-
Notifications
You must be signed in to change notification settings - Fork 1
/
train.log
90 lines (90 loc) · 3.96 KB
/
train.log
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
--------------------------------------
some config:
data_dir = ./data
output_dir = ./output
cache_dir = ./embedding
embedding_path = ./embedding/GoogleNews-vectors-negative300.txt
word_dim = 300
min_freq = 0
model_name = Attention_CNN
mode = 1
seed = 5782
cuda = 0
epoch = 30
dropout = 0.5
batch_size = 32
lr = 1.0
max_len = 100
pos_dis = 50
pos_dim = 5
tag_dim = 10
filter_num = 100
window = 3
device = cuda:0
model_dir = ./output/Attention_CNN
--------------------------------------
start to load data ...
embedding scale: 19215*300d
finish loading embeddng!
finish!
--------------------------------------
Attention_CNN(
(word_embedding): Embedding(19216, 300)
(pos1_embedding): Embedding(103, 5)
(pos2_embedding): Embedding(103, 5)
(tag_embedding): Embedding(16, 10)
(conv): Conv2d(1, 100, kernel_size=(3, 320), stride=(1, 1), padding=(1, 0))
(maxpool): MaxPool2d(kernel_size=(100, 1), stride=(100, 1), padding=0, dilation=1, ceil_mode=False)
(tanh): Tanh()
(we): Linear(in_features=640, out_features=640, bias=True)
(wa): Linear(in_features=640, out_features=1, bias=True)
(dense): Linear(in_features=740, out_features=19, bias=True)
)
traning model parameters:
word_embedding.weight : torch.Size([19216, 300])
pos1_embedding.weight : torch.Size([103, 5])
pos2_embedding.weight : torch.Size([103, 5])
tag_embedding.weight : torch.Size([16, 10])
conv.weight : torch.Size([100, 1, 3, 320])
conv.bias : torch.Size([100])
we.weight : torch.Size([640, 640])
we.bias : torch.Size([640])
wa.weight : torch.Size([1, 640])
wa.bias : torch.Size([1])
dense.weight : torch.Size([19, 740])
dense.bias : torch.Size([19])
--------------------------------------
start to train the model ...
[001] train_loss: 0.705 | dev_loss: 0.925 | micro f1 on dev: 0.7585 >>> save models!
[002] train_loss: 0.322 | dev_loss: 0.855 | micro f1 on dev: 0.7511
[003] train_loss: 0.120 | dev_loss: 0.835 | micro f1 on dev: 0.7933 >>> save models!
[004] train_loss: 0.029 | dev_loss: 0.825 | micro f1 on dev: 0.7976 >>> save models!
[005] train_loss: 0.010 | dev_loss: 0.829 | micro f1 on dev: 0.8065 >>> save models!
[006] train_loss: 0.006 | dev_loss: 0.848 | micro f1 on dev: 0.8110 >>> save models!
[007] train_loss: 0.005 | dev_loss: 0.861 | micro f1 on dev: 0.8110
[008] train_loss: 0.004 | dev_loss: 0.875 | micro f1 on dev: 0.8114 >>> save models!
[009] train_loss: 0.003 | dev_loss: 0.884 | micro f1 on dev: 0.8149 >>> save models!
[010] train_loss: 0.003 | dev_loss: 0.892 | micro f1 on dev: 0.8137
[011] train_loss: 0.003 | dev_loss: 0.899 | micro f1 on dev: 0.8142
[012] train_loss: 0.002 | dev_loss: 0.905 | micro f1 on dev: 0.8153 >>> save models!
[013] train_loss: 0.002 | dev_loss: 0.910 | micro f1 on dev: 0.8142
[014] train_loss: 0.002 | dev_loss: 0.916 | micro f1 on dev: 0.8152
[015] train_loss: 0.002 | dev_loss: 0.920 | micro f1 on dev: 0.8150
[016] train_loss: 0.002 | dev_loss: 0.925 | micro f1 on dev: 0.8153 >>> save models!
[017] train_loss: 0.002 | dev_loss: 0.928 | micro f1 on dev: 0.8152
[018] train_loss: 0.002 | dev_loss: 0.933 | micro f1 on dev: 0.8153
[019] train_loss: 0.002 | dev_loss: 0.935 | micro f1 on dev: 0.8156 >>> save models!
[020] train_loss: 0.002 | dev_loss: 0.937 | micro f1 on dev: 0.8155
[021] train_loss: 0.002 | dev_loss: 0.941 | micro f1 on dev: 0.8156 >>> save models!
[022] train_loss: 0.002 | dev_loss: 0.944 | micro f1 on dev: 0.8155
[023] train_loss: 0.001 | dev_loss: 0.945 | micro f1 on dev: 0.8155
[024] train_loss: 0.001 | dev_loss: 0.946 | micro f1 on dev: 0.8148
[025] train_loss: 0.001 | dev_loss: 0.950 | micro f1 on dev: 0.8153
[026] train_loss: 0.001 | dev_loss: 0.952 | micro f1 on dev: 0.8146
[027] train_loss: 0.001 | dev_loss: 0.953 | micro f1 on dev: 0.8146
[028] train_loss: 0.001 | dev_loss: 0.954 | micro f1 on dev: 0.8148
[029] train_loss: 0.001 | dev_loss: 0.956 | micro f1 on dev: 0.8148
[030] train_loss: 0.001 | dev_loss: 0.957 | micro f1 on dev: 0.8146
--------------------------------------
start test ...
test_loss: 0.941 | micro f1 on test: 0.8156