 15: ======
Average train loss: 0.0354
P = 0.9984, R = 0.9979, F1 = 0.9982
Accuracy: 0.9923263941650206
Average val loss: 1.3217
P = 0.9611, R = 0.9525, F1 = 0.9568
Accuracy: 0.8902777777777777
Best total val loss: 951.6284
Best Test Accuracy: 0.8931