Davlan commited on
Commit
1f7d5c7
1 Parent(s): ecc33e5

Upload trainer_state.json

Browse files
Files changed (1) hide show
  1. trainer_state.json +2815 -0
trainer_state.json ADDED
@@ -0,0 +1,2815 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.0,
5
+ "global_step": 232677,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 4.989255491518285e-05,
13
+ "loss": 2.956,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 4.97851098303657e-05,
19
+ "loss": 2.4622,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 0.02,
24
+ "learning_rate": 4.967766474554855e-05,
25
+ "loss": 2.3321,
26
+ "step": 1500
27
+ },
28
+ {
29
+ "epoch": 0.03,
30
+ "learning_rate": 4.9570219660731405e-05,
31
+ "loss": 2.2234,
32
+ "step": 2000
33
+ },
34
+ {
35
+ "epoch": 0.03,
36
+ "learning_rate": 4.946277457591425e-05,
37
+ "loss": 2.156,
38
+ "step": 2500
39
+ },
40
+ {
41
+ "epoch": 0.04,
42
+ "learning_rate": 4.93553294910971e-05,
43
+ "loss": 2.1099,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.05,
48
+ "learning_rate": 4.924788440627995e-05,
49
+ "loss": 2.0528,
50
+ "step": 3500
51
+ },
52
+ {
53
+ "epoch": 0.05,
54
+ "learning_rate": 4.91404393214628e-05,
55
+ "loss": 2.0336,
56
+ "step": 4000
57
+ },
58
+ {
59
+ "epoch": 0.06,
60
+ "learning_rate": 4.903299423664565e-05,
61
+ "loss": 1.9938,
62
+ "step": 4500
63
+ },
64
+ {
65
+ "epoch": 0.06,
66
+ "learning_rate": 4.8925549151828506e-05,
67
+ "loss": 1.9726,
68
+ "step": 5000
69
+ },
70
+ {
71
+ "epoch": 0.07,
72
+ "learning_rate": 4.8818104067011355e-05,
73
+ "loss": 1.9286,
74
+ "step": 5500
75
+ },
76
+ {
77
+ "epoch": 0.08,
78
+ "learning_rate": 4.8710658982194204e-05,
79
+ "loss": 1.9026,
80
+ "step": 6000
81
+ },
82
+ {
83
+ "epoch": 0.08,
84
+ "learning_rate": 4.860321389737705e-05,
85
+ "loss": 1.9147,
86
+ "step": 6500
87
+ },
88
+ {
89
+ "epoch": 0.09,
90
+ "learning_rate": 4.84957688125599e-05,
91
+ "loss": 1.8774,
92
+ "step": 7000
93
+ },
94
+ {
95
+ "epoch": 0.1,
96
+ "learning_rate": 4.838832372774275e-05,
97
+ "loss": 1.8614,
98
+ "step": 7500
99
+ },
100
+ {
101
+ "epoch": 0.1,
102
+ "learning_rate": 4.82808786429256e-05,
103
+ "loss": 1.8656,
104
+ "step": 8000
105
+ },
106
+ {
107
+ "epoch": 0.11,
108
+ "learning_rate": 4.8173433558108456e-05,
109
+ "loss": 1.8337,
110
+ "step": 8500
111
+ },
112
+ {
113
+ "epoch": 0.12,
114
+ "learning_rate": 4.80659884732913e-05,
115
+ "loss": 1.8322,
116
+ "step": 9000
117
+ },
118
+ {
119
+ "epoch": 0.12,
120
+ "learning_rate": 4.7958543388474154e-05,
121
+ "loss": 1.7918,
122
+ "step": 9500
123
+ },
124
+ {
125
+ "epoch": 0.13,
126
+ "learning_rate": 4.7851098303657e-05,
127
+ "loss": 1.7813,
128
+ "step": 10000
129
+ },
130
+ {
131
+ "epoch": 0.14,
132
+ "learning_rate": 4.774365321883985e-05,
133
+ "loss": 1.7649,
134
+ "step": 10500
135
+ },
136
+ {
137
+ "epoch": 0.14,
138
+ "learning_rate": 4.76362081340227e-05,
139
+ "loss": 1.7451,
140
+ "step": 11000
141
+ },
142
+ {
143
+ "epoch": 0.15,
144
+ "learning_rate": 4.752876304920556e-05,
145
+ "loss": 1.7723,
146
+ "step": 11500
147
+ },
148
+ {
149
+ "epoch": 0.15,
150
+ "learning_rate": 4.74213179643884e-05,
151
+ "loss": 1.7512,
152
+ "step": 12000
153
+ },
154
+ {
155
+ "epoch": 0.16,
156
+ "learning_rate": 4.7313872879571255e-05,
157
+ "loss": 1.728,
158
+ "step": 12500
159
+ },
160
+ {
161
+ "epoch": 0.17,
162
+ "learning_rate": 4.7206427794754104e-05,
163
+ "loss": 1.728,
164
+ "step": 13000
165
+ },
166
+ {
167
+ "epoch": 0.17,
168
+ "learning_rate": 4.709898270993695e-05,
169
+ "loss": 1.7277,
170
+ "step": 13500
171
+ },
172
+ {
173
+ "epoch": 0.18,
174
+ "learning_rate": 4.69915376251198e-05,
175
+ "loss": 1.7009,
176
+ "step": 14000
177
+ },
178
+ {
179
+ "epoch": 0.19,
180
+ "learning_rate": 4.688409254030265e-05,
181
+ "loss": 1.6931,
182
+ "step": 14500
183
+ },
184
+ {
185
+ "epoch": 0.19,
186
+ "learning_rate": 4.677664745548551e-05,
187
+ "loss": 1.734,
188
+ "step": 15000
189
+ },
190
+ {
191
+ "epoch": 0.2,
192
+ "learning_rate": 4.666920237066835e-05,
193
+ "loss": 1.6718,
194
+ "step": 15500
195
+ },
196
+ {
197
+ "epoch": 0.21,
198
+ "learning_rate": 4.6561757285851205e-05,
199
+ "loss": 1.6895,
200
+ "step": 16000
201
+ },
202
+ {
203
+ "epoch": 0.21,
204
+ "learning_rate": 4.6454312201034054e-05,
205
+ "loss": 1.7088,
206
+ "step": 16500
207
+ },
208
+ {
209
+ "epoch": 0.22,
210
+ "learning_rate": 4.63468671162169e-05,
211
+ "loss": 1.6694,
212
+ "step": 17000
213
+ },
214
+ {
215
+ "epoch": 0.23,
216
+ "learning_rate": 4.623942203139975e-05,
217
+ "loss": 1.6648,
218
+ "step": 17500
219
+ },
220
+ {
221
+ "epoch": 0.23,
222
+ "learning_rate": 4.613197694658261e-05,
223
+ "loss": 1.6718,
224
+ "step": 18000
225
+ },
226
+ {
227
+ "epoch": 0.24,
228
+ "learning_rate": 4.602453186176545e-05,
229
+ "loss": 1.6606,
230
+ "step": 18500
231
+ },
232
+ {
233
+ "epoch": 0.24,
234
+ "learning_rate": 4.5917086776948306e-05,
235
+ "loss": 1.6807,
236
+ "step": 19000
237
+ },
238
+ {
239
+ "epoch": 0.25,
240
+ "learning_rate": 4.5809641692131155e-05,
241
+ "loss": 1.6352,
242
+ "step": 19500
243
+ },
244
+ {
245
+ "epoch": 0.26,
246
+ "learning_rate": 4.5702196607314004e-05,
247
+ "loss": 1.6454,
248
+ "step": 20000
249
+ },
250
+ {
251
+ "epoch": 0.26,
252
+ "learning_rate": 4.559475152249685e-05,
253
+ "loss": 1.6509,
254
+ "step": 20500
255
+ },
256
+ {
257
+ "epoch": 0.27,
258
+ "learning_rate": 4.548730643767971e-05,
259
+ "loss": 1.6083,
260
+ "step": 21000
261
+ },
262
+ {
263
+ "epoch": 0.28,
264
+ "learning_rate": 4.537986135286255e-05,
265
+ "loss": 1.6177,
266
+ "step": 21500
267
+ },
268
+ {
269
+ "epoch": 0.28,
270
+ "learning_rate": 4.527241626804541e-05,
271
+ "loss": 1.6097,
272
+ "step": 22000
273
+ },
274
+ {
275
+ "epoch": 0.29,
276
+ "learning_rate": 4.5164971183228256e-05,
277
+ "loss": 1.6025,
278
+ "step": 22500
279
+ },
280
+ {
281
+ "epoch": 0.3,
282
+ "learning_rate": 4.5057526098411105e-05,
283
+ "loss": 1.6162,
284
+ "step": 23000
285
+ },
286
+ {
287
+ "epoch": 0.3,
288
+ "learning_rate": 4.4950081013593954e-05,
289
+ "loss": 1.6092,
290
+ "step": 23500
291
+ },
292
+ {
293
+ "epoch": 0.31,
294
+ "learning_rate": 4.48426359287768e-05,
295
+ "loss": 1.592,
296
+ "step": 24000
297
+ },
298
+ {
299
+ "epoch": 0.32,
300
+ "learning_rate": 4.473519084395966e-05,
301
+ "loss": 1.5961,
302
+ "step": 24500
303
+ },
304
+ {
305
+ "epoch": 0.32,
306
+ "learning_rate": 4.46277457591425e-05,
307
+ "loss": 1.5711,
308
+ "step": 25000
309
+ },
310
+ {
311
+ "epoch": 0.33,
312
+ "learning_rate": 4.4520300674325357e-05,
313
+ "loss": 1.5789,
314
+ "step": 25500
315
+ },
316
+ {
317
+ "epoch": 0.34,
318
+ "learning_rate": 4.4412855589508206e-05,
319
+ "loss": 1.5774,
320
+ "step": 26000
321
+ },
322
+ {
323
+ "epoch": 0.34,
324
+ "learning_rate": 4.4305410504691055e-05,
325
+ "loss": 1.5854,
326
+ "step": 26500
327
+ },
328
+ {
329
+ "epoch": 0.35,
330
+ "learning_rate": 4.4197965419873904e-05,
331
+ "loss": 1.5462,
332
+ "step": 27000
333
+ },
334
+ {
335
+ "epoch": 0.35,
336
+ "learning_rate": 4.409052033505676e-05,
337
+ "loss": 1.5591,
338
+ "step": 27500
339
+ },
340
+ {
341
+ "epoch": 0.36,
342
+ "learning_rate": 4.39830752502396e-05,
343
+ "loss": 1.5775,
344
+ "step": 28000
345
+ },
346
+ {
347
+ "epoch": 0.37,
348
+ "learning_rate": 4.387563016542246e-05,
349
+ "loss": 1.542,
350
+ "step": 28500
351
+ },
352
+ {
353
+ "epoch": 0.37,
354
+ "learning_rate": 4.3768185080605306e-05,
355
+ "loss": 1.5408,
356
+ "step": 29000
357
+ },
358
+ {
359
+ "epoch": 0.38,
360
+ "learning_rate": 4.3660739995788155e-05,
361
+ "loss": 1.5484,
362
+ "step": 29500
363
+ },
364
+ {
365
+ "epoch": 0.39,
366
+ "learning_rate": 4.3553294910971004e-05,
367
+ "loss": 1.5625,
368
+ "step": 30000
369
+ },
370
+ {
371
+ "epoch": 0.39,
372
+ "learning_rate": 4.3445849826153853e-05,
373
+ "loss": 1.5276,
374
+ "step": 30500
375
+ },
376
+ {
377
+ "epoch": 0.4,
378
+ "learning_rate": 4.33384047413367e-05,
379
+ "loss": 1.5524,
380
+ "step": 31000
381
+ },
382
+ {
383
+ "epoch": 0.41,
384
+ "learning_rate": 4.323095965651955e-05,
385
+ "loss": 1.5458,
386
+ "step": 31500
387
+ },
388
+ {
389
+ "epoch": 0.41,
390
+ "learning_rate": 4.312351457170241e-05,
391
+ "loss": 1.541,
392
+ "step": 32000
393
+ },
394
+ {
395
+ "epoch": 0.42,
396
+ "learning_rate": 4.3016069486885256e-05,
397
+ "loss": 1.5313,
398
+ "step": 32500
399
+ },
400
+ {
401
+ "epoch": 0.43,
402
+ "learning_rate": 4.2908624402068105e-05,
403
+ "loss": 1.5507,
404
+ "step": 33000
405
+ },
406
+ {
407
+ "epoch": 0.43,
408
+ "learning_rate": 4.2801179317250954e-05,
409
+ "loss": 1.5189,
410
+ "step": 33500
411
+ },
412
+ {
413
+ "epoch": 0.44,
414
+ "learning_rate": 4.269373423243381e-05,
415
+ "loss": 1.5292,
416
+ "step": 34000
417
+ },
418
+ {
419
+ "epoch": 0.44,
420
+ "learning_rate": 4.258628914761665e-05,
421
+ "loss": 1.5109,
422
+ "step": 34500
423
+ },
424
+ {
425
+ "epoch": 0.45,
426
+ "learning_rate": 4.247884406279951e-05,
427
+ "loss": 1.4981,
428
+ "step": 35000
429
+ },
430
+ {
431
+ "epoch": 0.46,
432
+ "learning_rate": 4.237139897798236e-05,
433
+ "loss": 1.5229,
434
+ "step": 35500
435
+ },
436
+ {
437
+ "epoch": 0.46,
438
+ "learning_rate": 4.2263953893165206e-05,
439
+ "loss": 1.496,
440
+ "step": 36000
441
+ },
442
+ {
443
+ "epoch": 0.47,
444
+ "learning_rate": 4.2156508808348055e-05,
445
+ "loss": 1.4987,
446
+ "step": 36500
447
+ },
448
+ {
449
+ "epoch": 0.48,
450
+ "learning_rate": 4.2049063723530904e-05,
451
+ "loss": 1.5046,
452
+ "step": 37000
453
+ },
454
+ {
455
+ "epoch": 0.48,
456
+ "learning_rate": 4.194161863871375e-05,
457
+ "loss": 1.5233,
458
+ "step": 37500
459
+ },
460
+ {
461
+ "epoch": 0.49,
462
+ "learning_rate": 4.18341735538966e-05,
463
+ "loss": 1.4931,
464
+ "step": 38000
465
+ },
466
+ {
467
+ "epoch": 0.5,
468
+ "learning_rate": 4.172672846907946e-05,
469
+ "loss": 1.5,
470
+ "step": 38500
471
+ },
472
+ {
473
+ "epoch": 0.5,
474
+ "learning_rate": 4.16192833842623e-05,
475
+ "loss": 1.5132,
476
+ "step": 39000
477
+ },
478
+ {
479
+ "epoch": 0.51,
480
+ "learning_rate": 4.1511838299445156e-05,
481
+ "loss": 1.4723,
482
+ "step": 39500
483
+ },
484
+ {
485
+ "epoch": 0.52,
486
+ "learning_rate": 4.1404393214628005e-05,
487
+ "loss": 1.5036,
488
+ "step": 40000
489
+ },
490
+ {
491
+ "epoch": 0.52,
492
+ "learning_rate": 4.1296948129810854e-05,
493
+ "loss": 1.4907,
494
+ "step": 40500
495
+ },
496
+ {
497
+ "epoch": 0.53,
498
+ "learning_rate": 4.11895030449937e-05,
499
+ "loss": 1.4806,
500
+ "step": 41000
501
+ },
502
+ {
503
+ "epoch": 0.54,
504
+ "learning_rate": 4.108205796017656e-05,
505
+ "loss": 1.4775,
506
+ "step": 41500
507
+ },
508
+ {
509
+ "epoch": 0.54,
510
+ "learning_rate": 4.097461287535941e-05,
511
+ "loss": 1.466,
512
+ "step": 42000
513
+ },
514
+ {
515
+ "epoch": 0.55,
516
+ "learning_rate": 4.086716779054226e-05,
517
+ "loss": 1.4855,
518
+ "step": 42500
519
+ },
520
+ {
521
+ "epoch": 0.55,
522
+ "learning_rate": 4.0759722705725106e-05,
523
+ "loss": 1.4749,
524
+ "step": 43000
525
+ },
526
+ {
527
+ "epoch": 0.56,
528
+ "learning_rate": 4.0652277620907955e-05,
529
+ "loss": 1.4617,
530
+ "step": 43500
531
+ },
532
+ {
533
+ "epoch": 0.57,
534
+ "learning_rate": 4.0544832536090804e-05,
535
+ "loss": 1.4674,
536
+ "step": 44000
537
+ },
538
+ {
539
+ "epoch": 0.57,
540
+ "learning_rate": 4.043738745127365e-05,
541
+ "loss": 1.437,
542
+ "step": 44500
543
+ },
544
+ {
545
+ "epoch": 0.58,
546
+ "learning_rate": 4.032994236645651e-05,
547
+ "loss": 1.4632,
548
+ "step": 45000
549
+ },
550
+ {
551
+ "epoch": 0.59,
552
+ "learning_rate": 4.022249728163935e-05,
553
+ "loss": 1.4692,
554
+ "step": 45500
555
+ },
556
+ {
557
+ "epoch": 0.59,
558
+ "learning_rate": 4.011505219682221e-05,
559
+ "loss": 1.4655,
560
+ "step": 46000
561
+ },
562
+ {
563
+ "epoch": 0.6,
564
+ "learning_rate": 4.0007607112005056e-05,
565
+ "loss": 1.4469,
566
+ "step": 46500
567
+ },
568
+ {
569
+ "epoch": 0.61,
570
+ "learning_rate": 3.9900162027187905e-05,
571
+ "loss": 1.4606,
572
+ "step": 47000
573
+ },
574
+ {
575
+ "epoch": 0.61,
576
+ "learning_rate": 3.9792716942370754e-05,
577
+ "loss": 1.4746,
578
+ "step": 47500
579
+ },
580
+ {
581
+ "epoch": 0.62,
582
+ "learning_rate": 3.968527185755361e-05,
583
+ "loss": 1.4479,
584
+ "step": 48000
585
+ },
586
+ {
587
+ "epoch": 0.63,
588
+ "learning_rate": 3.957782677273645e-05,
589
+ "loss": 1.4187,
590
+ "step": 48500
591
+ },
592
+ {
593
+ "epoch": 0.63,
594
+ "learning_rate": 3.947038168791931e-05,
595
+ "loss": 1.4497,
596
+ "step": 49000
597
+ },
598
+ {
599
+ "epoch": 0.64,
600
+ "learning_rate": 3.936293660310216e-05,
601
+ "loss": 1.4174,
602
+ "step": 49500
603
+ },
604
+ {
605
+ "epoch": 0.64,
606
+ "learning_rate": 3.9255491518285006e-05,
607
+ "loss": 1.4391,
608
+ "step": 50000
609
+ },
610
+ {
611
+ "epoch": 0.65,
612
+ "learning_rate": 3.9148046433467855e-05,
613
+ "loss": 1.4355,
614
+ "step": 50500
615
+ },
616
+ {
617
+ "epoch": 0.66,
618
+ "learning_rate": 3.9040601348650704e-05,
619
+ "loss": 1.441,
620
+ "step": 51000
621
+ },
622
+ {
623
+ "epoch": 0.66,
624
+ "learning_rate": 3.893315626383356e-05,
625
+ "loss": 1.4265,
626
+ "step": 51500
627
+ },
628
+ {
629
+ "epoch": 0.67,
630
+ "learning_rate": 3.88257111790164e-05,
631
+ "loss": 1.4499,
632
+ "step": 52000
633
+ },
634
+ {
635
+ "epoch": 0.68,
636
+ "learning_rate": 3.871826609419926e-05,
637
+ "loss": 1.4564,
638
+ "step": 52500
639
+ },
640
+ {
641
+ "epoch": 0.68,
642
+ "learning_rate": 3.861082100938211e-05,
643
+ "loss": 1.4362,
644
+ "step": 53000
645
+ },
646
+ {
647
+ "epoch": 0.69,
648
+ "learning_rate": 3.8503375924564956e-05,
649
+ "loss": 1.4338,
650
+ "step": 53500
651
+ },
652
+ {
653
+ "epoch": 0.7,
654
+ "learning_rate": 3.8395930839747805e-05,
655
+ "loss": 1.4099,
656
+ "step": 54000
657
+ },
658
+ {
659
+ "epoch": 0.7,
660
+ "learning_rate": 3.828848575493066e-05,
661
+ "loss": 1.4213,
662
+ "step": 54500
663
+ },
664
+ {
665
+ "epoch": 0.71,
666
+ "learning_rate": 3.81810406701135e-05,
667
+ "loss": 1.4142,
668
+ "step": 55000
669
+ },
670
+ {
671
+ "epoch": 0.72,
672
+ "learning_rate": 3.807359558529636e-05,
673
+ "loss": 1.411,
674
+ "step": 55500
675
+ },
676
+ {
677
+ "epoch": 0.72,
678
+ "learning_rate": 3.796615050047921e-05,
679
+ "loss": 1.4071,
680
+ "step": 56000
681
+ },
682
+ {
683
+ "epoch": 0.73,
684
+ "learning_rate": 3.785870541566206e-05,
685
+ "loss": 1.4536,
686
+ "step": 56500
687
+ },
688
+ {
689
+ "epoch": 0.73,
690
+ "learning_rate": 3.7751260330844906e-05,
691
+ "loss": 1.4205,
692
+ "step": 57000
693
+ },
694
+ {
695
+ "epoch": 0.74,
696
+ "learning_rate": 3.764381524602776e-05,
697
+ "loss": 1.4275,
698
+ "step": 57500
699
+ },
700
+ {
701
+ "epoch": 0.75,
702
+ "learning_rate": 3.7536370161210604e-05,
703
+ "loss": 1.4142,
704
+ "step": 58000
705
+ },
706
+ {
707
+ "epoch": 0.75,
708
+ "learning_rate": 3.742892507639346e-05,
709
+ "loss": 1.4007,
710
+ "step": 58500
711
+ },
712
+ {
713
+ "epoch": 0.76,
714
+ "learning_rate": 3.732147999157631e-05,
715
+ "loss": 1.3993,
716
+ "step": 59000
717
+ },
718
+ {
719
+ "epoch": 0.77,
720
+ "learning_rate": 3.721403490675916e-05,
721
+ "loss": 1.3797,
722
+ "step": 59500
723
+ },
724
+ {
725
+ "epoch": 0.77,
726
+ "learning_rate": 3.7106589821942007e-05,
727
+ "loss": 1.4012,
728
+ "step": 60000
729
+ },
730
+ {
731
+ "epoch": 0.78,
732
+ "learning_rate": 3.6999144737124856e-05,
733
+ "loss": 1.4133,
734
+ "step": 60500
735
+ },
736
+ {
737
+ "epoch": 0.79,
738
+ "learning_rate": 3.689169965230771e-05,
739
+ "loss": 1.4276,
740
+ "step": 61000
741
+ },
742
+ {
743
+ "epoch": 0.79,
744
+ "learning_rate": 3.6784254567490554e-05,
745
+ "loss": 1.4062,
746
+ "step": 61500
747
+ },
748
+ {
749
+ "epoch": 0.8,
750
+ "learning_rate": 3.667680948267341e-05,
751
+ "loss": 1.3683,
752
+ "step": 62000
753
+ },
754
+ {
755
+ "epoch": 0.81,
756
+ "learning_rate": 3.656936439785626e-05,
757
+ "loss": 1.3942,
758
+ "step": 62500
759
+ },
760
+ {
761
+ "epoch": 0.81,
762
+ "learning_rate": 3.646191931303911e-05,
763
+ "loss": 1.3861,
764
+ "step": 63000
765
+ },
766
+ {
767
+ "epoch": 0.82,
768
+ "learning_rate": 3.6354474228221956e-05,
769
+ "loss": 1.4005,
770
+ "step": 63500
771
+ },
772
+ {
773
+ "epoch": 0.83,
774
+ "learning_rate": 3.624702914340481e-05,
775
+ "loss": 1.3732,
776
+ "step": 64000
777
+ },
778
+ {
779
+ "epoch": 0.83,
780
+ "learning_rate": 3.6139584058587654e-05,
781
+ "loss": 1.4009,
782
+ "step": 64500
783
+ },
784
+ {
785
+ "epoch": 0.84,
786
+ "learning_rate": 3.603213897377051e-05,
787
+ "loss": 1.3785,
788
+ "step": 65000
789
+ },
790
+ {
791
+ "epoch": 0.84,
792
+ "learning_rate": 3.592469388895336e-05,
793
+ "loss": 1.3883,
794
+ "step": 65500
795
+ },
796
+ {
797
+ "epoch": 0.85,
798
+ "learning_rate": 3.581724880413621e-05,
799
+ "loss": 1.351,
800
+ "step": 66000
801
+ },
802
+ {
803
+ "epoch": 0.86,
804
+ "learning_rate": 3.570980371931906e-05,
805
+ "loss": 1.3579,
806
+ "step": 66500
807
+ },
808
+ {
809
+ "epoch": 0.86,
810
+ "learning_rate": 3.5602358634501906e-05,
811
+ "loss": 1.4154,
812
+ "step": 67000
813
+ },
814
+ {
815
+ "epoch": 0.87,
816
+ "learning_rate": 3.5494913549684755e-05,
817
+ "loss": 1.3607,
818
+ "step": 67500
819
+ },
820
+ {
821
+ "epoch": 0.88,
822
+ "learning_rate": 3.5387468464867604e-05,
823
+ "loss": 1.3604,
824
+ "step": 68000
825
+ },
826
+ {
827
+ "epoch": 0.88,
828
+ "learning_rate": 3.528002338005046e-05,
829
+ "loss": 1.3624,
830
+ "step": 68500
831
+ },
832
+ {
833
+ "epoch": 0.89,
834
+ "learning_rate": 3.517257829523331e-05,
835
+ "loss": 1.3805,
836
+ "step": 69000
837
+ },
838
+ {
839
+ "epoch": 0.9,
840
+ "learning_rate": 3.506513321041616e-05,
841
+ "loss": 1.3659,
842
+ "step": 69500
843
+ },
844
+ {
845
+ "epoch": 0.9,
846
+ "learning_rate": 3.495768812559901e-05,
847
+ "loss": 1.3777,
848
+ "step": 70000
849
+ },
850
+ {
851
+ "epoch": 0.91,
852
+ "learning_rate": 3.485024304078186e-05,
853
+ "loss": 1.3516,
854
+ "step": 70500
855
+ },
856
+ {
857
+ "epoch": 0.92,
858
+ "learning_rate": 3.4742797955964705e-05,
859
+ "loss": 1.3718,
860
+ "step": 71000
861
+ },
862
+ {
863
+ "epoch": 0.92,
864
+ "learning_rate": 3.463535287114756e-05,
865
+ "loss": 1.37,
866
+ "step": 71500
867
+ },
868
+ {
869
+ "epoch": 0.93,
870
+ "learning_rate": 3.452790778633041e-05,
871
+ "loss": 1.3502,
872
+ "step": 72000
873
+ },
874
+ {
875
+ "epoch": 0.93,
876
+ "learning_rate": 3.442046270151326e-05,
877
+ "loss": 1.3591,
878
+ "step": 72500
879
+ },
880
+ {
881
+ "epoch": 0.94,
882
+ "learning_rate": 3.431301761669611e-05,
883
+ "loss": 1.3483,
884
+ "step": 73000
885
+ },
886
+ {
887
+ "epoch": 0.95,
888
+ "learning_rate": 3.420557253187896e-05,
889
+ "loss": 1.3578,
890
+ "step": 73500
891
+ },
892
+ {
893
+ "epoch": 0.95,
894
+ "learning_rate": 3.4098127447061806e-05,
895
+ "loss": 1.3465,
896
+ "step": 74000
897
+ },
898
+ {
899
+ "epoch": 0.96,
900
+ "learning_rate": 3.3990682362244655e-05,
901
+ "loss": 1.3694,
902
+ "step": 74500
903
+ },
904
+ {
905
+ "epoch": 0.97,
906
+ "learning_rate": 3.388323727742751e-05,
907
+ "loss": 1.3631,
908
+ "step": 75000
909
+ },
910
+ {
911
+ "epoch": 0.97,
912
+ "learning_rate": 3.377579219261035e-05,
913
+ "loss": 1.3675,
914
+ "step": 75500
915
+ },
916
+ {
917
+ "epoch": 0.98,
918
+ "learning_rate": 3.366834710779321e-05,
919
+ "loss": 1.3631,
920
+ "step": 76000
921
+ },
922
+ {
923
+ "epoch": 0.99,
924
+ "learning_rate": 3.356090202297606e-05,
925
+ "loss": 1.3432,
926
+ "step": 76500
927
+ },
928
+ {
929
+ "epoch": 0.99,
930
+ "learning_rate": 3.3453456938158914e-05,
931
+ "loss": 1.3441,
932
+ "step": 77000
933
+ },
934
+ {
935
+ "epoch": 1.0,
936
+ "learning_rate": 3.3346011853341756e-05,
937
+ "loss": 1.3543,
938
+ "step": 77500
939
+ },
940
+ {
941
+ "epoch": 1.01,
942
+ "learning_rate": 3.323856676852461e-05,
943
+ "loss": 1.2648,
944
+ "step": 78000
945
+ },
946
+ {
947
+ "epoch": 1.01,
948
+ "learning_rate": 3.313112168370746e-05,
949
+ "loss": 1.2416,
950
+ "step": 78500
951
+ },
952
+ {
953
+ "epoch": 1.02,
954
+ "learning_rate": 3.302367659889031e-05,
955
+ "loss": 1.2426,
956
+ "step": 79000
957
+ },
958
+ {
959
+ "epoch": 1.03,
960
+ "learning_rate": 3.291623151407316e-05,
961
+ "loss": 1.2268,
962
+ "step": 79500
963
+ },
964
+ {
965
+ "epoch": 1.03,
966
+ "learning_rate": 3.280878642925601e-05,
967
+ "loss": 1.2366,
968
+ "step": 80000
969
+ },
970
+ {
971
+ "epoch": 1.04,
972
+ "learning_rate": 3.270134134443886e-05,
973
+ "loss": 1.2198,
974
+ "step": 80500
975
+ },
976
+ {
977
+ "epoch": 1.04,
978
+ "learning_rate": 3.2593896259621706e-05,
979
+ "loss": 1.2172,
980
+ "step": 81000
981
+ },
982
+ {
983
+ "epoch": 1.05,
984
+ "learning_rate": 3.248645117480456e-05,
985
+ "loss": 1.226,
986
+ "step": 81500
987
+ },
988
+ {
989
+ "epoch": 1.06,
990
+ "learning_rate": 3.2379006089987404e-05,
991
+ "loss": 1.2192,
992
+ "step": 82000
993
+ },
994
+ {
995
+ "epoch": 1.06,
996
+ "learning_rate": 3.227156100517026e-05,
997
+ "loss": 1.2385,
998
+ "step": 82500
999
+ },
1000
+ {
1001
+ "epoch": 1.07,
1002
+ "learning_rate": 3.216411592035311e-05,
1003
+ "loss": 1.2127,
1004
+ "step": 83000
1005
+ },
1006
+ {
1007
+ "epoch": 1.08,
1008
+ "learning_rate": 3.205667083553596e-05,
1009
+ "loss": 1.234,
1010
+ "step": 83500
1011
+ },
1012
+ {
1013
+ "epoch": 1.08,
1014
+ "learning_rate": 3.194922575071881e-05,
1015
+ "loss": 1.2267,
1016
+ "step": 84000
1017
+ },
1018
+ {
1019
+ "epoch": 1.09,
1020
+ "learning_rate": 3.184178066590166e-05,
1021
+ "loss": 1.2302,
1022
+ "step": 84500
1023
+ },
1024
+ {
1025
+ "epoch": 1.1,
1026
+ "learning_rate": 3.1734335581084505e-05,
1027
+ "loss": 1.2387,
1028
+ "step": 85000
1029
+ },
1030
+ {
1031
+ "epoch": 1.1,
1032
+ "learning_rate": 3.162689049626736e-05,
1033
+ "loss": 1.243,
1034
+ "step": 85500
1035
+ },
1036
+ {
1037
+ "epoch": 1.11,
1038
+ "learning_rate": 3.151944541145021e-05,
1039
+ "loss": 1.2139,
1040
+ "step": 86000
1041
+ },
1042
+ {
1043
+ "epoch": 1.12,
1044
+ "learning_rate": 3.141200032663306e-05,
1045
+ "loss": 1.2217,
1046
+ "step": 86500
1047
+ },
1048
+ {
1049
+ "epoch": 1.12,
1050
+ "learning_rate": 3.130455524181591e-05,
1051
+ "loss": 1.2122,
1052
+ "step": 87000
1053
+ },
1054
+ {
1055
+ "epoch": 1.13,
1056
+ "learning_rate": 3.119711015699876e-05,
1057
+ "loss": 1.2253,
1058
+ "step": 87500
1059
+ },
1060
+ {
1061
+ "epoch": 1.13,
1062
+ "learning_rate": 3.108966507218161e-05,
1063
+ "loss": 1.2172,
1064
+ "step": 88000
1065
+ },
1066
+ {
1067
+ "epoch": 1.14,
1068
+ "learning_rate": 3.0982219987364455e-05,
1069
+ "loss": 1.2296,
1070
+ "step": 88500
1071
+ },
1072
+ {
1073
+ "epoch": 1.15,
1074
+ "learning_rate": 3.087477490254731e-05,
1075
+ "loss": 1.2216,
1076
+ "step": 89000
1077
+ },
1078
+ {
1079
+ "epoch": 1.15,
1080
+ "learning_rate": 3.076732981773016e-05,
1081
+ "loss": 1.2205,
1082
+ "step": 89500
1083
+ },
1084
+ {
1085
+ "epoch": 1.16,
1086
+ "learning_rate": 3.065988473291301e-05,
1087
+ "loss": 1.2127,
1088
+ "step": 90000
1089
+ },
1090
+ {
1091
+ "epoch": 1.17,
1092
+ "learning_rate": 3.055243964809586e-05,
1093
+ "loss": 1.2324,
1094
+ "step": 90500
1095
+ },
1096
+ {
1097
+ "epoch": 1.17,
1098
+ "learning_rate": 3.044499456327871e-05,
1099
+ "loss": 1.2124,
1100
+ "step": 91000
1101
+ },
1102
+ {
1103
+ "epoch": 1.18,
1104
+ "learning_rate": 3.033754947846156e-05,
1105
+ "loss": 1.2364,
1106
+ "step": 91500
1107
+ },
1108
+ {
1109
+ "epoch": 1.19,
1110
+ "learning_rate": 3.0230104393644408e-05,
1111
+ "loss": 1.2111,
1112
+ "step": 92000
1113
+ },
1114
+ {
1115
+ "epoch": 1.19,
1116
+ "learning_rate": 3.012265930882726e-05,
1117
+ "loss": 1.2082,
1118
+ "step": 92500
1119
+ },
1120
+ {
1121
+ "epoch": 1.2,
1122
+ "learning_rate": 3.0015214224010106e-05,
1123
+ "loss": 1.2292,
1124
+ "step": 93000
1125
+ },
1126
+ {
1127
+ "epoch": 1.21,
1128
+ "learning_rate": 2.990776913919296e-05,
1129
+ "loss": 1.2236,
1130
+ "step": 93500
1131
+ },
1132
+ {
1133
+ "epoch": 1.21,
1134
+ "learning_rate": 2.980032405437581e-05,
1135
+ "loss": 1.2178,
1136
+ "step": 94000
1137
+ },
1138
+ {
1139
+ "epoch": 1.22,
1140
+ "learning_rate": 2.9692878969558657e-05,
1141
+ "loss": 1.2165,
1142
+ "step": 94500
1143
+ },
1144
+ {
1145
+ "epoch": 1.22,
1146
+ "learning_rate": 2.958543388474151e-05,
1147
+ "loss": 1.21,
1148
+ "step": 95000
1149
+ },
1150
+ {
1151
+ "epoch": 1.23,
1152
+ "learning_rate": 2.947798879992436e-05,
1153
+ "loss": 1.2239,
1154
+ "step": 95500
1155
+ },
1156
+ {
1157
+ "epoch": 1.24,
1158
+ "learning_rate": 2.9370543715107214e-05,
1159
+ "loss": 1.2094,
1160
+ "step": 96000
1161
+ },
1162
+ {
1163
+ "epoch": 1.24,
1164
+ "learning_rate": 2.926309863029006e-05,
1165
+ "loss": 1.1993,
1166
+ "step": 96500
1167
+ },
1168
+ {
1169
+ "epoch": 1.25,
1170
+ "learning_rate": 2.9155653545472912e-05,
1171
+ "loss": 1.2143,
1172
+ "step": 97000
1173
+ },
1174
+ {
1175
+ "epoch": 1.26,
1176
+ "learning_rate": 2.9048208460655764e-05,
1177
+ "loss": 1.2092,
1178
+ "step": 97500
1179
+ },
1180
+ {
1181
+ "epoch": 1.26,
1182
+ "learning_rate": 2.894076337583861e-05,
1183
+ "loss": 1.2128,
1184
+ "step": 98000
1185
+ },
1186
+ {
1187
+ "epoch": 1.27,
1188
+ "learning_rate": 2.8833318291021462e-05,
1189
+ "loss": 1.2339,
1190
+ "step": 98500
1191
+ },
1192
+ {
1193
+ "epoch": 1.28,
1194
+ "learning_rate": 2.872587320620431e-05,
1195
+ "loss": 1.2002,
1196
+ "step": 99000
1197
+ },
1198
+ {
1199
+ "epoch": 1.28,
1200
+ "learning_rate": 2.8618428121387157e-05,
1201
+ "loss": 1.2068,
1202
+ "step": 99500
1203
+ },
1204
+ {
1205
+ "epoch": 1.29,
1206
+ "learning_rate": 2.851098303657001e-05,
1207
+ "loss": 1.2026,
1208
+ "step": 100000
1209
+ },
1210
+ {
1211
+ "epoch": 1.3,
1212
+ "learning_rate": 2.840353795175286e-05,
1213
+ "loss": 1.2186,
1214
+ "step": 100500
1215
+ },
1216
+ {
1217
+ "epoch": 1.3,
1218
+ "learning_rate": 2.8296092866935707e-05,
1219
+ "loss": 1.2251,
1220
+ "step": 101000
1221
+ },
1222
+ {
1223
+ "epoch": 1.31,
1224
+ "learning_rate": 2.818864778211856e-05,
1225
+ "loss": 1.2054,
1226
+ "step": 101500
1227
+ },
1228
+ {
1229
+ "epoch": 1.32,
1230
+ "learning_rate": 2.8081202697301412e-05,
1231
+ "loss": 1.2136,
1232
+ "step": 102000
1233
+ },
1234
+ {
1235
+ "epoch": 1.32,
1236
+ "learning_rate": 2.7973757612484258e-05,
1237
+ "loss": 1.2064,
1238
+ "step": 102500
1239
+ },
1240
+ {
1241
+ "epoch": 1.33,
1242
+ "learning_rate": 2.786631252766711e-05,
1243
+ "loss": 1.1948,
1244
+ "step": 103000
1245
+ },
1246
+ {
1247
+ "epoch": 1.33,
1248
+ "learning_rate": 2.7758867442849963e-05,
1249
+ "loss": 1.1918,
1250
+ "step": 103500
1251
+ },
1252
+ {
1253
+ "epoch": 1.34,
1254
+ "learning_rate": 2.7651422358032808e-05,
1255
+ "loss": 1.2137,
1256
+ "step": 104000
1257
+ },
1258
+ {
1259
+ "epoch": 1.35,
1260
+ "learning_rate": 2.754397727321566e-05,
1261
+ "loss": 1.1977,
1262
+ "step": 104500
1263
+ },
1264
+ {
1265
+ "epoch": 1.35,
1266
+ "learning_rate": 2.7436532188398513e-05,
1267
+ "loss": 1.2128,
1268
+ "step": 105000
1269
+ },
1270
+ {
1271
+ "epoch": 1.36,
1272
+ "learning_rate": 2.7329087103581362e-05,
1273
+ "loss": 1.1802,
1274
+ "step": 105500
1275
+ },
1276
+ {
1277
+ "epoch": 1.37,
1278
+ "learning_rate": 2.722164201876421e-05,
1279
+ "loss": 1.1854,
1280
+ "step": 106000
1281
+ },
1282
+ {
1283
+ "epoch": 1.37,
1284
+ "learning_rate": 2.711419693394706e-05,
1285
+ "loss": 1.1809,
1286
+ "step": 106500
1287
+ },
1288
+ {
1289
+ "epoch": 1.38,
1290
+ "learning_rate": 2.7006751849129912e-05,
1291
+ "loss": 1.1923,
1292
+ "step": 107000
1293
+ },
1294
+ {
1295
+ "epoch": 1.39,
1296
+ "learning_rate": 2.6899306764312758e-05,
1297
+ "loss": 1.1918,
1298
+ "step": 107500
1299
+ },
1300
+ {
1301
+ "epoch": 1.39,
1302
+ "learning_rate": 2.679186167949561e-05,
1303
+ "loss": 1.1987,
1304
+ "step": 108000
1305
+ },
1306
+ {
1307
+ "epoch": 1.4,
1308
+ "learning_rate": 2.6684416594678463e-05,
1309
+ "loss": 1.1786,
1310
+ "step": 108500
1311
+ },
1312
+ {
1313
+ "epoch": 1.41,
1314
+ "learning_rate": 2.657697150986131e-05,
1315
+ "loss": 1.179,
1316
+ "step": 109000
1317
+ },
1318
+ {
1319
+ "epoch": 1.41,
1320
+ "learning_rate": 2.646952642504416e-05,
1321
+ "loss": 1.194,
1322
+ "step": 109500
1323
+ },
1324
+ {
1325
+ "epoch": 1.42,
1326
+ "learning_rate": 2.6362081340227013e-05,
1327
+ "loss": 1.1954,
1328
+ "step": 110000
1329
+ },
1330
+ {
1331
+ "epoch": 1.42,
1332
+ "learning_rate": 2.625463625540986e-05,
1333
+ "loss": 1.1996,
1334
+ "step": 110500
1335
+ },
1336
+ {
1337
+ "epoch": 1.43,
1338
+ "learning_rate": 2.614719117059271e-05,
1339
+ "loss": 1.2026,
1340
+ "step": 111000
1341
+ },
1342
+ {
1343
+ "epoch": 1.44,
1344
+ "learning_rate": 2.6039746085775564e-05,
1345
+ "loss": 1.1892,
1346
+ "step": 111500
1347
+ },
1348
+ {
1349
+ "epoch": 1.44,
1350
+ "learning_rate": 2.593230100095841e-05,
1351
+ "loss": 1.1968,
1352
+ "step": 112000
1353
+ },
1354
+ {
1355
+ "epoch": 1.45,
1356
+ "learning_rate": 2.5824855916141262e-05,
1357
+ "loss": 1.1719,
1358
+ "step": 112500
1359
+ },
1360
+ {
1361
+ "epoch": 1.46,
1362
+ "learning_rate": 2.571741083132411e-05,
1363
+ "loss": 1.1855,
1364
+ "step": 113000
1365
+ },
1366
+ {
1367
+ "epoch": 1.46,
1368
+ "learning_rate": 2.5609965746506963e-05,
1369
+ "loss": 1.1762,
1370
+ "step": 113500
1371
+ },
1372
+ {
1373
+ "epoch": 1.47,
1374
+ "learning_rate": 2.550252066168981e-05,
1375
+ "loss": 1.187,
1376
+ "step": 114000
1377
+ },
1378
+ {
1379
+ "epoch": 1.48,
1380
+ "learning_rate": 2.539507557687266e-05,
1381
+ "loss": 1.1892,
1382
+ "step": 114500
1383
+ },
1384
+ {
1385
+ "epoch": 1.48,
1386
+ "learning_rate": 2.5287630492055514e-05,
1387
+ "loss": 1.2085,
1388
+ "step": 115000
1389
+ },
1390
+ {
1391
+ "epoch": 1.49,
1392
+ "learning_rate": 2.518018540723836e-05,
1393
+ "loss": 1.181,
1394
+ "step": 115500
1395
+ },
1396
+ {
1397
+ "epoch": 1.5,
1398
+ "learning_rate": 2.5072740322421212e-05,
1399
+ "loss": 1.1654,
1400
+ "step": 116000
1401
+ },
1402
+ {
1403
+ "epoch": 1.5,
1404
+ "learning_rate": 2.496529523760406e-05,
1405
+ "loss": 1.1677,
1406
+ "step": 116500
1407
+ },
1408
+ {
1409
+ "epoch": 1.51,
1410
+ "learning_rate": 2.4857850152786913e-05,
1411
+ "loss": 1.2037,
1412
+ "step": 117000
1413
+ },
1414
+ {
1415
+ "epoch": 1.51,
1416
+ "learning_rate": 2.4750405067969762e-05,
1417
+ "loss": 1.1787,
1418
+ "step": 117500
1419
+ },
1420
+ {
1421
+ "epoch": 1.52,
1422
+ "learning_rate": 2.464295998315261e-05,
1423
+ "loss": 1.179,
1424
+ "step": 118000
1425
+ },
1426
+ {
1427
+ "epoch": 1.53,
1428
+ "learning_rate": 2.4535514898335464e-05,
1429
+ "loss": 1.2019,
1430
+ "step": 118500
1431
+ },
1432
+ {
1433
+ "epoch": 1.53,
1434
+ "learning_rate": 2.4428069813518313e-05,
1435
+ "loss": 1.1787,
1436
+ "step": 119000
1437
+ },
1438
+ {
1439
+ "epoch": 1.54,
1440
+ "learning_rate": 2.432062472870116e-05,
1441
+ "loss": 1.1677,
1442
+ "step": 119500
1443
+ },
1444
+ {
1445
+ "epoch": 1.55,
1446
+ "learning_rate": 2.4213179643884014e-05,
1447
+ "loss": 1.1919,
1448
+ "step": 120000
1449
+ },
1450
+ {
1451
+ "epoch": 1.55,
1452
+ "learning_rate": 2.4105734559066863e-05,
1453
+ "loss": 1.1859,
1454
+ "step": 120500
1455
+ },
1456
+ {
1457
+ "epoch": 1.56,
1458
+ "learning_rate": 2.3998289474249712e-05,
1459
+ "loss": 1.1597,
1460
+ "step": 121000
1461
+ },
1462
+ {
1463
+ "epoch": 1.57,
1464
+ "learning_rate": 2.389084438943256e-05,
1465
+ "loss": 1.1806,
1466
+ "step": 121500
1467
+ },
1468
+ {
1469
+ "epoch": 1.57,
1470
+ "learning_rate": 2.378339930461541e-05,
1471
+ "loss": 1.1694,
1472
+ "step": 122000
1473
+ },
1474
+ {
1475
+ "epoch": 1.58,
1476
+ "learning_rate": 2.3675954219798263e-05,
1477
+ "loss": 1.1896,
1478
+ "step": 122500
1479
+ },
1480
+ {
1481
+ "epoch": 1.59,
1482
+ "learning_rate": 2.356850913498111e-05,
1483
+ "loss": 1.1897,
1484
+ "step": 123000
1485
+ },
1486
+ {
1487
+ "epoch": 1.59,
1488
+ "learning_rate": 2.346106405016396e-05,
1489
+ "loss": 1.1796,
1490
+ "step": 123500
1491
+ },
1492
+ {
1493
+ "epoch": 1.6,
1494
+ "learning_rate": 2.3353618965346813e-05,
1495
+ "loss": 1.1745,
1496
+ "step": 124000
1497
+ },
1498
+ {
1499
+ "epoch": 1.61,
1500
+ "learning_rate": 2.3246173880529662e-05,
1501
+ "loss": 1.1828,
1502
+ "step": 124500
1503
+ },
1504
+ {
1505
+ "epoch": 1.61,
1506
+ "learning_rate": 2.3138728795712514e-05,
1507
+ "loss": 1.1875,
1508
+ "step": 125000
1509
+ },
1510
+ {
1511
+ "epoch": 1.62,
1512
+ "learning_rate": 2.3031283710895363e-05,
1513
+ "loss": 1.1716,
1514
+ "step": 125500
1515
+ },
1516
+ {
1517
+ "epoch": 1.62,
1518
+ "learning_rate": 2.2923838626078212e-05,
1519
+ "loss": 1.1768,
1520
+ "step": 126000
1521
+ },
1522
+ {
1523
+ "epoch": 1.63,
1524
+ "learning_rate": 2.2816393541261065e-05,
1525
+ "loss": 1.1821,
1526
+ "step": 126500
1527
+ },
1528
+ {
1529
+ "epoch": 1.64,
1530
+ "learning_rate": 2.2708948456443914e-05,
1531
+ "loss": 1.1672,
1532
+ "step": 127000
1533
+ },
1534
+ {
1535
+ "epoch": 1.64,
1536
+ "learning_rate": 2.2601503371626763e-05,
1537
+ "loss": 1.173,
1538
+ "step": 127500
1539
+ },
1540
+ {
1541
+ "epoch": 1.65,
1542
+ "learning_rate": 2.2494058286809612e-05,
1543
+ "loss": 1.1751,
1544
+ "step": 128000
1545
+ },
1546
+ {
1547
+ "epoch": 1.66,
1548
+ "learning_rate": 2.238661320199246e-05,
1549
+ "loss": 1.1967,
1550
+ "step": 128500
1551
+ },
1552
+ {
1553
+ "epoch": 1.66,
1554
+ "learning_rate": 2.227916811717531e-05,
1555
+ "loss": 1.1695,
1556
+ "step": 129000
1557
+ },
1558
+ {
1559
+ "epoch": 1.67,
1560
+ "learning_rate": 2.2171723032358162e-05,
1561
+ "loss": 1.179,
1562
+ "step": 129500
1563
+ },
1564
+ {
1565
+ "epoch": 1.68,
1566
+ "learning_rate": 2.206427794754101e-05,
1567
+ "loss": 1.1793,
1568
+ "step": 130000
1569
+ },
1570
+ {
1571
+ "epoch": 1.68,
1572
+ "learning_rate": 2.1956832862723864e-05,
1573
+ "loss": 1.1579,
1574
+ "step": 130500
1575
+ },
1576
+ {
1577
+ "epoch": 1.69,
1578
+ "learning_rate": 2.1849387777906713e-05,
1579
+ "loss": 1.1466,
1580
+ "step": 131000
1581
+ },
1582
+ {
1583
+ "epoch": 1.7,
1584
+ "learning_rate": 2.1741942693089562e-05,
1585
+ "loss": 1.208,
1586
+ "step": 131500
1587
+ },
1588
+ {
1589
+ "epoch": 1.7,
1590
+ "learning_rate": 2.1634497608272414e-05,
1591
+ "loss": 1.1592,
1592
+ "step": 132000
1593
+ },
1594
+ {
1595
+ "epoch": 1.71,
1596
+ "learning_rate": 2.1527052523455263e-05,
1597
+ "loss": 1.1558,
1598
+ "step": 132500
1599
+ },
1600
+ {
1601
+ "epoch": 1.71,
1602
+ "learning_rate": 2.1419607438638112e-05,
1603
+ "loss": 1.158,
1604
+ "step": 133000
1605
+ },
1606
+ {
1607
+ "epoch": 1.72,
1608
+ "learning_rate": 2.1312162353820965e-05,
1609
+ "loss": 1.1566,
1610
+ "step": 133500
1611
+ },
1612
+ {
1613
+ "epoch": 1.73,
1614
+ "learning_rate": 2.1204717269003814e-05,
1615
+ "loss": 1.1722,
1616
+ "step": 134000
1617
+ },
1618
+ {
1619
+ "epoch": 1.73,
1620
+ "learning_rate": 2.1097272184186666e-05,
1621
+ "loss": 1.1612,
1622
+ "step": 134500
1623
+ },
1624
+ {
1625
+ "epoch": 1.74,
1626
+ "learning_rate": 2.0989827099369515e-05,
1627
+ "loss": 1.1706,
1628
+ "step": 135000
1629
+ },
1630
+ {
1631
+ "epoch": 1.75,
1632
+ "learning_rate": 2.0882382014552364e-05,
1633
+ "loss": 1.1746,
1634
+ "step": 135500
1635
+ },
1636
+ {
1637
+ "epoch": 1.75,
1638
+ "learning_rate": 2.0774936929735213e-05,
1639
+ "loss": 1.1568,
1640
+ "step": 136000
1641
+ },
1642
+ {
1643
+ "epoch": 1.76,
1644
+ "learning_rate": 2.0667491844918062e-05,
1645
+ "loss": 1.1564,
1646
+ "step": 136500
1647
+ },
1648
+ {
1649
+ "epoch": 1.77,
1650
+ "learning_rate": 2.056004676010091e-05,
1651
+ "loss": 1.1566,
1652
+ "step": 137000
1653
+ },
1654
+ {
1655
+ "epoch": 1.77,
1656
+ "learning_rate": 2.0452601675283764e-05,
1657
+ "loss": 1.1684,
1658
+ "step": 137500
1659
+ },
1660
+ {
1661
+ "epoch": 1.78,
1662
+ "learning_rate": 2.0345156590466613e-05,
1663
+ "loss": 1.166,
1664
+ "step": 138000
1665
+ },
1666
+ {
1667
+ "epoch": 1.79,
1668
+ "learning_rate": 2.0237711505649465e-05,
1669
+ "loss": 1.1439,
1670
+ "step": 138500
1671
+ },
1672
+ {
1673
+ "epoch": 1.79,
1674
+ "learning_rate": 2.0130266420832314e-05,
1675
+ "loss": 1.1535,
1676
+ "step": 139000
1677
+ },
1678
+ {
1679
+ "epoch": 1.8,
1680
+ "learning_rate": 2.0022821336015163e-05,
1681
+ "loss": 1.1571,
1682
+ "step": 139500
1683
+ },
1684
+ {
1685
+ "epoch": 1.81,
1686
+ "learning_rate": 1.9915376251198015e-05,
1687
+ "loss": 1.1573,
1688
+ "step": 140000
1689
+ },
1690
+ {
1691
+ "epoch": 1.81,
1692
+ "learning_rate": 1.9807931166380864e-05,
1693
+ "loss": 1.1682,
1694
+ "step": 140500
1695
+ },
1696
+ {
1697
+ "epoch": 1.82,
1698
+ "learning_rate": 1.9700486081563713e-05,
1699
+ "loss": 1.1506,
1700
+ "step": 141000
1701
+ },
1702
+ {
1703
+ "epoch": 1.82,
1704
+ "learning_rate": 1.9593040996746566e-05,
1705
+ "loss": 1.1513,
1706
+ "step": 141500
1707
+ },
1708
+ {
1709
+ "epoch": 1.83,
1710
+ "learning_rate": 1.9485595911929415e-05,
1711
+ "loss": 1.17,
1712
+ "step": 142000
1713
+ },
1714
+ {
1715
+ "epoch": 1.84,
1716
+ "learning_rate": 1.9378150827112264e-05,
1717
+ "loss": 1.1538,
1718
+ "step": 142500
1719
+ },
1720
+ {
1721
+ "epoch": 1.84,
1722
+ "learning_rate": 1.9270705742295113e-05,
1723
+ "loss": 1.1461,
1724
+ "step": 143000
1725
+ },
1726
+ {
1727
+ "epoch": 1.85,
1728
+ "learning_rate": 1.9163260657477962e-05,
1729
+ "loss": 1.1353,
1730
+ "step": 143500
1731
+ },
1732
+ {
1733
+ "epoch": 1.86,
1734
+ "learning_rate": 1.9055815572660814e-05,
1735
+ "loss": 1.1523,
1736
+ "step": 144000
1737
+ },
1738
+ {
1739
+ "epoch": 1.86,
1740
+ "learning_rate": 1.8948370487843663e-05,
1741
+ "loss": 1.1547,
1742
+ "step": 144500
1743
+ },
1744
+ {
1745
+ "epoch": 1.87,
1746
+ "learning_rate": 1.8840925403026512e-05,
1747
+ "loss": 1.1092,
1748
+ "step": 145000
1749
+ },
1750
+ {
1751
+ "epoch": 1.88,
1752
+ "learning_rate": 1.8733480318209365e-05,
1753
+ "loss": 1.1662,
1754
+ "step": 145500
1755
+ },
1756
+ {
1757
+ "epoch": 1.88,
1758
+ "learning_rate": 1.8626035233392214e-05,
1759
+ "loss": 1.1539,
1760
+ "step": 146000
1761
+ },
1762
+ {
1763
+ "epoch": 1.89,
1764
+ "learning_rate": 1.8518590148575063e-05,
1765
+ "loss": 1.1571,
1766
+ "step": 146500
1767
+ },
1768
+ {
1769
+ "epoch": 1.9,
1770
+ "learning_rate": 1.8411145063757915e-05,
1771
+ "loss": 1.1228,
1772
+ "step": 147000
1773
+ },
1774
+ {
1775
+ "epoch": 1.9,
1776
+ "learning_rate": 1.8303699978940764e-05,
1777
+ "loss": 1.1421,
1778
+ "step": 147500
1779
+ },
1780
+ {
1781
+ "epoch": 1.91,
1782
+ "learning_rate": 1.8196254894123617e-05,
1783
+ "loss": 1.1603,
1784
+ "step": 148000
1785
+ },
1786
+ {
1787
+ "epoch": 1.91,
1788
+ "learning_rate": 1.8088809809306466e-05,
1789
+ "loss": 1.1051,
1790
+ "step": 148500
1791
+ },
1792
+ {
1793
+ "epoch": 1.92,
1794
+ "learning_rate": 1.7981364724489315e-05,
1795
+ "loss": 1.1423,
1796
+ "step": 149000
1797
+ },
1798
+ {
1799
+ "epoch": 1.93,
1800
+ "learning_rate": 1.7873919639672164e-05,
1801
+ "loss": 1.1548,
1802
+ "step": 149500
1803
+ },
1804
+ {
1805
+ "epoch": 1.93,
1806
+ "learning_rate": 1.7766474554855013e-05,
1807
+ "loss": 1.1424,
1808
+ "step": 150000
1809
+ },
1810
+ {
1811
+ "epoch": 1.94,
1812
+ "learning_rate": 1.7659029470037862e-05,
1813
+ "loss": 1.1435,
1814
+ "step": 150500
1815
+ },
1816
+ {
1817
+ "epoch": 1.95,
1818
+ "learning_rate": 1.7551584385220714e-05,
1819
+ "loss": 1.1592,
1820
+ "step": 151000
1821
+ },
1822
+ {
1823
+ "epoch": 1.95,
1824
+ "learning_rate": 1.7444139300403563e-05,
1825
+ "loss": 1.1364,
1826
+ "step": 151500
1827
+ },
1828
+ {
1829
+ "epoch": 1.96,
1830
+ "learning_rate": 1.7336694215586416e-05,
1831
+ "loss": 1.1412,
1832
+ "step": 152000
1833
+ },
1834
+ {
1835
+ "epoch": 1.97,
1836
+ "learning_rate": 1.7229249130769265e-05,
1837
+ "loss": 1.1428,
1838
+ "step": 152500
1839
+ },
1840
+ {
1841
+ "epoch": 1.97,
1842
+ "learning_rate": 1.7121804045952114e-05,
1843
+ "loss": 1.131,
1844
+ "step": 153000
1845
+ },
1846
+ {
1847
+ "epoch": 1.98,
1848
+ "learning_rate": 1.7014358961134966e-05,
1849
+ "loss": 1.1438,
1850
+ "step": 153500
1851
+ },
1852
+ {
1853
+ "epoch": 1.99,
1854
+ "learning_rate": 1.6906913876317815e-05,
1855
+ "loss": 1.117,
1856
+ "step": 154000
1857
+ },
1858
+ {
1859
+ "epoch": 1.99,
1860
+ "learning_rate": 1.6799468791500664e-05,
1861
+ "loss": 1.1552,
1862
+ "step": 154500
1863
+ },
1864
+ {
1865
+ "epoch": 2.0,
1866
+ "learning_rate": 1.6692023706683516e-05,
1867
+ "loss": 1.1362,
1868
+ "step": 155000
1869
+ },
1870
+ {
1871
+ "epoch": 2.0,
1872
+ "learning_rate": 1.6584578621866365e-05,
1873
+ "loss": 1.0437,
1874
+ "step": 155500
1875
+ },
1876
+ {
1877
+ "epoch": 2.01,
1878
+ "learning_rate": 1.6477133537049214e-05,
1879
+ "loss": 0.9948,
1880
+ "step": 156000
1881
+ },
1882
+ {
1883
+ "epoch": 2.02,
1884
+ "learning_rate": 1.6369688452232067e-05,
1885
+ "loss": 1.0094,
1886
+ "step": 156500
1887
+ },
1888
+ {
1889
+ "epoch": 2.02,
1890
+ "learning_rate": 1.6262243367414916e-05,
1891
+ "loss": 1.0137,
1892
+ "step": 157000
1893
+ },
1894
+ {
1895
+ "epoch": 2.03,
1896
+ "learning_rate": 1.6154798282597765e-05,
1897
+ "loss": 1.0197,
1898
+ "step": 157500
1899
+ },
1900
+ {
1901
+ "epoch": 2.04,
1902
+ "learning_rate": 1.6047353197780614e-05,
1903
+ "loss": 1.0113,
1904
+ "step": 158000
1905
+ },
1906
+ {
1907
+ "epoch": 2.04,
1908
+ "learning_rate": 1.5939908112963463e-05,
1909
+ "loss": 1.0123,
1910
+ "step": 158500
1911
+ },
1912
+ {
1913
+ "epoch": 2.05,
1914
+ "learning_rate": 1.5832463028146315e-05,
1915
+ "loss": 1.018,
1916
+ "step": 159000
1917
+ },
1918
+ {
1919
+ "epoch": 2.06,
1920
+ "learning_rate": 1.5725017943329164e-05,
1921
+ "loss": 1.0256,
1922
+ "step": 159500
1923
+ },
1924
+ {
1925
+ "epoch": 2.06,
1926
+ "learning_rate": 1.5617572858512013e-05,
1927
+ "loss": 0.9943,
1928
+ "step": 160000
1929
+ },
1930
+ {
1931
+ "epoch": 2.07,
1932
+ "learning_rate": 1.5510127773694866e-05,
1933
+ "loss": 1.0117,
1934
+ "step": 160500
1935
+ },
1936
+ {
1937
+ "epoch": 2.08,
1938
+ "learning_rate": 1.5402682688877715e-05,
1939
+ "loss": 1.0127,
1940
+ "step": 161000
1941
+ },
1942
+ {
1943
+ "epoch": 2.08,
1944
+ "learning_rate": 1.5295237604060567e-05,
1945
+ "loss": 1.012,
1946
+ "step": 161500
1947
+ },
1948
+ {
1949
+ "epoch": 2.09,
1950
+ "learning_rate": 1.5187792519243416e-05,
1951
+ "loss": 0.9863,
1952
+ "step": 162000
1953
+ },
1954
+ {
1955
+ "epoch": 2.1,
1956
+ "learning_rate": 1.5080347434426265e-05,
1957
+ "loss": 1.0253,
1958
+ "step": 162500
1959
+ },
1960
+ {
1961
+ "epoch": 2.1,
1962
+ "learning_rate": 1.4972902349609116e-05,
1963
+ "loss": 1.0143,
1964
+ "step": 163000
1965
+ },
1966
+ {
1967
+ "epoch": 2.11,
1968
+ "learning_rate": 1.4865457264791965e-05,
1969
+ "loss": 0.9931,
1970
+ "step": 163500
1971
+ },
1972
+ {
1973
+ "epoch": 2.11,
1974
+ "learning_rate": 1.4758012179974814e-05,
1975
+ "loss": 1.0074,
1976
+ "step": 164000
1977
+ },
1978
+ {
1979
+ "epoch": 2.12,
1980
+ "learning_rate": 1.4650567095157666e-05,
1981
+ "loss": 0.9954,
1982
+ "step": 164500
1983
+ },
1984
+ {
1985
+ "epoch": 2.13,
1986
+ "learning_rate": 1.4543122010340515e-05,
1987
+ "loss": 1.0081,
1988
+ "step": 165000
1989
+ },
1990
+ {
1991
+ "epoch": 2.13,
1992
+ "learning_rate": 1.4435676925523368e-05,
1993
+ "loss": 1.0072,
1994
+ "step": 165500
1995
+ },
1996
+ {
1997
+ "epoch": 2.14,
1998
+ "learning_rate": 1.4328231840706217e-05,
1999
+ "loss": 1.0019,
2000
+ "step": 166000
2001
+ },
2002
+ {
2003
+ "epoch": 2.15,
2004
+ "learning_rate": 1.4220786755889066e-05,
2005
+ "loss": 1.0164,
2006
+ "step": 166500
2007
+ },
2008
+ {
2009
+ "epoch": 2.15,
2010
+ "learning_rate": 1.4113341671071917e-05,
2011
+ "loss": 1.0449,
2012
+ "step": 167000
2013
+ },
2014
+ {
2015
+ "epoch": 2.16,
2016
+ "learning_rate": 1.4005896586254766e-05,
2017
+ "loss": 0.9966,
2018
+ "step": 167500
2019
+ },
2020
+ {
2021
+ "epoch": 2.17,
2022
+ "learning_rate": 1.3898451501437615e-05,
2023
+ "loss": 1.0105,
2024
+ "step": 168000
2025
+ },
2026
+ {
2027
+ "epoch": 2.17,
2028
+ "learning_rate": 1.3791006416620467e-05,
2029
+ "loss": 1.0485,
2030
+ "step": 168500
2031
+ },
2032
+ {
2033
+ "epoch": 2.18,
2034
+ "learning_rate": 1.3683561331803316e-05,
2035
+ "loss": 1.0047,
2036
+ "step": 169000
2037
+ },
2038
+ {
2039
+ "epoch": 2.19,
2040
+ "learning_rate": 1.3576116246986165e-05,
2041
+ "loss": 1.0103,
2042
+ "step": 169500
2043
+ },
2044
+ {
2045
+ "epoch": 2.19,
2046
+ "learning_rate": 1.3468671162169016e-05,
2047
+ "loss": 1.0032,
2048
+ "step": 170000
2049
+ },
2050
+ {
2051
+ "epoch": 2.2,
2052
+ "learning_rate": 1.3361226077351865e-05,
2053
+ "loss": 1.0013,
2054
+ "step": 170500
2055
+ },
2056
+ {
2057
+ "epoch": 2.2,
2058
+ "learning_rate": 1.3253780992534717e-05,
2059
+ "loss": 1.0028,
2060
+ "step": 171000
2061
+ },
2062
+ {
2063
+ "epoch": 2.21,
2064
+ "learning_rate": 1.3146335907717566e-05,
2065
+ "loss": 1.0134,
2066
+ "step": 171500
2067
+ },
2068
+ {
2069
+ "epoch": 2.22,
2070
+ "learning_rate": 1.3038890822900415e-05,
2071
+ "loss": 1.036,
2072
+ "step": 172000
2073
+ },
2074
+ {
2075
+ "epoch": 2.22,
2076
+ "learning_rate": 1.2931445738083268e-05,
2077
+ "loss": 1.0139,
2078
+ "step": 172500
2079
+ },
2080
+ {
2081
+ "epoch": 2.23,
2082
+ "learning_rate": 1.2824000653266117e-05,
2083
+ "loss": 0.9908,
2084
+ "step": 173000
2085
+ },
2086
+ {
2087
+ "epoch": 2.24,
2088
+ "learning_rate": 1.2716555568448966e-05,
2089
+ "loss": 0.9896,
2090
+ "step": 173500
2091
+ },
2092
+ {
2093
+ "epoch": 2.24,
2094
+ "learning_rate": 1.2609110483631816e-05,
2095
+ "loss": 0.9982,
2096
+ "step": 174000
2097
+ },
2098
+ {
2099
+ "epoch": 2.25,
2100
+ "learning_rate": 1.2501665398814665e-05,
2101
+ "loss": 1.0115,
2102
+ "step": 174500
2103
+ },
2104
+ {
2105
+ "epoch": 2.26,
2106
+ "learning_rate": 1.2394220313997516e-05,
2107
+ "loss": 0.9939,
2108
+ "step": 175000
2109
+ },
2110
+ {
2111
+ "epoch": 2.26,
2112
+ "learning_rate": 1.2286775229180367e-05,
2113
+ "loss": 0.9957,
2114
+ "step": 175500
2115
+ },
2116
+ {
2117
+ "epoch": 2.27,
2118
+ "learning_rate": 1.2179330144363218e-05,
2119
+ "loss": 1.0009,
2120
+ "step": 176000
2121
+ },
2122
+ {
2123
+ "epoch": 2.28,
2124
+ "learning_rate": 1.2071885059546067e-05,
2125
+ "loss": 1.0102,
2126
+ "step": 176500
2127
+ },
2128
+ {
2129
+ "epoch": 2.28,
2130
+ "learning_rate": 1.1964439974728917e-05,
2131
+ "loss": 1.0058,
2132
+ "step": 177000
2133
+ },
2134
+ {
2135
+ "epoch": 2.29,
2136
+ "learning_rate": 1.1856994889911766e-05,
2137
+ "loss": 0.9999,
2138
+ "step": 177500
2139
+ },
2140
+ {
2141
+ "epoch": 2.3,
2142
+ "learning_rate": 1.1749549805094617e-05,
2143
+ "loss": 1.0141,
2144
+ "step": 178000
2145
+ },
2146
+ {
2147
+ "epoch": 2.3,
2148
+ "learning_rate": 1.1642104720277466e-05,
2149
+ "loss": 1.0004,
2150
+ "step": 178500
2151
+ },
2152
+ {
2153
+ "epoch": 2.31,
2154
+ "learning_rate": 1.1534659635460317e-05,
2155
+ "loss": 0.9727,
2156
+ "step": 179000
2157
+ },
2158
+ {
2159
+ "epoch": 2.31,
2160
+ "learning_rate": 1.1427214550643167e-05,
2161
+ "loss": 0.9989,
2162
+ "step": 179500
2163
+ },
2164
+ {
2165
+ "epoch": 2.32,
2166
+ "learning_rate": 1.1319769465826018e-05,
2167
+ "loss": 0.9968,
2168
+ "step": 180000
2169
+ },
2170
+ {
2171
+ "epoch": 2.33,
2172
+ "learning_rate": 1.1212324381008867e-05,
2173
+ "loss": 1.0138,
2174
+ "step": 180500
2175
+ },
2176
+ {
2177
+ "epoch": 2.33,
2178
+ "learning_rate": 1.1104879296191716e-05,
2179
+ "loss": 1.0079,
2180
+ "step": 181000
2181
+ },
2182
+ {
2183
+ "epoch": 2.34,
2184
+ "learning_rate": 1.0997434211374567e-05,
2185
+ "loss": 0.9866,
2186
+ "step": 181500
2187
+ },
2188
+ {
2189
+ "epoch": 2.35,
2190
+ "learning_rate": 1.0889989126557416e-05,
2191
+ "loss": 1.0007,
2192
+ "step": 182000
2193
+ },
2194
+ {
2195
+ "epoch": 2.35,
2196
+ "learning_rate": 1.0782544041740267e-05,
2197
+ "loss": 1.004,
2198
+ "step": 182500
2199
+ },
2200
+ {
2201
+ "epoch": 2.36,
2202
+ "learning_rate": 1.0675098956923117e-05,
2203
+ "loss": 0.9913,
2204
+ "step": 183000
2205
+ },
2206
+ {
2207
+ "epoch": 2.37,
2208
+ "learning_rate": 1.0567653872105968e-05,
2209
+ "loss": 1.0002,
2210
+ "step": 183500
2211
+ },
2212
+ {
2213
+ "epoch": 2.37,
2214
+ "learning_rate": 1.0460208787288817e-05,
2215
+ "loss": 1.0069,
2216
+ "step": 184000
2217
+ },
2218
+ {
2219
+ "epoch": 2.38,
2220
+ "learning_rate": 1.0352763702471668e-05,
2221
+ "loss": 0.9947,
2222
+ "step": 184500
2223
+ },
2224
+ {
2225
+ "epoch": 2.39,
2226
+ "learning_rate": 1.0245318617654517e-05,
2227
+ "loss": 0.9987,
2228
+ "step": 185000
2229
+ },
2230
+ {
2231
+ "epoch": 2.39,
2232
+ "learning_rate": 1.0137873532837368e-05,
2233
+ "loss": 0.9895,
2234
+ "step": 185500
2235
+ },
2236
+ {
2237
+ "epoch": 2.4,
2238
+ "learning_rate": 1.0030428448020217e-05,
2239
+ "loss": 0.9888,
2240
+ "step": 186000
2241
+ },
2242
+ {
2243
+ "epoch": 2.4,
2244
+ "learning_rate": 9.922983363203067e-06,
2245
+ "loss": 0.9905,
2246
+ "step": 186500
2247
+ },
2248
+ {
2249
+ "epoch": 2.41,
2250
+ "learning_rate": 9.815538278385918e-06,
2251
+ "loss": 0.9872,
2252
+ "step": 187000
2253
+ },
2254
+ {
2255
+ "epoch": 2.42,
2256
+ "learning_rate": 9.708093193568769e-06,
2257
+ "loss": 0.9855,
2258
+ "step": 187500
2259
+ },
2260
+ {
2261
+ "epoch": 2.42,
2262
+ "learning_rate": 9.600648108751618e-06,
2263
+ "loss": 1.0014,
2264
+ "step": 188000
2265
+ },
2266
+ {
2267
+ "epoch": 2.43,
2268
+ "learning_rate": 9.493203023934467e-06,
2269
+ "loss": 0.9987,
2270
+ "step": 188500
2271
+ },
2272
+ {
2273
+ "epoch": 2.44,
2274
+ "learning_rate": 9.385757939117317e-06,
2275
+ "loss": 0.9986,
2276
+ "step": 189000
2277
+ },
2278
+ {
2279
+ "epoch": 2.44,
2280
+ "learning_rate": 9.278312854300168e-06,
2281
+ "loss": 1.0028,
2282
+ "step": 189500
2283
+ },
2284
+ {
2285
+ "epoch": 2.45,
2286
+ "learning_rate": 9.170867769483017e-06,
2287
+ "loss": 0.9882,
2288
+ "step": 190000
2289
+ },
2290
+ {
2291
+ "epoch": 2.46,
2292
+ "learning_rate": 9.063422684665868e-06,
2293
+ "loss": 0.983,
2294
+ "step": 190500
2295
+ },
2296
+ {
2297
+ "epoch": 2.46,
2298
+ "learning_rate": 8.955977599848719e-06,
2299
+ "loss": 0.984,
2300
+ "step": 191000
2301
+ },
2302
+ {
2303
+ "epoch": 2.47,
2304
+ "learning_rate": 8.848532515031568e-06,
2305
+ "loss": 1.0114,
2306
+ "step": 191500
2307
+ },
2308
+ {
2309
+ "epoch": 2.48,
2310
+ "learning_rate": 8.741087430214417e-06,
2311
+ "loss": 0.9921,
2312
+ "step": 192000
2313
+ },
2314
+ {
2315
+ "epoch": 2.48,
2316
+ "learning_rate": 8.633642345397267e-06,
2317
+ "loss": 0.9994,
2318
+ "step": 192500
2319
+ },
2320
+ {
2321
+ "epoch": 2.49,
2322
+ "learning_rate": 8.526197260580118e-06,
2323
+ "loss": 0.9947,
2324
+ "step": 193000
2325
+ },
2326
+ {
2327
+ "epoch": 2.49,
2328
+ "learning_rate": 8.418752175762967e-06,
2329
+ "loss": 0.9933,
2330
+ "step": 193500
2331
+ },
2332
+ {
2333
+ "epoch": 2.5,
2334
+ "learning_rate": 8.311307090945818e-06,
2335
+ "loss": 0.985,
2336
+ "step": 194000
2337
+ },
2338
+ {
2339
+ "epoch": 2.51,
2340
+ "learning_rate": 8.203862006128668e-06,
2341
+ "loss": 0.9863,
2342
+ "step": 194500
2343
+ },
2344
+ {
2345
+ "epoch": 2.51,
2346
+ "learning_rate": 8.09641692131152e-06,
2347
+ "loss": 0.9924,
2348
+ "step": 195000
2349
+ },
2350
+ {
2351
+ "epoch": 2.52,
2352
+ "learning_rate": 7.988971836494368e-06,
2353
+ "loss": 0.9814,
2354
+ "step": 195500
2355
+ },
2356
+ {
2357
+ "epoch": 2.53,
2358
+ "learning_rate": 7.881526751677217e-06,
2359
+ "loss": 0.9837,
2360
+ "step": 196000
2361
+ },
2362
+ {
2363
+ "epoch": 2.53,
2364
+ "learning_rate": 7.774081666860068e-06,
2365
+ "loss": 0.9849,
2366
+ "step": 196500
2367
+ },
2368
+ {
2369
+ "epoch": 2.54,
2370
+ "learning_rate": 7.666636582042919e-06,
2371
+ "loss": 1.0009,
2372
+ "step": 197000
2373
+ },
2374
+ {
2375
+ "epoch": 2.55,
2376
+ "learning_rate": 7.559191497225768e-06,
2377
+ "loss": 1.0074,
2378
+ "step": 197500
2379
+ },
2380
+ {
2381
+ "epoch": 2.55,
2382
+ "learning_rate": 7.451746412408618e-06,
2383
+ "loss": 0.9983,
2384
+ "step": 198000
2385
+ },
2386
+ {
2387
+ "epoch": 2.56,
2388
+ "learning_rate": 7.344301327591468e-06,
2389
+ "loss": 1.0009,
2390
+ "step": 198500
2391
+ },
2392
+ {
2393
+ "epoch": 2.57,
2394
+ "learning_rate": 7.236856242774319e-06,
2395
+ "loss": 0.9858,
2396
+ "step": 199000
2397
+ },
2398
+ {
2399
+ "epoch": 2.57,
2400
+ "learning_rate": 7.129411157957168e-06,
2401
+ "loss": 0.9908,
2402
+ "step": 199500
2403
+ },
2404
+ {
2405
+ "epoch": 2.58,
2406
+ "learning_rate": 7.021966073140019e-06,
2407
+ "loss": 0.973,
2408
+ "step": 200000
2409
+ },
2410
+ {
2411
+ "epoch": 2.59,
2412
+ "learning_rate": 6.9145209883228686e-06,
2413
+ "loss": 0.98,
2414
+ "step": 200500
2415
+ },
2416
+ {
2417
+ "epoch": 2.59,
2418
+ "learning_rate": 6.807075903505719e-06,
2419
+ "loss": 0.9783,
2420
+ "step": 201000
2421
+ },
2422
+ {
2423
+ "epoch": 2.6,
2424
+ "learning_rate": 6.699630818688568e-06,
2425
+ "loss": 0.9697,
2426
+ "step": 201500
2427
+ },
2428
+ {
2429
+ "epoch": 2.6,
2430
+ "learning_rate": 6.592185733871418e-06,
2431
+ "loss": 0.9787,
2432
+ "step": 202000
2433
+ },
2434
+ {
2435
+ "epoch": 2.61,
2436
+ "learning_rate": 6.484740649054269e-06,
2437
+ "loss": 0.9739,
2438
+ "step": 202500
2439
+ },
2440
+ {
2441
+ "epoch": 2.62,
2442
+ "learning_rate": 6.37729556423712e-06,
2443
+ "loss": 0.9966,
2444
+ "step": 203000
2445
+ },
2446
+ {
2447
+ "epoch": 2.62,
2448
+ "learning_rate": 6.269850479419969e-06,
2449
+ "loss": 0.9894,
2450
+ "step": 203500
2451
+ },
2452
+ {
2453
+ "epoch": 2.63,
2454
+ "learning_rate": 6.1624053946028185e-06,
2455
+ "loss": 0.965,
2456
+ "step": 204000
2457
+ },
2458
+ {
2459
+ "epoch": 2.64,
2460
+ "learning_rate": 6.054960309785669e-06,
2461
+ "loss": 0.9836,
2462
+ "step": 204500
2463
+ },
2464
+ {
2465
+ "epoch": 2.64,
2466
+ "learning_rate": 5.947515224968519e-06,
2467
+ "loss": 0.9728,
2468
+ "step": 205000
2469
+ },
2470
+ {
2471
+ "epoch": 2.65,
2472
+ "learning_rate": 5.840070140151369e-06,
2473
+ "loss": 0.977,
2474
+ "step": 205500
2475
+ },
2476
+ {
2477
+ "epoch": 2.66,
2478
+ "learning_rate": 5.732625055334219e-06,
2479
+ "loss": 0.9655,
2480
+ "step": 206000
2481
+ },
2482
+ {
2483
+ "epoch": 2.66,
2484
+ "learning_rate": 5.625179970517069e-06,
2485
+ "loss": 0.9802,
2486
+ "step": 206500
2487
+ },
2488
+ {
2489
+ "epoch": 2.67,
2490
+ "learning_rate": 5.517734885699919e-06,
2491
+ "loss": 0.9992,
2492
+ "step": 207000
2493
+ },
2494
+ {
2495
+ "epoch": 2.68,
2496
+ "learning_rate": 5.410289800882768e-06,
2497
+ "loss": 0.9975,
2498
+ "step": 207500
2499
+ },
2500
+ {
2501
+ "epoch": 2.68,
2502
+ "learning_rate": 5.302844716065619e-06,
2503
+ "loss": 0.9721,
2504
+ "step": 208000
2505
+ },
2506
+ {
2507
+ "epoch": 2.69,
2508
+ "learning_rate": 5.195399631248469e-06,
2509
+ "loss": 0.9695,
2510
+ "step": 208500
2511
+ },
2512
+ {
2513
+ "epoch": 2.69,
2514
+ "learning_rate": 5.08795454643132e-06,
2515
+ "loss": 0.9832,
2516
+ "step": 209000
2517
+ },
2518
+ {
2519
+ "epoch": 2.7,
2520
+ "learning_rate": 4.980509461614169e-06,
2521
+ "loss": 0.9873,
2522
+ "step": 209500
2523
+ },
2524
+ {
2525
+ "epoch": 2.71,
2526
+ "learning_rate": 4.873064376797019e-06,
2527
+ "loss": 0.9886,
2528
+ "step": 210000
2529
+ },
2530
+ {
2531
+ "epoch": 2.71,
2532
+ "learning_rate": 4.765619291979869e-06,
2533
+ "loss": 0.9753,
2534
+ "step": 210500
2535
+ },
2536
+ {
2537
+ "epoch": 2.72,
2538
+ "learning_rate": 4.658174207162719e-06,
2539
+ "loss": 0.9785,
2540
+ "step": 211000
2541
+ },
2542
+ {
2543
+ "epoch": 2.73,
2544
+ "learning_rate": 4.550729122345569e-06,
2545
+ "loss": 0.9704,
2546
+ "step": 211500
2547
+ },
2548
+ {
2549
+ "epoch": 2.73,
2550
+ "learning_rate": 4.44328403752842e-06,
2551
+ "loss": 0.9855,
2552
+ "step": 212000
2553
+ },
2554
+ {
2555
+ "epoch": 2.74,
2556
+ "learning_rate": 4.3358389527112696e-06,
2557
+ "loss": 0.9773,
2558
+ "step": 212500
2559
+ },
2560
+ {
2561
+ "epoch": 2.75,
2562
+ "learning_rate": 4.2283938678941194e-06,
2563
+ "loss": 0.9894,
2564
+ "step": 213000
2565
+ },
2566
+ {
2567
+ "epoch": 2.75,
2568
+ "learning_rate": 4.120948783076969e-06,
2569
+ "loss": 0.9701,
2570
+ "step": 213500
2571
+ },
2572
+ {
2573
+ "epoch": 2.76,
2574
+ "learning_rate": 4.01350369825982e-06,
2575
+ "loss": 0.9865,
2576
+ "step": 214000
2577
+ },
2578
+ {
2579
+ "epoch": 2.77,
2580
+ "learning_rate": 3.90605861344267e-06,
2581
+ "loss": 0.9934,
2582
+ "step": 214500
2583
+ },
2584
+ {
2585
+ "epoch": 2.77,
2586
+ "learning_rate": 3.7986135286255197e-06,
2587
+ "loss": 0.9765,
2588
+ "step": 215000
2589
+ },
2590
+ {
2591
+ "epoch": 2.78,
2592
+ "learning_rate": 3.6911684438083696e-06,
2593
+ "loss": 0.9633,
2594
+ "step": 215500
2595
+ },
2596
+ {
2597
+ "epoch": 2.78,
2598
+ "learning_rate": 3.58372335899122e-06,
2599
+ "loss": 0.9625,
2600
+ "step": 216000
2601
+ },
2602
+ {
2603
+ "epoch": 2.79,
2604
+ "learning_rate": 3.4762782741740698e-06,
2605
+ "loss": 0.9676,
2606
+ "step": 216500
2607
+ },
2608
+ {
2609
+ "epoch": 2.8,
2610
+ "learning_rate": 3.36883318935692e-06,
2611
+ "loss": 0.9642,
2612
+ "step": 217000
2613
+ },
2614
+ {
2615
+ "epoch": 2.8,
2616
+ "learning_rate": 3.26138810453977e-06,
2617
+ "loss": 0.9923,
2618
+ "step": 217500
2619
+ },
2620
+ {
2621
+ "epoch": 2.81,
2622
+ "learning_rate": 3.15394301972262e-06,
2623
+ "loss": 0.9641,
2624
+ "step": 218000
2625
+ },
2626
+ {
2627
+ "epoch": 2.82,
2628
+ "learning_rate": 3.04649793490547e-06,
2629
+ "loss": 0.9727,
2630
+ "step": 218500
2631
+ },
2632
+ {
2633
+ "epoch": 2.82,
2634
+ "learning_rate": 2.93905285008832e-06,
2635
+ "loss": 0.9868,
2636
+ "step": 219000
2637
+ },
2638
+ {
2639
+ "epoch": 2.83,
2640
+ "learning_rate": 2.8316077652711702e-06,
2641
+ "loss": 0.9504,
2642
+ "step": 219500
2643
+ },
2644
+ {
2645
+ "epoch": 2.84,
2646
+ "learning_rate": 2.72416268045402e-06,
2647
+ "loss": 0.9811,
2648
+ "step": 220000
2649
+ },
2650
+ {
2651
+ "epoch": 2.84,
2652
+ "learning_rate": 2.6167175956368704e-06,
2653
+ "loss": 0.9813,
2654
+ "step": 220500
2655
+ },
2656
+ {
2657
+ "epoch": 2.85,
2658
+ "learning_rate": 2.5092725108197202e-06,
2659
+ "loss": 0.9687,
2660
+ "step": 221000
2661
+ },
2662
+ {
2663
+ "epoch": 2.86,
2664
+ "learning_rate": 2.40182742600257e-06,
2665
+ "loss": 0.9673,
2666
+ "step": 221500
2667
+ },
2668
+ {
2669
+ "epoch": 2.86,
2670
+ "learning_rate": 2.2943823411854204e-06,
2671
+ "loss": 0.9653,
2672
+ "step": 222000
2673
+ },
2674
+ {
2675
+ "epoch": 2.87,
2676
+ "learning_rate": 2.1869372563682703e-06,
2677
+ "loss": 0.9479,
2678
+ "step": 222500
2679
+ },
2680
+ {
2681
+ "epoch": 2.88,
2682
+ "learning_rate": 2.0794921715511205e-06,
2683
+ "loss": 0.9802,
2684
+ "step": 223000
2685
+ },
2686
+ {
2687
+ "epoch": 2.88,
2688
+ "learning_rate": 1.9720470867339704e-06,
2689
+ "loss": 0.9624,
2690
+ "step": 223500
2691
+ },
2692
+ {
2693
+ "epoch": 2.89,
2694
+ "learning_rate": 1.8646020019168205e-06,
2695
+ "loss": 0.9822,
2696
+ "step": 224000
2697
+ },
2698
+ {
2699
+ "epoch": 2.89,
2700
+ "learning_rate": 1.7571569170996706e-06,
2701
+ "loss": 0.9747,
2702
+ "step": 224500
2703
+ },
2704
+ {
2705
+ "epoch": 2.9,
2706
+ "learning_rate": 1.6497118322825206e-06,
2707
+ "loss": 0.9614,
2708
+ "step": 225000
2709
+ },
2710
+ {
2711
+ "epoch": 2.91,
2712
+ "learning_rate": 1.5422667474653705e-06,
2713
+ "loss": 0.9729,
2714
+ "step": 225500
2715
+ },
2716
+ {
2717
+ "epoch": 2.91,
2718
+ "learning_rate": 1.4348216626482206e-06,
2719
+ "loss": 0.9679,
2720
+ "step": 226000
2721
+ },
2722
+ {
2723
+ "epoch": 2.92,
2724
+ "learning_rate": 1.3273765778310707e-06,
2725
+ "loss": 0.9551,
2726
+ "step": 226500
2727
+ },
2728
+ {
2729
+ "epoch": 2.93,
2730
+ "learning_rate": 1.2199314930139207e-06,
2731
+ "loss": 0.9762,
2732
+ "step": 227000
2733
+ },
2734
+ {
2735
+ "epoch": 2.93,
2736
+ "learning_rate": 1.1124864081967708e-06,
2737
+ "loss": 0.9737,
2738
+ "step": 227500
2739
+ },
2740
+ {
2741
+ "epoch": 2.94,
2742
+ "learning_rate": 1.0050413233796209e-06,
2743
+ "loss": 0.9718,
2744
+ "step": 228000
2745
+ },
2746
+ {
2747
+ "epoch": 2.95,
2748
+ "learning_rate": 8.975962385624709e-07,
2749
+ "loss": 0.9666,
2750
+ "step": 228500
2751
+ },
2752
+ {
2753
+ "epoch": 2.95,
2754
+ "learning_rate": 7.901511537453207e-07,
2755
+ "loss": 0.9854,
2756
+ "step": 229000
2757
+ },
2758
+ {
2759
+ "epoch": 2.96,
2760
+ "learning_rate": 6.827060689281709e-07,
2761
+ "loss": 0.9535,
2762
+ "step": 229500
2763
+ },
2764
+ {
2765
+ "epoch": 2.97,
2766
+ "learning_rate": 5.752609841110209e-07,
2767
+ "loss": 0.9665,
2768
+ "step": 230000
2769
+ },
2770
+ {
2771
+ "epoch": 2.97,
2772
+ "learning_rate": 4.678158992938709e-07,
2773
+ "loss": 0.9463,
2774
+ "step": 230500
2775
+ },
2776
+ {
2777
+ "epoch": 2.98,
2778
+ "learning_rate": 3.60370814476721e-07,
2779
+ "loss": 0.9664,
2780
+ "step": 231000
2781
+ },
2782
+ {
2783
+ "epoch": 2.98,
2784
+ "learning_rate": 2.52925729659571e-07,
2785
+ "loss": 0.9573,
2786
+ "step": 231500
2787
+ },
2788
+ {
2789
+ "epoch": 2.99,
2790
+ "learning_rate": 1.4548064484242105e-07,
2791
+ "loss": 0.95,
2792
+ "step": 232000
2793
+ },
2794
+ {
2795
+ "epoch": 3.0,
2796
+ "learning_rate": 3.8035560025271085e-08,
2797
+ "loss": 0.9636,
2798
+ "step": 232500
2799
+ },
2800
+ {
2801
+ "epoch": 3.0,
2802
+ "step": 232677,
2803
+ "total_flos": 3.713762606980301e+17,
2804
+ "train_loss": 1.2468314008385242,
2805
+ "train_runtime": 62368.4442,
2806
+ "train_samples_per_second": 37.307,
2807
+ "train_steps_per_second": 3.731
2808
+ }
2809
+ ],
2810
+ "max_steps": 232677,
2811
+ "num_train_epochs": 3,
2812
+ "total_flos": 3.713762606980301e+17,
2813
+ "trial_name": null,
2814
+ "trial_params": null
2815
+ }