Hugging Face's logo

ConvLab
/
mle-policy-multiwoz21

English
dialogue policy
task-oriented dialog
Model card Files Community
ChrisGeishauser commited on
Commit
1e7648e
•
1 Parent(s): d6cb9ea

Upload 3 files

Browse files
Files changed (3) hide show
  1. config_saved.json +1 -0
  2. supervised.pol.mdl +0 -0
  3. train_INFO.log +205 -0
config_saved.json ADDED
@@ -0,0 +1 @@
 
1
+ {"args": {"seed": 0, "eval_freq": 2, "dataset_name": "multiwoz21"}, "config": {"batchsz": 32, "epoch": 24, "lr_supervised": 0.0001, "save_dir": "save", "log_dir": "log", "print_per_batch": 400, "save_per_epoch": 1, "h_dim": 100, "load": "save/best", "pos_weight": 5, "hidden_size": 256, "weight_decay": 1e-05, "lambda": 1, "tau": 0.005, "policy_freq": 2, "entropy_weight": 0.001}, "policy_config": null}
supervised.pol.mdl ADDED
Binary file (271 kB). View file
train_INFO.log ADDED
@@ -0,0 +1,205 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Visible device: cuda
2
+ Seed used: 0
3
+ Vectorizer: Data set used is multiwoz21
4
+ Start training
5
+ Epoch: 0
6
+ Precision: 0
7
+ Recall: 0
8
+ F1: 0
9
+ Best Precision: 0.0
10
+ Best Recall: 0.0
11
+ Best F1: 0.0
12
+ Epoch: 1
13
+ Precision: 0
14
+ Recall: 0
15
+ F1: 0
16
+ Best Precision: 0.0
17
+ Best Recall: 0.0
18
+ Best F1: 0.0
19
+ Epoch: 2
20
+ Average actions: 3.803938627243042
21
+ Average target actions: 2.6072394847869873
22
+ Precision: 0.36443668246783334
23
+ Recall: 0.5317489209007229
24
+ F1: 0.43247472824937616
25
+ <<dialog policy>> epoch 2: saved network to mdl
26
+ Best Precision: 0.36443668246783334
27
+ Best Recall: 0.5317489209007229
28
+ Best F1: 0.43247472824937616
29
+ Epoch: 3
30
+ Precision: 0.36443668246783334
31
+ Recall: 0.5317489209007229
32
+ F1: 0.43247472824937616
33
+ Best Precision: 0.36443668246783334
34
+ Best Recall: 0.5317489209007229
35
+ Best F1: 0.43247472824937616
36
+ Epoch: 4
37
+ Average actions: 4.113307952880859
38
+ Average target actions: 2.6075873374938965
39
+ Precision: 0.3832530835696854
40
+ Recall: 0.6043475999791981
41
+ F1: 0.46905208774797685
42
+ <<dialog policy>> epoch 4: saved network to mdl
43
+ Best Precision: 0.3832530835696854
44
+ Best Recall: 0.6043475999791981
45
+ Best F1: 0.46905208774797685
46
+ Epoch: 5
47
+ Precision: 0.3832530835696854
48
+ Recall: 0.6043475999791981
49
+ F1: 0.46905208774797685
50
+ Best Precision: 0.3832530835696854
51
+ Best Recall: 0.6043475999791981
52
+ Best F1: 0.46905208774797685
53
+ Epoch: 6
54
+ Average actions: 4.202342510223389
55
+ Average target actions: 2.6075873374938965
56
+ Precision: 0.3931234866828087
57
+ Recall: 0.6332622601279317
58
+ F1: 0.4851007887817704
59
+ <<dialog policy>> epoch 6: saved network to mdl
60
+ Best Precision: 0.3931234866828087
61
+ Best Recall: 0.6332622601279317
62
+ Best F1: 0.4851007887817704
63
+ Epoch: 7
64
+ Precision: 0.3931234866828087
65
+ Recall: 0.6332622601279317
66
+ F1: 0.4851007887817704
67
+ Best Precision: 0.3931234866828087
68
+ Best Recall: 0.6332622601279317
69
+ Best F1: 0.4851007887817704
70
+ Epoch: 8
71
+ Average actions: 4.356949806213379
72
+ Average target actions: 2.6075873374938965
73
+ Precision: 0.3951788491446345
74
+ Recall: 0.6607207863123408
75
+ F1: 0.4945600342552405
76
+ <<dialog policy>> epoch 8: saved network to mdl
77
+ Best Precision: 0.3951788491446345
78
+ Best Recall: 0.6607207863123408
79
+ Best F1: 0.4945600342552405
80
+ Epoch: 9
81
+ Precision: 0.3951788491446345
82
+ Recall: 0.6607207863123408
83
+ F1: 0.4945600342552405
84
+ Best Precision: 0.3951788491446345
85
+ Best Recall: 0.6607207863123408
86
+ Best F1: 0.4945600342552405
87
+ Epoch: 10
88
+ Average actions: 4.292381763458252
89
+ Average target actions: 2.6075873374938965
90
+ Precision: 0.4069264069264069
91
+ Recall: 0.6697176140204899
92
+ F1: 0.5062504913908326
93
+ <<dialog policy>> epoch 10: saved network to mdl
94
+ Best Precision: 0.4069264069264069
95
+ Best Recall: 0.6697176140204899
96
+ Best F1: 0.5062504913908326
97
+ Epoch: 11
98
+ Precision: 0.4069264069264069
99
+ Recall: 0.6697176140204899
100
+ F1: 0.5062504913908326
101
+ Best Precision: 0.4069264069264069
102
+ Best Recall: 0.6697176140204899
103
+ Best F1: 0.5062504913908326
104
+ Epoch: 12
105
+ Average actions: 4.411757946014404
106
+ Average target actions: 2.608457088470459
107
+ Precision: 0.4065842862412394
108
+ Recall: 0.6878672837901086
109
+ F1: 0.5110797704835687
110
+ <<dialog policy>> epoch 12: saved network to mdl
111
+ Best Precision: 0.4069264069264069
112
+ Best Recall: 0.6878672837901086
113
+ Best F1: 0.5110797704835687
114
+ Epoch: 13
115
+ Precision: 0.4065842862412394
116
+ Recall: 0.6878672837901086
117
+ F1: 0.5110797704835687
118
+ Best Precision: 0.4069264069264069
119
+ Best Recall: 0.6878672837901086
120
+ Best F1: 0.5110797704835687
121
+ Epoch: 14
122
+ Average actions: 4.343286514282227
123
+ Average target actions: 2.608804702758789
124
+ Precision: 0.4146211979264256
125
+ Recall: 0.6904675230121171
126
+ F1: 0.5181167196737625
127
+ <<dialog policy>> epoch 14: saved network to mdl
128
+ Best Precision: 0.4146211979264256
129
+ Best Recall: 0.6904675230121171
130
+ Best F1: 0.5181167196737625
131
+ Epoch: 15
132
+ Precision: 0.4146211979264256
133
+ Recall: 0.6904675230121171
134
+ F1: 0.5181167196737625
135
+ Best Precision: 0.4146211979264256
136
+ Best Recall: 0.6904675230121171
137
+ Best F1: 0.5181167196737625
138
+ Epoch: 16
139
+ Average actions: 4.276244640350342
140
+ Average target actions: 2.608457088470459
141
+ Precision: 0.4216435662406039
142
+ Recall: 0.6913516043476
143
+ F1: 0.5238189053942235
144
+ <<dialog policy>> epoch 16: saved network to mdl
145
+ Best Precision: 0.4216435662406039
146
+ Best Recall: 0.6913516043476
147
+ Best F1: 0.5238189053942235
148
+ Epoch: 17
149
+ Precision: 0.4216435662406039
150
+ Recall: 0.6913516043476
151
+ F1: 0.5238189053942235
152
+ Best Precision: 0.4216435662406039
153
+ Best Recall: 0.6913516043476
154
+ Best F1: 0.5238189053942235
155
+ Epoch: 18
156
+ Average actions: 4.305194854736328
157
+ Average target actions: 2.6089789867401123
158
+ Precision: 0.4217372134038801
159
+ Recall: 0.6963960684382964
160
+ F1: 0.5253329671838528
161
+ <<dialog policy>> epoch 18: saved network to mdl
162
+ Best Precision: 0.4217372134038801
163
+ Best Recall: 0.6963960684382964
164
+ Best F1: 0.5253329671838528
165
+ Epoch: 19
166
+ Precision: 0.4217372134038801
167
+ Recall: 0.6963960684382964
168
+ F1: 0.5253329671838528
169
+ Best Precision: 0.4217372134038801
170
+ Best Recall: 0.6963960684382964
171
+ Best F1: 0.5253329671838528
172
+ Epoch: 20
173
+ Average actions: 4.321138858795166
174
+ Average target actions: 2.6060221195220947
175
+ Precision: 0.42330383480825956
176
+ Recall: 0.7014925373134329
177
+ F1: 0.5279968685781387
178
+ <<dialog policy>> epoch 20: saved network to mdl
179
+ Best Precision: 0.42330383480825956
180
+ Best Recall: 0.7014925373134329
181
+ Best F1: 0.5279968685781387
182
+ Epoch: 21
183
+ Precision: 0.42330383480825956
184
+ Recall: 0.7014925373134329
185
+ F1: 0.5279968685781387
186
+ Best Precision: 0.42330383480825956
187
+ Best Recall: 0.7014925373134329
188
+ Best F1: 0.5279968685781387
189
+ Epoch: 22
190
+ Average actions: 4.332869529724121
191
+ Average target actions: 2.6077613830566406
192
+ Precision: 0.42460478948192204
193
+ Recall: 0.7053928961464455
194
+ F1: 0.5301129479813969
195
+ <<dialog policy>> epoch 22: saved network to mdl
196
+ Best Precision: 0.42460478948192204
197
+ Best Recall: 0.7053928961464455
198
+ Best F1: 0.5301129479813969
199
+ Epoch: 23
200
+ Precision: 0.42460478948192204
201
+ Recall: 0.7053928961464455
202
+ F1: 0.5301129479813969
203
+ Best Precision: 0.42460478948192204
204
+ Best Recall: 0.7053928961464455
205
+ Best F1: 0.5301129479813969