-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathmulti-for-ko.out
More file actions
117 lines (111 loc) · 4.9 KB
/
multi-for-ko.out
File metadata and controls
117 lines (111 loc) · 4.9 KB
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
### Korean FrameNet ###
# contact: hahmyg@kaist, hahmyg@gmail.com #
###DEVICE: cuda:0
###multilingual-for-ko-masking
### Your result would be saved to: /disk/data/models/results/framenet/mulModel-100/en_ko_for_ko_with_masking_result.txt
### loading Korean FrameNet 1.1 data...
# of instances in training data: 17838
# of instances in dev data: 2548
# of instances in test data: 5097
# of instances in trn: 17838
# of instances in dev: 2548
# of instances in tst: 5097
data example: [['태풍', 'Hugo가', '남긴', '피해들과', '회사', '내', '몇몇', '주요', '부서들의', '저조한', '실적들을', '반영하여,', 'Aetna', 'Life', 'and', 'Casualty', 'Co.의', '3분기', '<tgt>', '순이익이', '</tgt>', '182.6', '백만', '달러', '또는', '주당', '1.63', '달러로', '22', '%', '하락하였다.'], ['_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '이익.n', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_'], ['_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', 'Earnings_and_losses', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_', '_'], ['O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'B-Earner', 'I-Earner', 'I-Earner', 'I-Earner', 'I-Earner', 'B-Time', 'X', 'O', 'X', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O', 'O']]
### EVALUATION
MODE: framenet
target LANGUAGE: ko
trained LANGUAGE: en_ko
Viterbi: False
masking: True
using TGT token: True
### model dir: /disk/data/models/dict_framenet/mulModel-100/36/
### TARGET LANGUAGE: ko
srl model: framenet
language: multilingual
version: 1.1
using viterbi: False
using masking: True
pretrained BERT: bert-base-multilingual-cased
using TGT special token: True
used dictionary:
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_lu2idx.json
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_lufrmap.json
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_bio_frargmap.json
...loaded model path: /disk/data/models/dict_framenet/mulModel-100/36/
/disk/data/models/dict_framenet/mulModel-100/36/
...model is loaded
# EPOCH: 36
SenseId Accuracy: 0.8112615263880714
ArgId Precision: 0.4539265742766837
ArgId Recall: 0.46195719411109737
ArgId F1: 0.45790667729474527
full-structure Precision: 0.6031634938409854
full-structure Recall: 0.6098216812906878
full-structure F1: 0.6064743138634764
-----processing time: 0hour:25min:35sec
### model dir: /disk/data/models/dict_framenet/mulModel-100/3/
### TARGET LANGUAGE: ko
srl model: framenet
language: multilingual
version: 1.1
using viterbi: False
using masking: True
pretrained BERT: bert-base-multilingual-cased
using TGT special token: True
used dictionary:
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_lu2idx.json
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_lufrmap.json
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_bio_frargmap.json
...loaded model path: /disk/data/models/dict_framenet/mulModel-100/3/
/disk/data/models/dict_framenet/mulModel-100/3/
...model is loaded
# EPOCH: 3
SenseId Accuracy: 0.8069452619187758
ArgId Precision: 0.395957358438322
ArgId Recall: 0.3538290238772733
ArgId F1: 0.37370965634391745
full-structure Precision: 0.5812081546678703
full-structure Recall: 0.546702519105576
full-structure F1: 0.5634275296262534
-----processing time: 0hour:43min:23sec
### model dir: /disk/data/models/dict_framenet/mulModel-100/15/
### TARGET LANGUAGE: ko
srl model: framenet
language: multilingual
version: 1.1
using viterbi: False
using masking: True
pretrained BERT: bert-base-multilingual-cased
using TGT special token: True
used dictionary:
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_lu2idx.json
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_lufrmap.json
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_bio_frargmap.json
...loaded model path: /disk/data/models/dict_framenet/mulModel-100/15/
/disk/data/models/dict_framenet/mulModel-100/15/
...model is loaded
# EPOCH: 15
SenseId Accuracy: 0.8161663723759074
ArgId Precision: 0.4414283990831222
ArgId Recall: 0.45267846096746256
ArgId F1: 0.44698265331053016
full-structure Precision: 0.5975337884910129
full-structure Recall: 0.6069204641947353
full-structure F1: 0.6021905497437338
-----processing time: 1hour:0min:19sec
### model dir: /disk/data/models/dict_framenet/mulModel-100/37/
### TARGET LANGUAGE: ko
srl model: framenet
language: multilingual
version: 1.1
using viterbi: False
using masking: True
pretrained BERT: bert-base-multilingual-cased
using TGT special token: True
used dictionary:
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_lu2idx.json
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_lufrmap.json
/disk/kaiser/kaiser/src/../koreanframenet/resource/info/mul_bio_frargmap.json
...loaded model path: /disk/data/models/dict_framenet/mulModel-100/37/
/disk/data/models/dict_framenet/mulModel-100/37/
...model is loaded