chavinlo commited on
Commit
5a3b2f2
1 Parent(s): fa02585

Upload checkpoint-500/trainer_state.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. checkpoint-500/trainer_state.json +3016 -0
checkpoint-500/trainer_state.json ADDED
@@ -0,0 +1,3016 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.2313668179370818,
5
+ "global_step": 500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 2.0000000000000002e-07,
13
+ "loss": 2.0749,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 4.0000000000000003e-07,
19
+ "loss": 2.0939,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.01,
24
+ "learning_rate": 6.000000000000001e-07,
25
+ "loss": 2.0808,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.01,
30
+ "learning_rate": 8.000000000000001e-07,
31
+ "loss": 2.0523,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.01,
36
+ "learning_rate": 1.0000000000000002e-06,
37
+ "loss": 2.0731,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.01,
42
+ "learning_rate": 1.2000000000000002e-06,
43
+ "loss": 2.0781,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.02,
48
+ "learning_rate": 1.4000000000000001e-06,
49
+ "loss": 2.0222,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.02,
54
+ "learning_rate": 1.6000000000000001e-06,
55
+ "loss": 2.0379,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.02,
60
+ "learning_rate": 1.8000000000000001e-06,
61
+ "loss": 2.0547,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.02,
66
+ "learning_rate": 2.0000000000000003e-06,
67
+ "loss": 2.088,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.03,
72
+ "learning_rate": 2.2e-06,
73
+ "loss": 2.0938,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.03,
78
+ "learning_rate": 2.4000000000000003e-06,
79
+ "loss": 1.9722,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.03,
84
+ "learning_rate": 2.6e-06,
85
+ "loss": 2.0697,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.03,
90
+ "learning_rate": 2.8000000000000003e-06,
91
+ "loss": 2.0278,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.04,
96
+ "learning_rate": 3e-06,
97
+ "loss": 2.0449,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.04,
102
+ "learning_rate": 3.2000000000000003e-06,
103
+ "loss": 2.0585,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.04,
108
+ "learning_rate": 3.4000000000000005e-06,
109
+ "loss": 2.0547,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.04,
114
+ "learning_rate": 3.6000000000000003e-06,
115
+ "loss": 2.0735,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.05,
120
+ "learning_rate": 3.8000000000000005e-06,
121
+ "loss": 2.1028,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.05,
126
+ "learning_rate": 4.000000000000001e-06,
127
+ "loss": 2.1025,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.05,
132
+ "learning_rate": 4.2000000000000004e-06,
133
+ "loss": 2.0129,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.05,
138
+ "learning_rate": 4.4e-06,
139
+ "loss": 2.0452,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.06,
144
+ "learning_rate": 4.600000000000001e-06,
145
+ "loss": 2.0978,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.06,
150
+ "learning_rate": 4.800000000000001e-06,
151
+ "loss": 2.0609,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.06,
156
+ "learning_rate": 5e-06,
157
+ "loss": 2.0783,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.06,
162
+ "learning_rate": 5.2e-06,
163
+ "loss": 2.0535,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.07,
168
+ "learning_rate": 5.400000000000001e-06,
169
+ "loss": 2.0383,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.07,
174
+ "learning_rate": 5.600000000000001e-06,
175
+ "loss": 2.0492,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.07,
180
+ "learning_rate": 5.8e-06,
181
+ "loss": 2.097,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.07,
186
+ "learning_rate": 6e-06,
187
+ "loss": 2.0641,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.08,
192
+ "learning_rate": 6.200000000000001e-06,
193
+ "loss": 2.0511,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.08,
198
+ "learning_rate": 6.4000000000000006e-06,
199
+ "loss": 2.0927,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.08,
204
+ "learning_rate": 6.600000000000001e-06,
205
+ "loss": 1.9922,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.08,
210
+ "learning_rate": 6.800000000000001e-06,
211
+ "loss": 2.0557,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.09,
216
+ "learning_rate": 7e-06,
217
+ "loss": 1.992,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.09,
222
+ "learning_rate": 7.2000000000000005e-06,
223
+ "loss": 2.0997,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.09,
228
+ "learning_rate": 7.4e-06,
229
+ "loss": 2.0116,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.09,
234
+ "learning_rate": 7.600000000000001e-06,
235
+ "loss": 1.9953,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.1,
240
+ "learning_rate": 7.800000000000002e-06,
241
+ "loss": 2.043,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.1,
246
+ "learning_rate": 8.000000000000001e-06,
247
+ "loss": 2.0136,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.1,
252
+ "learning_rate": 8.2e-06,
253
+ "loss": 2.0295,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.1,
258
+ "learning_rate": 8.400000000000001e-06,
259
+ "loss": 2.0217,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.11,
264
+ "learning_rate": 8.6e-06,
265
+ "loss": 2.0433,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.11,
270
+ "learning_rate": 8.8e-06,
271
+ "loss": 1.937,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.11,
276
+ "learning_rate": 9e-06,
277
+ "loss": 1.9604,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 0.11,
282
+ "learning_rate": 9.200000000000002e-06,
283
+ "loss": 1.9755,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 0.12,
288
+ "learning_rate": 9.4e-06,
289
+ "loss": 1.9808,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 0.12,
294
+ "learning_rate": 9.600000000000001e-06,
295
+ "loss": 2.0126,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 0.12,
300
+ "learning_rate": 9.800000000000001e-06,
301
+ "loss": 2.0072,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 0.12,
306
+ "learning_rate": 1e-05,
307
+ "loss": 2.0112,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 0.13,
312
+ "learning_rate": 1.02e-05,
313
+ "loss": 2.0066,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 0.13,
318
+ "learning_rate": 1.04e-05,
319
+ "loss": 1.9487,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 0.13,
324
+ "learning_rate": 1.0600000000000002e-05,
325
+ "loss": 2.0321,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 0.13,
330
+ "learning_rate": 1.0800000000000002e-05,
331
+ "loss": 1.9578,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 0.14,
336
+ "learning_rate": 1.1000000000000001e-05,
337
+ "loss": 2.0002,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 0.14,
342
+ "learning_rate": 1.1200000000000001e-05,
343
+ "loss": 1.9924,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 0.14,
348
+ "learning_rate": 1.14e-05,
349
+ "loss": 1.9803,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 0.14,
354
+ "learning_rate": 1.16e-05,
355
+ "loss": 1.9154,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 0.15,
360
+ "learning_rate": 1.18e-05,
361
+ "loss": 1.9307,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 0.15,
366
+ "learning_rate": 1.2e-05,
367
+ "loss": 1.9131,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 0.15,
372
+ "learning_rate": 1.22e-05,
373
+ "loss": 1.9555,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 0.15,
378
+ "learning_rate": 1.2400000000000002e-05,
379
+ "loss": 1.9246,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 0.16,
384
+ "learning_rate": 1.2600000000000001e-05,
385
+ "loss": 1.9316,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 0.16,
390
+ "learning_rate": 1.2800000000000001e-05,
391
+ "loss": 1.8861,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 0.16,
396
+ "learning_rate": 1.3000000000000001e-05,
397
+ "loss": 1.9391,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 0.16,
402
+ "learning_rate": 1.3200000000000002e-05,
403
+ "loss": 1.8988,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 0.16,
408
+ "learning_rate": 1.3400000000000002e-05,
409
+ "loss": 1.8759,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 0.17,
414
+ "learning_rate": 1.3600000000000002e-05,
415
+ "loss": 1.8406,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 0.17,
420
+ "learning_rate": 1.38e-05,
421
+ "loss": 1.8932,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 0.17,
426
+ "learning_rate": 1.4e-05,
427
+ "loss": 1.8809,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 0.17,
432
+ "learning_rate": 1.4200000000000001e-05,
433
+ "loss": 1.877,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 0.18,
438
+ "learning_rate": 1.4400000000000001e-05,
439
+ "loss": 1.8344,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 0.18,
444
+ "learning_rate": 1.46e-05,
445
+ "loss": 1.8449,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 0.18,
450
+ "learning_rate": 1.48e-05,
451
+ "loss": 1.7725,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 0.18,
456
+ "learning_rate": 1.5000000000000002e-05,
457
+ "loss": 1.7643,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 0.19,
462
+ "learning_rate": 1.5200000000000002e-05,
463
+ "loss": 1.7966,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 0.19,
468
+ "learning_rate": 1.54e-05,
469
+ "loss": 1.8069,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 0.19,
474
+ "learning_rate": 1.5600000000000003e-05,
475
+ "loss": 1.7931,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 0.19,
480
+ "learning_rate": 1.58e-05,
481
+ "loss": 1.7639,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 0.2,
486
+ "learning_rate": 1.6000000000000003e-05,
487
+ "loss": 1.7514,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 0.2,
492
+ "learning_rate": 1.62e-05,
493
+ "loss": 1.7457,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 0.2,
498
+ "learning_rate": 1.64e-05,
499
+ "loss": 1.7195,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 0.2,
504
+ "learning_rate": 1.66e-05,
505
+ "loss": 1.7307,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 0.21,
510
+ "learning_rate": 1.6800000000000002e-05,
511
+ "loss": 1.6572,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 0.21,
516
+ "learning_rate": 1.7e-05,
517
+ "loss": 1.6728,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 0.21,
522
+ "learning_rate": 1.72e-05,
523
+ "loss": 1.6499,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 0.21,
528
+ "learning_rate": 1.7400000000000003e-05,
529
+ "loss": 1.6669,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 0.22,
534
+ "learning_rate": 1.76e-05,
535
+ "loss": 1.5629,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 0.22,
540
+ "learning_rate": 1.7800000000000002e-05,
541
+ "loss": 1.5887,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 0.22,
546
+ "learning_rate": 1.8e-05,
547
+ "loss": 1.5662,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 0.22,
552
+ "learning_rate": 1.8200000000000002e-05,
553
+ "loss": 1.574,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 0.23,
558
+ "learning_rate": 1.8400000000000003e-05,
559
+ "loss": 1.5435,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 0.23,
564
+ "learning_rate": 1.86e-05,
565
+ "loss": 1.5675,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 0.23,
570
+ "learning_rate": 1.88e-05,
571
+ "loss": 1.5504,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 0.23,
576
+ "learning_rate": 1.9e-05,
577
+ "loss": 1.5344,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 0.24,
582
+ "learning_rate": 1.9200000000000003e-05,
583
+ "loss": 1.5012,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 0.24,
588
+ "learning_rate": 1.94e-05,
589
+ "loss": 1.4572,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 0.24,
594
+ "learning_rate": 1.9600000000000002e-05,
595
+ "loss": 1.4746,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 0.24,
600
+ "learning_rate": 1.98e-05,
601
+ "loss": 1.3997,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 0.25,
606
+ "learning_rate": 2e-05,
607
+ "loss": 1.4291,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 0.25,
612
+ "learning_rate": 1.9982110912343474e-05,
613
+ "loss": 1.3956,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 0.25,
618
+ "learning_rate": 1.9964221824686943e-05,
619
+ "loss": 1.3893,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 0.25,
624
+ "learning_rate": 1.9946332737030415e-05,
625
+ "loss": 1.3612,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 0.26,
630
+ "learning_rate": 1.9928443649373884e-05,
631
+ "loss": 1.3851,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 0.26,
636
+ "learning_rate": 1.9910554561717356e-05,
637
+ "loss": 1.4022,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 0.26,
642
+ "learning_rate": 1.9892665474060825e-05,
643
+ "loss": 1.3556,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 0.26,
648
+ "learning_rate": 1.9874776386404297e-05,
649
+ "loss": 1.3112,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 0.27,
654
+ "learning_rate": 1.9856887298747766e-05,
655
+ "loss": 1.3367,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 0.27,
660
+ "learning_rate": 1.9838998211091234e-05,
661
+ "loss": 1.2876,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 0.27,
666
+ "learning_rate": 1.9821109123434707e-05,
667
+ "loss": 1.2824,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 0.27,
672
+ "learning_rate": 1.9803220035778175e-05,
673
+ "loss": 1.2417,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 0.28,
678
+ "learning_rate": 1.9785330948121648e-05,
679
+ "loss": 1.2683,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 0.28,
684
+ "learning_rate": 1.9767441860465116e-05,
685
+ "loss": 1.2332,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 0.28,
690
+ "learning_rate": 1.974955277280859e-05,
691
+ "loss": 1.2492,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 0.28,
696
+ "learning_rate": 1.9731663685152057e-05,
697
+ "loss": 1.2397,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 0.29,
702
+ "learning_rate": 1.971377459749553e-05,
703
+ "loss": 1.2537,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 0.29,
708
+ "learning_rate": 1.9695885509839e-05,
709
+ "loss": 1.2242,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 0.29,
714
+ "learning_rate": 1.967799642218247e-05,
715
+ "loss": 1.211,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 0.29,
720
+ "learning_rate": 1.9660107334525943e-05,
721
+ "loss": 1.1657,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 0.3,
726
+ "learning_rate": 1.964221824686941e-05,
727
+ "loss": 1.1705,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 0.3,
732
+ "learning_rate": 1.9624329159212884e-05,
733
+ "loss": 1.1687,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 0.3,
738
+ "learning_rate": 1.9606440071556353e-05,
739
+ "loss": 1.1469,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 0.3,
744
+ "learning_rate": 1.9588550983899825e-05,
745
+ "loss": 1.1621,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 0.31,
750
+ "learning_rate": 1.9570661896243294e-05,
751
+ "loss": 1.1559,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 0.31,
756
+ "learning_rate": 1.9552772808586762e-05,
757
+ "loss": 1.1086,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 0.31,
762
+ "learning_rate": 1.9534883720930235e-05,
763
+ "loss": 1.089,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 0.31,
768
+ "learning_rate": 1.9516994633273703e-05,
769
+ "loss": 1.0888,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 0.32,
774
+ "learning_rate": 1.9499105545617176e-05,
775
+ "loss": 1.0849,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 0.32,
780
+ "learning_rate": 1.9481216457960644e-05,
781
+ "loss": 1.0562,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 0.32,
786
+ "learning_rate": 1.9463327370304117e-05,
787
+ "loss": 1.0523,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 0.32,
792
+ "learning_rate": 1.9445438282647585e-05,
793
+ "loss": 1.0164,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 0.32,
798
+ "learning_rate": 1.9427549194991058e-05,
799
+ "loss": 1.0467,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 0.33,
804
+ "learning_rate": 1.9409660107334526e-05,
805
+ "loss": 1.0242,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 0.33,
810
+ "learning_rate": 1.9391771019678e-05,
811
+ "loss": 1.0218,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 0.33,
816
+ "learning_rate": 1.9373881932021467e-05,
817
+ "loss": 0.988,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 0.33,
822
+ "learning_rate": 1.935599284436494e-05,
823
+ "loss": 1.0138,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 0.34,
828
+ "learning_rate": 1.933810375670841e-05,
829
+ "loss": 1.0288,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 0.34,
834
+ "learning_rate": 1.932021466905188e-05,
835
+ "loss": 0.9849,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 0.34,
840
+ "learning_rate": 1.9302325581395353e-05,
841
+ "loss": 1.0165,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 0.34,
846
+ "learning_rate": 1.928443649373882e-05,
847
+ "loss": 0.979,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 0.35,
852
+ "learning_rate": 1.9266547406082294e-05,
853
+ "loss": 0.9418,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 0.35,
858
+ "learning_rate": 1.9248658318425763e-05,
859
+ "loss": 0.9832,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 0.35,
864
+ "learning_rate": 1.923076923076923e-05,
865
+ "loss": 0.9422,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 0.35,
870
+ "learning_rate": 1.9212880143112704e-05,
871
+ "loss": 1.0011,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 0.36,
876
+ "learning_rate": 1.9194991055456172e-05,
877
+ "loss": 0.9562,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 0.36,
882
+ "learning_rate": 1.917710196779964e-05,
883
+ "loss": 0.9214,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 0.36,
888
+ "learning_rate": 1.9159212880143113e-05,
889
+ "loss": 0.9713,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 0.36,
894
+ "learning_rate": 1.9141323792486586e-05,
895
+ "loss": 0.9227,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 0.37,
900
+ "learning_rate": 1.9123434704830054e-05,
901
+ "loss": 0.944,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 0.37,
906
+ "learning_rate": 1.9105545617173527e-05,
907
+ "loss": 0.9623,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 0.37,
912
+ "learning_rate": 1.9087656529516995e-05,
913
+ "loss": 0.9324,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 0.37,
918
+ "learning_rate": 1.9069767441860468e-05,
919
+ "loss": 0.941,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 0.38,
924
+ "learning_rate": 1.9051878354203936e-05,
925
+ "loss": 0.9442,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 0.38,
930
+ "learning_rate": 1.903398926654741e-05,
931
+ "loss": 0.8992,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 0.38,
936
+ "learning_rate": 1.9016100178890877e-05,
937
+ "loss": 0.8882,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 0.38,
942
+ "learning_rate": 1.899821109123435e-05,
943
+ "loss": 0.9508,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 0.39,
948
+ "learning_rate": 1.8980322003577822e-05,
949
+ "loss": 0.9295,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 0.39,
954
+ "learning_rate": 1.896243291592129e-05,
955
+ "loss": 0.9519,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 0.39,
960
+ "learning_rate": 1.894454382826476e-05,
961
+ "loss": 0.9342,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 0.39,
966
+ "learning_rate": 1.892665474060823e-05,
967
+ "loss": 0.9766,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 0.4,
972
+ "learning_rate": 1.89087656529517e-05,
973
+ "loss": 0.9625,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 0.4,
978
+ "learning_rate": 1.889087656529517e-05,
979
+ "loss": 0.8774,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 0.4,
984
+ "learning_rate": 1.887298747763864e-05,
985
+ "loss": 0.9137,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 0.4,
990
+ "learning_rate": 1.885509838998211e-05,
991
+ "loss": 0.9574,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 0.41,
996
+ "learning_rate": 1.8837209302325582e-05,
997
+ "loss": 0.9246,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 0.41,
1002
+ "learning_rate": 1.881932021466905e-05,
1003
+ "loss": 0.8908,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 0.41,
1008
+ "learning_rate": 1.8801431127012523e-05,
1009
+ "loss": 0.8991,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 0.41,
1014
+ "learning_rate": 1.8783542039355996e-05,
1015
+ "loss": 0.9756,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 0.42,
1020
+ "learning_rate": 1.8765652951699464e-05,
1021
+ "loss": 0.9422,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 0.42,
1026
+ "learning_rate": 1.8747763864042937e-05,
1027
+ "loss": 0.9082,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 0.42,
1032
+ "learning_rate": 1.8729874776386405e-05,
1033
+ "loss": 0.9071,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 0.42,
1038
+ "learning_rate": 1.8711985688729878e-05,
1039
+ "loss": 0.9027,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 0.43,
1044
+ "learning_rate": 1.8694096601073346e-05,
1045
+ "loss": 0.8683,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 0.43,
1050
+ "learning_rate": 1.867620751341682e-05,
1051
+ "loss": 0.8829,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 0.43,
1056
+ "learning_rate": 1.8658318425760287e-05,
1057
+ "loss": 0.9089,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 0.43,
1062
+ "learning_rate": 1.864042933810376e-05,
1063
+ "loss": 0.9022,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 0.44,
1068
+ "learning_rate": 1.862254025044723e-05,
1069
+ "loss": 0.8839,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 0.44,
1074
+ "learning_rate": 1.86046511627907e-05,
1075
+ "loss": 0.9115,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 0.44,
1080
+ "learning_rate": 1.858676207513417e-05,
1081
+ "loss": 0.869,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 0.44,
1086
+ "learning_rate": 1.8568872987477638e-05,
1087
+ "loss": 0.9228,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 0.45,
1092
+ "learning_rate": 1.855098389982111e-05,
1093
+ "loss": 0.9513,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 0.45,
1098
+ "learning_rate": 1.853309481216458e-05,
1099
+ "loss": 0.8936,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 0.45,
1104
+ "learning_rate": 1.851520572450805e-05,
1105
+ "loss": 0.883,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 0.45,
1110
+ "learning_rate": 1.849731663685152e-05,
1111
+ "loss": 0.8078,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 0.46,
1116
+ "learning_rate": 1.8479427549194992e-05,
1117
+ "loss": 0.8923,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 0.46,
1122
+ "learning_rate": 1.8461538461538465e-05,
1123
+ "loss": 0.8929,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 0.46,
1128
+ "learning_rate": 1.8443649373881933e-05,
1129
+ "loss": 0.889,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 0.46,
1134
+ "learning_rate": 1.8425760286225406e-05,
1135
+ "loss": 0.8927,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 0.47,
1140
+ "learning_rate": 1.8407871198568874e-05,
1141
+ "loss": 0.8774,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 0.47,
1146
+ "learning_rate": 1.8389982110912347e-05,
1147
+ "loss": 0.8691,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 0.47,
1152
+ "learning_rate": 1.8372093023255815e-05,
1153
+ "loss": 0.9064,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 0.47,
1158
+ "learning_rate": 1.8354203935599288e-05,
1159
+ "loss": 0.8597,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 0.48,
1164
+ "learning_rate": 1.8336314847942756e-05,
1165
+ "loss": 0.8927,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 0.48,
1170
+ "learning_rate": 1.831842576028623e-05,
1171
+ "loss": 0.8337,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 0.48,
1176
+ "learning_rate": 1.8300536672629697e-05,
1177
+ "loss": 0.8601,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 0.48,
1182
+ "learning_rate": 1.8282647584973166e-05,
1183
+ "loss": 0.8689,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 0.48,
1188
+ "learning_rate": 1.826475849731664e-05,
1189
+ "loss": 0.8508,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 0.49,
1194
+ "learning_rate": 1.8246869409660107e-05,
1195
+ "loss": 0.8423,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 0.49,
1200
+ "learning_rate": 1.822898032200358e-05,
1201
+ "loss": 0.8897,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 0.49,
1206
+ "learning_rate": 1.8211091234347048e-05,
1207
+ "loss": 0.867,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 0.49,
1212
+ "learning_rate": 1.819320214669052e-05,
1213
+ "loss": 0.8933,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 0.5,
1218
+ "learning_rate": 1.817531305903399e-05,
1219
+ "loss": 0.9027,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 0.5,
1224
+ "learning_rate": 1.815742397137746e-05,
1225
+ "loss": 0.8983,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 0.5,
1230
+ "learning_rate": 1.813953488372093e-05,
1231
+ "loss": 0.8857,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 0.5,
1236
+ "learning_rate": 1.8121645796064402e-05,
1237
+ "loss": 0.8266,
1238
+ "step": 205
1239
+ },
1240
+ {
1241
+ "epoch": 0.51,
1242
+ "learning_rate": 1.8103756708407875e-05,
1243
+ "loss": 0.8404,
1244
+ "step": 206
1245
+ },
1246
+ {
1247
+ "epoch": 0.51,
1248
+ "learning_rate": 1.8085867620751343e-05,
1249
+ "loss": 0.8731,
1250
+ "step": 207
1251
+ },
1252
+ {
1253
+ "epoch": 0.51,
1254
+ "learning_rate": 1.8067978533094816e-05,
1255
+ "loss": 0.8429,
1256
+ "step": 208
1257
+ },
1258
+ {
1259
+ "epoch": 0.51,
1260
+ "learning_rate": 1.8050089445438284e-05,
1261
+ "loss": 0.8586,
1262
+ "step": 209
1263
+ },
1264
+ {
1265
+ "epoch": 0.52,
1266
+ "learning_rate": 1.8032200357781757e-05,
1267
+ "loss": 0.8492,
1268
+ "step": 210
1269
+ },
1270
+ {
1271
+ "epoch": 0.52,
1272
+ "learning_rate": 1.8014311270125225e-05,
1273
+ "loss": 0.8351,
1274
+ "step": 211
1275
+ },
1276
+ {
1277
+ "epoch": 0.52,
1278
+ "learning_rate": 1.7996422182468694e-05,
1279
+ "loss": 0.8566,
1280
+ "step": 212
1281
+ },
1282
+ {
1283
+ "epoch": 0.52,
1284
+ "learning_rate": 1.7978533094812166e-05,
1285
+ "loss": 0.872,
1286
+ "step": 213
1287
+ },
1288
+ {
1289
+ "epoch": 0.53,
1290
+ "learning_rate": 1.7960644007155635e-05,
1291
+ "loss": 0.8946,
1292
+ "step": 214
1293
+ },
1294
+ {
1295
+ "epoch": 0.53,
1296
+ "learning_rate": 1.7942754919499107e-05,
1297
+ "loss": 0.8784,
1298
+ "step": 215
1299
+ },
1300
+ {
1301
+ "epoch": 0.53,
1302
+ "learning_rate": 1.7924865831842576e-05,
1303
+ "loss": 0.8371,
1304
+ "step": 216
1305
+ },
1306
+ {
1307
+ "epoch": 0.53,
1308
+ "learning_rate": 1.790697674418605e-05,
1309
+ "loss": 0.8649,
1310
+ "step": 217
1311
+ },
1312
+ {
1313
+ "epoch": 0.54,
1314
+ "learning_rate": 1.7889087656529517e-05,
1315
+ "loss": 0.819,
1316
+ "step": 218
1317
+ },
1318
+ {
1319
+ "epoch": 0.54,
1320
+ "learning_rate": 1.787119856887299e-05,
1321
+ "loss": 0.8903,
1322
+ "step": 219
1323
+ },
1324
+ {
1325
+ "epoch": 0.54,
1326
+ "learning_rate": 1.7853309481216458e-05,
1327
+ "loss": 0.8675,
1328
+ "step": 220
1329
+ },
1330
+ {
1331
+ "epoch": 0.54,
1332
+ "learning_rate": 1.783542039355993e-05,
1333
+ "loss": 0.8728,
1334
+ "step": 221
1335
+ },
1336
+ {
1337
+ "epoch": 0.55,
1338
+ "learning_rate": 1.78175313059034e-05,
1339
+ "loss": 0.8753,
1340
+ "step": 222
1341
+ },
1342
+ {
1343
+ "epoch": 0.55,
1344
+ "learning_rate": 1.779964221824687e-05,
1345
+ "loss": 0.8156,
1346
+ "step": 223
1347
+ },
1348
+ {
1349
+ "epoch": 0.55,
1350
+ "learning_rate": 1.7781753130590344e-05,
1351
+ "loss": 0.8388,
1352
+ "step": 224
1353
+ },
1354
+ {
1355
+ "epoch": 0.55,
1356
+ "learning_rate": 1.7763864042933812e-05,
1357
+ "loss": 0.8928,
1358
+ "step": 225
1359
+ },
1360
+ {
1361
+ "epoch": 0.56,
1362
+ "learning_rate": 1.7745974955277285e-05,
1363
+ "loss": 0.8225,
1364
+ "step": 226
1365
+ },
1366
+ {
1367
+ "epoch": 0.56,
1368
+ "learning_rate": 1.7728085867620753e-05,
1369
+ "loss": 0.8366,
1370
+ "step": 227
1371
+ },
1372
+ {
1373
+ "epoch": 0.56,
1374
+ "learning_rate": 1.7710196779964222e-05,
1375
+ "loss": 0.8131,
1376
+ "step": 228
1377
+ },
1378
+ {
1379
+ "epoch": 0.56,
1380
+ "learning_rate": 1.7692307692307694e-05,
1381
+ "loss": 0.8717,
1382
+ "step": 229
1383
+ },
1384
+ {
1385
+ "epoch": 0.57,
1386
+ "learning_rate": 1.7674418604651163e-05,
1387
+ "loss": 0.8281,
1388
+ "step": 230
1389
+ },
1390
+ {
1391
+ "epoch": 0.57,
1392
+ "learning_rate": 1.7656529516994635e-05,
1393
+ "loss": 0.8203,
1394
+ "step": 231
1395
+ },
1396
+ {
1397
+ "epoch": 0.57,
1398
+ "learning_rate": 1.7638640429338104e-05,
1399
+ "loss": 0.8409,
1400
+ "step": 232
1401
+ },
1402
+ {
1403
+ "epoch": 0.57,
1404
+ "learning_rate": 1.7620751341681576e-05,
1405
+ "loss": 0.9037,
1406
+ "step": 233
1407
+ },
1408
+ {
1409
+ "epoch": 0.58,
1410
+ "learning_rate": 1.7602862254025045e-05,
1411
+ "loss": 0.9087,
1412
+ "step": 234
1413
+ },
1414
+ {
1415
+ "epoch": 0.58,
1416
+ "learning_rate": 1.7584973166368517e-05,
1417
+ "loss": 0.8616,
1418
+ "step": 235
1419
+ },
1420
+ {
1421
+ "epoch": 0.58,
1422
+ "learning_rate": 1.7567084078711986e-05,
1423
+ "loss": 0.8293,
1424
+ "step": 236
1425
+ },
1426
+ {
1427
+ "epoch": 0.58,
1428
+ "learning_rate": 1.754919499105546e-05,
1429
+ "loss": 0.9028,
1430
+ "step": 237
1431
+ },
1432
+ {
1433
+ "epoch": 0.59,
1434
+ "learning_rate": 1.7531305903398927e-05,
1435
+ "loss": 0.8728,
1436
+ "step": 238
1437
+ },
1438
+ {
1439
+ "epoch": 0.59,
1440
+ "learning_rate": 1.75134168157424e-05,
1441
+ "loss": 0.8617,
1442
+ "step": 239
1443
+ },
1444
+ {
1445
+ "epoch": 0.59,
1446
+ "learning_rate": 1.7495527728085868e-05,
1447
+ "loss": 0.8362,
1448
+ "step": 240
1449
+ },
1450
+ {
1451
+ "epoch": 0.59,
1452
+ "learning_rate": 1.747763864042934e-05,
1453
+ "loss": 0.9042,
1454
+ "step": 241
1455
+ },
1456
+ {
1457
+ "epoch": 0.6,
1458
+ "learning_rate": 1.745974955277281e-05,
1459
+ "loss": 0.8496,
1460
+ "step": 242
1461
+ },
1462
+ {
1463
+ "epoch": 0.6,
1464
+ "learning_rate": 1.744186046511628e-05,
1465
+ "loss": 0.8953,
1466
+ "step": 243
1467
+ },
1468
+ {
1469
+ "epoch": 0.6,
1470
+ "learning_rate": 1.742397137745975e-05,
1471
+ "loss": 0.8925,
1472
+ "step": 244
1473
+ },
1474
+ {
1475
+ "epoch": 0.6,
1476
+ "learning_rate": 1.7406082289803222e-05,
1477
+ "loss": 0.8732,
1478
+ "step": 245
1479
+ },
1480
+ {
1481
+ "epoch": 0.61,
1482
+ "learning_rate": 1.738819320214669e-05,
1483
+ "loss": 0.8479,
1484
+ "step": 246
1485
+ },
1486
+ {
1487
+ "epoch": 0.61,
1488
+ "learning_rate": 1.7370304114490163e-05,
1489
+ "loss": 0.841,
1490
+ "step": 247
1491
+ },
1492
+ {
1493
+ "epoch": 0.61,
1494
+ "learning_rate": 1.7352415026833632e-05,
1495
+ "loss": 0.8621,
1496
+ "step": 248
1497
+ },
1498
+ {
1499
+ "epoch": 0.61,
1500
+ "learning_rate": 1.73345259391771e-05,
1501
+ "loss": 0.8883,
1502
+ "step": 249
1503
+ },
1504
+ {
1505
+ "epoch": 0.62,
1506
+ "learning_rate": 1.7316636851520573e-05,
1507
+ "loss": 0.9141,
1508
+ "step": 250
1509
+ },
1510
+ {
1511
+ "epoch": 0.62,
1512
+ "learning_rate": 1.7298747763864042e-05,
1513
+ "loss": 0.8723,
1514
+ "step": 251
1515
+ },
1516
+ {
1517
+ "epoch": 0.62,
1518
+ "learning_rate": 1.7280858676207514e-05,
1519
+ "loss": 0.8421,
1520
+ "step": 252
1521
+ },
1522
+ {
1523
+ "epoch": 0.62,
1524
+ "learning_rate": 1.7262969588550986e-05,
1525
+ "loss": 0.81,
1526
+ "step": 253
1527
+ },
1528
+ {
1529
+ "epoch": 0.63,
1530
+ "learning_rate": 1.7245080500894455e-05,
1531
+ "loss": 0.8013,
1532
+ "step": 254
1533
+ },
1534
+ {
1535
+ "epoch": 0.63,
1536
+ "learning_rate": 1.7227191413237927e-05,
1537
+ "loss": 0.8855,
1538
+ "step": 255
1539
+ },
1540
+ {
1541
+ "epoch": 0.63,
1542
+ "learning_rate": 1.7209302325581396e-05,
1543
+ "loss": 0.8604,
1544
+ "step": 256
1545
+ },
1546
+ {
1547
+ "epoch": 0.63,
1548
+ "learning_rate": 1.719141323792487e-05,
1549
+ "loss": 0.8795,
1550
+ "step": 257
1551
+ },
1552
+ {
1553
+ "epoch": 0.64,
1554
+ "learning_rate": 1.7173524150268337e-05,
1555
+ "loss": 0.8324,
1556
+ "step": 258
1557
+ },
1558
+ {
1559
+ "epoch": 0.64,
1560
+ "learning_rate": 1.715563506261181e-05,
1561
+ "loss": 0.8615,
1562
+ "step": 259
1563
+ },
1564
+ {
1565
+ "epoch": 0.64,
1566
+ "learning_rate": 1.7137745974955278e-05,
1567
+ "loss": 0.8045,
1568
+ "step": 260
1569
+ },
1570
+ {
1571
+ "epoch": 0.64,
1572
+ "learning_rate": 1.711985688729875e-05,
1573
+ "loss": 0.8664,
1574
+ "step": 261
1575
+ },
1576
+ {
1577
+ "epoch": 0.64,
1578
+ "learning_rate": 1.710196779964222e-05,
1579
+ "loss": 0.848,
1580
+ "step": 262
1581
+ },
1582
+ {
1583
+ "epoch": 0.65,
1584
+ "learning_rate": 1.708407871198569e-05,
1585
+ "loss": 0.8489,
1586
+ "step": 263
1587
+ },
1588
+ {
1589
+ "epoch": 0.65,
1590
+ "learning_rate": 1.706618962432916e-05,
1591
+ "loss": 0.8425,
1592
+ "step": 264
1593
+ },
1594
+ {
1595
+ "epoch": 0.65,
1596
+ "learning_rate": 1.704830053667263e-05,
1597
+ "loss": 0.881,
1598
+ "step": 265
1599
+ },
1600
+ {
1601
+ "epoch": 0.65,
1602
+ "learning_rate": 1.70304114490161e-05,
1603
+ "loss": 0.877,
1604
+ "step": 266
1605
+ },
1606
+ {
1607
+ "epoch": 0.66,
1608
+ "learning_rate": 1.701252236135957e-05,
1609
+ "loss": 0.8441,
1610
+ "step": 267
1611
+ },
1612
+ {
1613
+ "epoch": 0.66,
1614
+ "learning_rate": 1.6994633273703042e-05,
1615
+ "loss": 0.8549,
1616
+ "step": 268
1617
+ },
1618
+ {
1619
+ "epoch": 0.66,
1620
+ "learning_rate": 1.697674418604651e-05,
1621
+ "loss": 0.8613,
1622
+ "step": 269
1623
+ },
1624
+ {
1625
+ "epoch": 0.66,
1626
+ "learning_rate": 1.6958855098389983e-05,
1627
+ "loss": 0.8643,
1628
+ "step": 270
1629
+ },
1630
+ {
1631
+ "epoch": 0.67,
1632
+ "learning_rate": 1.6940966010733455e-05,
1633
+ "loss": 0.87,
1634
+ "step": 271
1635
+ },
1636
+ {
1637
+ "epoch": 0.67,
1638
+ "learning_rate": 1.6923076923076924e-05,
1639
+ "loss": 0.9223,
1640
+ "step": 272
1641
+ },
1642
+ {
1643
+ "epoch": 0.67,
1644
+ "learning_rate": 1.6905187835420396e-05,
1645
+ "loss": 0.8748,
1646
+ "step": 273
1647
+ },
1648
+ {
1649
+ "epoch": 0.67,
1650
+ "learning_rate": 1.6887298747763865e-05,
1651
+ "loss": 0.8234,
1652
+ "step": 274
1653
+ },
1654
+ {
1655
+ "epoch": 0.68,
1656
+ "learning_rate": 1.6869409660107337e-05,
1657
+ "loss": 0.8259,
1658
+ "step": 275
1659
+ },
1660
+ {
1661
+ "epoch": 0.68,
1662
+ "learning_rate": 1.6851520572450806e-05,
1663
+ "loss": 0.8406,
1664
+ "step": 276
1665
+ },
1666
+ {
1667
+ "epoch": 0.68,
1668
+ "learning_rate": 1.683363148479428e-05,
1669
+ "loss": 0.8218,
1670
+ "step": 277
1671
+ },
1672
+ {
1673
+ "epoch": 0.68,
1674
+ "learning_rate": 1.6815742397137747e-05,
1675
+ "loss": 0.8564,
1676
+ "step": 278
1677
+ },
1678
+ {
1679
+ "epoch": 0.69,
1680
+ "learning_rate": 1.679785330948122e-05,
1681
+ "loss": 0.885,
1682
+ "step": 279
1683
+ },
1684
+ {
1685
+ "epoch": 0.69,
1686
+ "learning_rate": 1.6779964221824688e-05,
1687
+ "loss": 0.8218,
1688
+ "step": 280
1689
+ },
1690
+ {
1691
+ "epoch": 0.69,
1692
+ "learning_rate": 1.6762075134168157e-05,
1693
+ "loss": 0.8206,
1694
+ "step": 281
1695
+ },
1696
+ {
1697
+ "epoch": 0.69,
1698
+ "learning_rate": 1.674418604651163e-05,
1699
+ "loss": 0.8428,
1700
+ "step": 282
1701
+ },
1702
+ {
1703
+ "epoch": 0.7,
1704
+ "learning_rate": 1.6726296958855098e-05,
1705
+ "loss": 0.8816,
1706
+ "step": 283
1707
+ },
1708
+ {
1709
+ "epoch": 0.7,
1710
+ "learning_rate": 1.670840787119857e-05,
1711
+ "loss": 0.8458,
1712
+ "step": 284
1713
+ },
1714
+ {
1715
+ "epoch": 0.7,
1716
+ "learning_rate": 1.669051878354204e-05,
1717
+ "loss": 0.8412,
1718
+ "step": 285
1719
+ },
1720
+ {
1721
+ "epoch": 0.7,
1722
+ "learning_rate": 1.667262969588551e-05,
1723
+ "loss": 0.8458,
1724
+ "step": 286
1725
+ },
1726
+ {
1727
+ "epoch": 0.71,
1728
+ "learning_rate": 1.665474060822898e-05,
1729
+ "loss": 0.854,
1730
+ "step": 287
1731
+ },
1732
+ {
1733
+ "epoch": 0.71,
1734
+ "learning_rate": 1.6636851520572452e-05,
1735
+ "loss": 0.8831,
1736
+ "step": 288
1737
+ },
1738
+ {
1739
+ "epoch": 0.71,
1740
+ "learning_rate": 1.661896243291592e-05,
1741
+ "loss": 0.9012,
1742
+ "step": 289
1743
+ },
1744
+ {
1745
+ "epoch": 0.71,
1746
+ "learning_rate": 1.6601073345259393e-05,
1747
+ "loss": 0.8148,
1748
+ "step": 290
1749
+ },
1750
+ {
1751
+ "epoch": 0.72,
1752
+ "learning_rate": 1.6583184257602865e-05,
1753
+ "loss": 0.8257,
1754
+ "step": 291
1755
+ },
1756
+ {
1757
+ "epoch": 0.72,
1758
+ "learning_rate": 1.6565295169946334e-05,
1759
+ "loss": 0.8451,
1760
+ "step": 292
1761
+ },
1762
+ {
1763
+ "epoch": 0.72,
1764
+ "learning_rate": 1.6547406082289806e-05,
1765
+ "loss": 0.8355,
1766
+ "step": 293
1767
+ },
1768
+ {
1769
+ "epoch": 0.72,
1770
+ "learning_rate": 1.6529516994633275e-05,
1771
+ "loss": 0.8564,
1772
+ "step": 294
1773
+ },
1774
+ {
1775
+ "epoch": 0.73,
1776
+ "learning_rate": 1.6511627906976747e-05,
1777
+ "loss": 0.8922,
1778
+ "step": 295
1779
+ },
1780
+ {
1781
+ "epoch": 0.73,
1782
+ "learning_rate": 1.6493738819320216e-05,
1783
+ "loss": 0.87,
1784
+ "step": 296
1785
+ },
1786
+ {
1787
+ "epoch": 0.73,
1788
+ "learning_rate": 1.647584973166369e-05,
1789
+ "loss": 0.8307,
1790
+ "step": 297
1791
+ },
1792
+ {
1793
+ "epoch": 0.73,
1794
+ "learning_rate": 1.6457960644007157e-05,
1795
+ "loss": 0.8233,
1796
+ "step": 298
1797
+ },
1798
+ {
1799
+ "epoch": 0.74,
1800
+ "learning_rate": 1.6440071556350626e-05,
1801
+ "loss": 0.8436,
1802
+ "step": 299
1803
+ },
1804
+ {
1805
+ "epoch": 0.74,
1806
+ "learning_rate": 1.6422182468694098e-05,
1807
+ "loss": 0.8316,
1808
+ "step": 300
1809
+ },
1810
+ {
1811
+ "epoch": 0.74,
1812
+ "learning_rate": 1.6404293381037567e-05,
1813
+ "loss": 0.8468,
1814
+ "step": 301
1815
+ },
1816
+ {
1817
+ "epoch": 0.74,
1818
+ "learning_rate": 1.638640429338104e-05,
1819
+ "loss": 0.851,
1820
+ "step": 302
1821
+ },
1822
+ {
1823
+ "epoch": 0.75,
1824
+ "learning_rate": 1.6368515205724508e-05,
1825
+ "loss": 0.8277,
1826
+ "step": 303
1827
+ },
1828
+ {
1829
+ "epoch": 0.75,
1830
+ "learning_rate": 1.635062611806798e-05,
1831
+ "loss": 0.8296,
1832
+ "step": 304
1833
+ },
1834
+ {
1835
+ "epoch": 0.75,
1836
+ "learning_rate": 1.633273703041145e-05,
1837
+ "loss": 0.8571,
1838
+ "step": 305
1839
+ },
1840
+ {
1841
+ "epoch": 0.75,
1842
+ "learning_rate": 1.631484794275492e-05,
1843
+ "loss": 0.8691,
1844
+ "step": 306
1845
+ },
1846
+ {
1847
+ "epoch": 0.76,
1848
+ "learning_rate": 1.629695885509839e-05,
1849
+ "loss": 0.8477,
1850
+ "step": 307
1851
+ },
1852
+ {
1853
+ "epoch": 0.76,
1854
+ "learning_rate": 1.6279069767441862e-05,
1855
+ "loss": 0.8399,
1856
+ "step": 308
1857
+ },
1858
+ {
1859
+ "epoch": 0.76,
1860
+ "learning_rate": 1.6261180679785334e-05,
1861
+ "loss": 0.8243,
1862
+ "step": 309
1863
+ },
1864
+ {
1865
+ "epoch": 0.76,
1866
+ "learning_rate": 1.6243291592128803e-05,
1867
+ "loss": 0.8439,
1868
+ "step": 310
1869
+ },
1870
+ {
1871
+ "epoch": 0.77,
1872
+ "learning_rate": 1.6225402504472275e-05,
1873
+ "loss": 0.8175,
1874
+ "step": 311
1875
+ },
1876
+ {
1877
+ "epoch": 0.77,
1878
+ "learning_rate": 1.6207513416815744e-05,
1879
+ "loss": 0.8134,
1880
+ "step": 312
1881
+ },
1882
+ {
1883
+ "epoch": 0.77,
1884
+ "learning_rate": 1.6189624329159216e-05,
1885
+ "loss": 0.8575,
1886
+ "step": 313
1887
+ },
1888
+ {
1889
+ "epoch": 0.77,
1890
+ "learning_rate": 1.6171735241502685e-05,
1891
+ "loss": 0.8559,
1892
+ "step": 314
1893
+ },
1894
+ {
1895
+ "epoch": 0.78,
1896
+ "learning_rate": 1.6153846153846154e-05,
1897
+ "loss": 0.8374,
1898
+ "step": 315
1899
+ },
1900
+ {
1901
+ "epoch": 0.78,
1902
+ "learning_rate": 1.6135957066189626e-05,
1903
+ "loss": 0.7871,
1904
+ "step": 316
1905
+ },
1906
+ {
1907
+ "epoch": 0.78,
1908
+ "learning_rate": 1.6118067978533095e-05,
1909
+ "loss": 0.8712,
1910
+ "step": 317
1911
+ },
1912
+ {
1913
+ "epoch": 0.78,
1914
+ "learning_rate": 1.6100178890876567e-05,
1915
+ "loss": 0.8224,
1916
+ "step": 318
1917
+ },
1918
+ {
1919
+ "epoch": 0.79,
1920
+ "learning_rate": 1.6082289803220036e-05,
1921
+ "loss": 0.8406,
1922
+ "step": 319
1923
+ },
1924
+ {
1925
+ "epoch": 0.79,
1926
+ "learning_rate": 1.6064400715563508e-05,
1927
+ "loss": 0.8155,
1928
+ "step": 320
1929
+ },
1930
+ {
1931
+ "epoch": 0.79,
1932
+ "learning_rate": 1.6046511627906977e-05,
1933
+ "loss": 0.8617,
1934
+ "step": 321
1935
+ },
1936
+ {
1937
+ "epoch": 0.79,
1938
+ "learning_rate": 1.602862254025045e-05,
1939
+ "loss": 0.8328,
1940
+ "step": 322
1941
+ },
1942
+ {
1943
+ "epoch": 0.8,
1944
+ "learning_rate": 1.6010733452593918e-05,
1945
+ "loss": 0.829,
1946
+ "step": 323
1947
+ },
1948
+ {
1949
+ "epoch": 0.8,
1950
+ "learning_rate": 1.599284436493739e-05,
1951
+ "loss": 0.8293,
1952
+ "step": 324
1953
+ },
1954
+ {
1955
+ "epoch": 0.8,
1956
+ "learning_rate": 1.597495527728086e-05,
1957
+ "loss": 0.8227,
1958
+ "step": 325
1959
+ },
1960
+ {
1961
+ "epoch": 0.8,
1962
+ "learning_rate": 1.595706618962433e-05,
1963
+ "loss": 0.8653,
1964
+ "step": 326
1965
+ },
1966
+ {
1967
+ "epoch": 0.8,
1968
+ "learning_rate": 1.59391771019678e-05,
1969
+ "loss": 0.8151,
1970
+ "step": 327
1971
+ },
1972
+ {
1973
+ "epoch": 0.81,
1974
+ "learning_rate": 1.5921288014311272e-05,
1975
+ "loss": 0.7727,
1976
+ "step": 328
1977
+ },
1978
+ {
1979
+ "epoch": 0.81,
1980
+ "learning_rate": 1.5903398926654744e-05,
1981
+ "loss": 0.862,
1982
+ "step": 329
1983
+ },
1984
+ {
1985
+ "epoch": 0.81,
1986
+ "learning_rate": 1.5885509838998213e-05,
1987
+ "loss": 0.7711,
1988
+ "step": 330
1989
+ },
1990
+ {
1991
+ "epoch": 0.81,
1992
+ "learning_rate": 1.5867620751341682e-05,
1993
+ "loss": 0.7793,
1994
+ "step": 331
1995
+ },
1996
+ {
1997
+ "epoch": 0.82,
1998
+ "learning_rate": 1.5849731663685154e-05,
1999
+ "loss": 0.8516,
2000
+ "step": 332
2001
+ },
2002
+ {
2003
+ "epoch": 0.82,
2004
+ "learning_rate": 1.5831842576028623e-05,
2005
+ "loss": 0.8307,
2006
+ "step": 333
2007
+ },
2008
+ {
2009
+ "epoch": 0.82,
2010
+ "learning_rate": 1.5813953488372095e-05,
2011
+ "loss": 0.8396,
2012
+ "step": 334
2013
+ },
2014
+ {
2015
+ "epoch": 0.82,
2016
+ "learning_rate": 1.5796064400715564e-05,
2017
+ "loss": 0.8286,
2018
+ "step": 335
2019
+ },
2020
+ {
2021
+ "epoch": 0.83,
2022
+ "learning_rate": 1.5778175313059033e-05,
2023
+ "loss": 0.8009,
2024
+ "step": 336
2025
+ },
2026
+ {
2027
+ "epoch": 0.83,
2028
+ "learning_rate": 1.5760286225402505e-05,
2029
+ "loss": 0.8019,
2030
+ "step": 337
2031
+ },
2032
+ {
2033
+ "epoch": 0.83,
2034
+ "learning_rate": 1.5742397137745977e-05,
2035
+ "loss": 0.8213,
2036
+ "step": 338
2037
+ },
2038
+ {
2039
+ "epoch": 0.83,
2040
+ "learning_rate": 1.5724508050089446e-05,
2041
+ "loss": 0.8219,
2042
+ "step": 339
2043
+ },
2044
+ {
2045
+ "epoch": 0.84,
2046
+ "learning_rate": 1.5706618962432918e-05,
2047
+ "loss": 0.8349,
2048
+ "step": 340
2049
+ },
2050
+ {
2051
+ "epoch": 0.84,
2052
+ "learning_rate": 1.5688729874776387e-05,
2053
+ "loss": 0.8143,
2054
+ "step": 341
2055
+ },
2056
+ {
2057
+ "epoch": 0.84,
2058
+ "learning_rate": 1.567084078711986e-05,
2059
+ "loss": 0.8195,
2060
+ "step": 342
2061
+ },
2062
+ {
2063
+ "epoch": 0.84,
2064
+ "learning_rate": 1.5652951699463328e-05,
2065
+ "loss": 0.7934,
2066
+ "step": 343
2067
+ },
2068
+ {
2069
+ "epoch": 0.85,
2070
+ "learning_rate": 1.56350626118068e-05,
2071
+ "loss": 0.8411,
2072
+ "step": 344
2073
+ },
2074
+ {
2075
+ "epoch": 0.85,
2076
+ "learning_rate": 1.561717352415027e-05,
2077
+ "loss": 0.8399,
2078
+ "step": 345
2079
+ },
2080
+ {
2081
+ "epoch": 0.85,
2082
+ "learning_rate": 1.559928443649374e-05,
2083
+ "loss": 0.8198,
2084
+ "step": 346
2085
+ },
2086
+ {
2087
+ "epoch": 0.85,
2088
+ "learning_rate": 1.558139534883721e-05,
2089
+ "loss": 0.8247,
2090
+ "step": 347
2091
+ },
2092
+ {
2093
+ "epoch": 0.86,
2094
+ "learning_rate": 1.5563506261180682e-05,
2095
+ "loss": 0.8273,
2096
+ "step": 348
2097
+ },
2098
+ {
2099
+ "epoch": 0.86,
2100
+ "learning_rate": 1.554561717352415e-05,
2101
+ "loss": 0.8559,
2102
+ "step": 349
2103
+ },
2104
+ {
2105
+ "epoch": 0.86,
2106
+ "learning_rate": 1.5527728085867623e-05,
2107
+ "loss": 0.8711,
2108
+ "step": 350
2109
+ },
2110
+ {
2111
+ "epoch": 0.86,
2112
+ "learning_rate": 1.5509838998211092e-05,
2113
+ "loss": 0.7908,
2114
+ "step": 351
2115
+ },
2116
+ {
2117
+ "epoch": 0.87,
2118
+ "learning_rate": 1.549194991055456e-05,
2119
+ "loss": 0.77,
2120
+ "step": 352
2121
+ },
2122
+ {
2123
+ "epoch": 0.87,
2124
+ "learning_rate": 1.5474060822898033e-05,
2125
+ "loss": 0.8534,
2126
+ "step": 353
2127
+ },
2128
+ {
2129
+ "epoch": 0.87,
2130
+ "learning_rate": 1.5456171735241502e-05,
2131
+ "loss": 0.8307,
2132
+ "step": 354
2133
+ },
2134
+ {
2135
+ "epoch": 0.87,
2136
+ "learning_rate": 1.5438282647584974e-05,
2137
+ "loss": 0.833,
2138
+ "step": 355
2139
+ },
2140
+ {
2141
+ "epoch": 0.88,
2142
+ "learning_rate": 1.5420393559928446e-05,
2143
+ "loss": 0.8106,
2144
+ "step": 356
2145
+ },
2146
+ {
2147
+ "epoch": 0.88,
2148
+ "learning_rate": 1.5402504472271915e-05,
2149
+ "loss": 0.8492,
2150
+ "step": 357
2151
+ },
2152
+ {
2153
+ "epoch": 0.88,
2154
+ "learning_rate": 1.5384615384615387e-05,
2155
+ "loss": 0.8324,
2156
+ "step": 358
2157
+ },
2158
+ {
2159
+ "epoch": 0.88,
2160
+ "learning_rate": 1.5366726296958856e-05,
2161
+ "loss": 0.7787,
2162
+ "step": 359
2163
+ },
2164
+ {
2165
+ "epoch": 0.89,
2166
+ "learning_rate": 1.5348837209302328e-05,
2167
+ "loss": 0.8056,
2168
+ "step": 360
2169
+ },
2170
+ {
2171
+ "epoch": 0.89,
2172
+ "learning_rate": 1.5330948121645797e-05,
2173
+ "loss": 0.8535,
2174
+ "step": 361
2175
+ },
2176
+ {
2177
+ "epoch": 0.89,
2178
+ "learning_rate": 1.531305903398927e-05,
2179
+ "loss": 0.8351,
2180
+ "step": 362
2181
+ },
2182
+ {
2183
+ "epoch": 0.89,
2184
+ "learning_rate": 1.5295169946332738e-05,
2185
+ "loss": 0.8457,
2186
+ "step": 363
2187
+ },
2188
+ {
2189
+ "epoch": 0.9,
2190
+ "learning_rate": 1.527728085867621e-05,
2191
+ "loss": 0.8336,
2192
+ "step": 364
2193
+ },
2194
+ {
2195
+ "epoch": 0.9,
2196
+ "learning_rate": 1.525939177101968e-05,
2197
+ "loss": 0.8459,
2198
+ "step": 365
2199
+ },
2200
+ {
2201
+ "epoch": 0.9,
2202
+ "learning_rate": 1.524150268336315e-05,
2203
+ "loss": 0.8714,
2204
+ "step": 366
2205
+ },
2206
+ {
2207
+ "epoch": 0.9,
2208
+ "learning_rate": 1.5223613595706622e-05,
2209
+ "loss": 0.8258,
2210
+ "step": 367
2211
+ },
2212
+ {
2213
+ "epoch": 0.91,
2214
+ "learning_rate": 1.520572450805009e-05,
2215
+ "loss": 0.9213,
2216
+ "step": 368
2217
+ },
2218
+ {
2219
+ "epoch": 0.91,
2220
+ "learning_rate": 1.5187835420393563e-05,
2221
+ "loss": 0.8341,
2222
+ "step": 369
2223
+ },
2224
+ {
2225
+ "epoch": 0.91,
2226
+ "learning_rate": 1.5169946332737031e-05,
2227
+ "loss": 0.8022,
2228
+ "step": 370
2229
+ },
2230
+ {
2231
+ "epoch": 0.91,
2232
+ "learning_rate": 1.5152057245080502e-05,
2233
+ "loss": 0.7943,
2234
+ "step": 371
2235
+ },
2236
+ {
2237
+ "epoch": 0.92,
2238
+ "learning_rate": 1.5134168157423972e-05,
2239
+ "loss": 0.852,
2240
+ "step": 372
2241
+ },
2242
+ {
2243
+ "epoch": 0.92,
2244
+ "learning_rate": 1.5116279069767443e-05,
2245
+ "loss": 0.8684,
2246
+ "step": 373
2247
+ },
2248
+ {
2249
+ "epoch": 0.92,
2250
+ "learning_rate": 1.5098389982110912e-05,
2251
+ "loss": 0.8951,
2252
+ "step": 374
2253
+ },
2254
+ {
2255
+ "epoch": 0.92,
2256
+ "learning_rate": 1.5080500894454384e-05,
2257
+ "loss": 0.8065,
2258
+ "step": 375
2259
+ },
2260
+ {
2261
+ "epoch": 0.93,
2262
+ "learning_rate": 1.5062611806797856e-05,
2263
+ "loss": 0.8308,
2264
+ "step": 376
2265
+ },
2266
+ {
2267
+ "epoch": 0.93,
2268
+ "learning_rate": 1.5044722719141325e-05,
2269
+ "loss": 0.8609,
2270
+ "step": 377
2271
+ },
2272
+ {
2273
+ "epoch": 0.93,
2274
+ "learning_rate": 1.5026833631484795e-05,
2275
+ "loss": 0.8266,
2276
+ "step": 378
2277
+ },
2278
+ {
2279
+ "epoch": 0.93,
2280
+ "learning_rate": 1.5008944543828266e-05,
2281
+ "loss": 0.8265,
2282
+ "step": 379
2283
+ },
2284
+ {
2285
+ "epoch": 0.94,
2286
+ "learning_rate": 1.4991055456171736e-05,
2287
+ "loss": 0.8416,
2288
+ "step": 380
2289
+ },
2290
+ {
2291
+ "epoch": 0.94,
2292
+ "learning_rate": 1.4973166368515207e-05,
2293
+ "loss": 0.8167,
2294
+ "step": 381
2295
+ },
2296
+ {
2297
+ "epoch": 0.94,
2298
+ "learning_rate": 1.4955277280858677e-05,
2299
+ "loss": 0.8447,
2300
+ "step": 382
2301
+ },
2302
+ {
2303
+ "epoch": 0.94,
2304
+ "learning_rate": 1.4937388193202146e-05,
2305
+ "loss": 0.8328,
2306
+ "step": 383
2307
+ },
2308
+ {
2309
+ "epoch": 0.95,
2310
+ "learning_rate": 1.4919499105545618e-05,
2311
+ "loss": 0.8347,
2312
+ "step": 384
2313
+ },
2314
+ {
2315
+ "epoch": 0.95,
2316
+ "learning_rate": 1.490161001788909e-05,
2317
+ "loss": 0.8552,
2318
+ "step": 385
2319
+ },
2320
+ {
2321
+ "epoch": 0.95,
2322
+ "learning_rate": 1.488372093023256e-05,
2323
+ "loss": 0.8543,
2324
+ "step": 386
2325
+ },
2326
+ {
2327
+ "epoch": 0.95,
2328
+ "learning_rate": 1.486583184257603e-05,
2329
+ "loss": 0.8652,
2330
+ "step": 387
2331
+ },
2332
+ {
2333
+ "epoch": 0.96,
2334
+ "learning_rate": 1.48479427549195e-05,
2335
+ "loss": 0.8058,
2336
+ "step": 388
2337
+ },
2338
+ {
2339
+ "epoch": 0.96,
2340
+ "learning_rate": 1.4830053667262971e-05,
2341
+ "loss": 0.8092,
2342
+ "step": 389
2343
+ },
2344
+ {
2345
+ "epoch": 0.96,
2346
+ "learning_rate": 1.481216457960644e-05,
2347
+ "loss": 0.8588,
2348
+ "step": 390
2349
+ },
2350
+ {
2351
+ "epoch": 0.96,
2352
+ "learning_rate": 1.4794275491949912e-05,
2353
+ "loss": 0.8591,
2354
+ "step": 391
2355
+ },
2356
+ {
2357
+ "epoch": 0.96,
2358
+ "learning_rate": 1.477638640429338e-05,
2359
+ "loss": 0.8499,
2360
+ "step": 392
2361
+ },
2362
+ {
2363
+ "epoch": 0.97,
2364
+ "learning_rate": 1.4758497316636853e-05,
2365
+ "loss": 0.8223,
2366
+ "step": 393
2367
+ },
2368
+ {
2369
+ "epoch": 0.97,
2370
+ "learning_rate": 1.4740608228980323e-05,
2371
+ "loss": 0.8023,
2372
+ "step": 394
2373
+ },
2374
+ {
2375
+ "epoch": 0.97,
2376
+ "learning_rate": 1.4722719141323794e-05,
2377
+ "loss": 0.8077,
2378
+ "step": 395
2379
+ },
2380
+ {
2381
+ "epoch": 0.97,
2382
+ "learning_rate": 1.4704830053667264e-05,
2383
+ "loss": 0.8578,
2384
+ "step": 396
2385
+ },
2386
+ {
2387
+ "epoch": 0.98,
2388
+ "learning_rate": 1.4686940966010735e-05,
2389
+ "loss": 0.8287,
2390
+ "step": 397
2391
+ },
2392
+ {
2393
+ "epoch": 0.98,
2394
+ "learning_rate": 1.4669051878354205e-05,
2395
+ "loss": 0.8481,
2396
+ "step": 398
2397
+ },
2398
+ {
2399
+ "epoch": 0.98,
2400
+ "learning_rate": 1.4651162790697674e-05,
2401
+ "loss": 0.8197,
2402
+ "step": 399
2403
+ },
2404
+ {
2405
+ "epoch": 0.98,
2406
+ "learning_rate": 1.4633273703041146e-05,
2407
+ "loss": 0.811,
2408
+ "step": 400
2409
+ },
2410
+ {
2411
+ "epoch": 0.99,
2412
+ "learning_rate": 1.4615384615384615e-05,
2413
+ "loss": 0.7848,
2414
+ "step": 401
2415
+ },
2416
+ {
2417
+ "epoch": 0.99,
2418
+ "learning_rate": 1.4597495527728087e-05,
2419
+ "loss": 0.7713,
2420
+ "step": 402
2421
+ },
2422
+ {
2423
+ "epoch": 0.99,
2424
+ "learning_rate": 1.4579606440071556e-05,
2425
+ "loss": 0.8476,
2426
+ "step": 403
2427
+ },
2428
+ {
2429
+ "epoch": 0.99,
2430
+ "learning_rate": 1.4561717352415028e-05,
2431
+ "loss": 0.829,
2432
+ "step": 404
2433
+ },
2434
+ {
2435
+ "epoch": 1.0,
2436
+ "learning_rate": 1.4543828264758499e-05,
2437
+ "loss": 0.8703,
2438
+ "step": 405
2439
+ },
2440
+ {
2441
+ "epoch": 1.0,
2442
+ "learning_rate": 1.452593917710197e-05,
2443
+ "loss": 0.8499,
2444
+ "step": 406
2445
+ },
2446
+ {
2447
+ "epoch": 1.0,
2448
+ "learning_rate": 1.450805008944544e-05,
2449
+ "loss": 1.0286,
2450
+ "step": 407
2451
+ },
2452
+ {
2453
+ "epoch": 1.0,
2454
+ "learning_rate": 1.4490161001788909e-05,
2455
+ "loss": 0.817,
2456
+ "step": 408
2457
+ },
2458
+ {
2459
+ "epoch": 1.01,
2460
+ "learning_rate": 1.4472271914132381e-05,
2461
+ "loss": 0.8247,
2462
+ "step": 409
2463
+ },
2464
+ {
2465
+ "epoch": 1.01,
2466
+ "learning_rate": 1.445438282647585e-05,
2467
+ "loss": 0.8081,
2468
+ "step": 410
2469
+ },
2470
+ {
2471
+ "epoch": 1.01,
2472
+ "learning_rate": 1.4436493738819322e-05,
2473
+ "loss": 0.8309,
2474
+ "step": 411
2475
+ },
2476
+ {
2477
+ "epoch": 1.01,
2478
+ "learning_rate": 1.441860465116279e-05,
2479
+ "loss": 0.8615,
2480
+ "step": 412
2481
+ },
2482
+ {
2483
+ "epoch": 1.02,
2484
+ "learning_rate": 1.4400715563506263e-05,
2485
+ "loss": 0.8196,
2486
+ "step": 413
2487
+ },
2488
+ {
2489
+ "epoch": 1.02,
2490
+ "learning_rate": 1.4382826475849733e-05,
2491
+ "loss": 0.8586,
2492
+ "step": 414
2493
+ },
2494
+ {
2495
+ "epoch": 1.02,
2496
+ "learning_rate": 1.4364937388193202e-05,
2497
+ "loss": 0.8293,
2498
+ "step": 415
2499
+ },
2500
+ {
2501
+ "epoch": 1.02,
2502
+ "learning_rate": 1.4347048300536674e-05,
2503
+ "loss": 0.8287,
2504
+ "step": 416
2505
+ },
2506
+ {
2507
+ "epoch": 1.03,
2508
+ "learning_rate": 1.4329159212880143e-05,
2509
+ "loss": 0.8315,
2510
+ "step": 417
2511
+ },
2512
+ {
2513
+ "epoch": 1.03,
2514
+ "learning_rate": 1.4311270125223615e-05,
2515
+ "loss": 0.8359,
2516
+ "step": 418
2517
+ },
2518
+ {
2519
+ "epoch": 1.03,
2520
+ "learning_rate": 1.4293381037567084e-05,
2521
+ "loss": 0.8281,
2522
+ "step": 419
2523
+ },
2524
+ {
2525
+ "epoch": 1.03,
2526
+ "learning_rate": 1.4275491949910556e-05,
2527
+ "loss": 0.8205,
2528
+ "step": 420
2529
+ },
2530
+ {
2531
+ "epoch": 1.04,
2532
+ "learning_rate": 1.4257602862254025e-05,
2533
+ "loss": 0.8177,
2534
+ "step": 421
2535
+ },
2536
+ {
2537
+ "epoch": 1.04,
2538
+ "learning_rate": 1.4239713774597497e-05,
2539
+ "loss": 0.8261,
2540
+ "step": 422
2541
+ },
2542
+ {
2543
+ "epoch": 1.04,
2544
+ "learning_rate": 1.4221824686940968e-05,
2545
+ "loss": 0.8496,
2546
+ "step": 423
2547
+ },
2548
+ {
2549
+ "epoch": 1.04,
2550
+ "learning_rate": 1.4203935599284437e-05,
2551
+ "loss": 0.8056,
2552
+ "step": 424
2553
+ },
2554
+ {
2555
+ "epoch": 1.05,
2556
+ "learning_rate": 1.4186046511627909e-05,
2557
+ "loss": 0.877,
2558
+ "step": 425
2559
+ },
2560
+ {
2561
+ "epoch": 1.05,
2562
+ "learning_rate": 1.4168157423971378e-05,
2563
+ "loss": 0.8444,
2564
+ "step": 426
2565
+ },
2566
+ {
2567
+ "epoch": 1.05,
2568
+ "learning_rate": 1.415026833631485e-05,
2569
+ "loss": 0.8212,
2570
+ "step": 427
2571
+ },
2572
+ {
2573
+ "epoch": 1.05,
2574
+ "learning_rate": 1.4132379248658319e-05,
2575
+ "loss": 0.8268,
2576
+ "step": 428
2577
+ },
2578
+ {
2579
+ "epoch": 1.06,
2580
+ "learning_rate": 1.4114490161001791e-05,
2581
+ "loss": 0.8419,
2582
+ "step": 429
2583
+ },
2584
+ {
2585
+ "epoch": 1.06,
2586
+ "learning_rate": 1.409660107334526e-05,
2587
+ "loss": 0.863,
2588
+ "step": 430
2589
+ },
2590
+ {
2591
+ "epoch": 1.06,
2592
+ "learning_rate": 1.407871198568873e-05,
2593
+ "loss": 0.8222,
2594
+ "step": 431
2595
+ },
2596
+ {
2597
+ "epoch": 1.06,
2598
+ "learning_rate": 1.4060822898032202e-05,
2599
+ "loss": 0.8259,
2600
+ "step": 432
2601
+ },
2602
+ {
2603
+ "epoch": 1.07,
2604
+ "learning_rate": 1.4042933810375671e-05,
2605
+ "loss": 0.7908,
2606
+ "step": 433
2607
+ },
2608
+ {
2609
+ "epoch": 1.07,
2610
+ "learning_rate": 1.4025044722719143e-05,
2611
+ "loss": 0.8292,
2612
+ "step": 434
2613
+ },
2614
+ {
2615
+ "epoch": 1.07,
2616
+ "learning_rate": 1.4007155635062612e-05,
2617
+ "loss": 0.8414,
2618
+ "step": 435
2619
+ },
2620
+ {
2621
+ "epoch": 1.07,
2622
+ "learning_rate": 1.3989266547406084e-05,
2623
+ "loss": 0.8316,
2624
+ "step": 436
2625
+ },
2626
+ {
2627
+ "epoch": 1.08,
2628
+ "learning_rate": 1.3971377459749553e-05,
2629
+ "loss": 0.8278,
2630
+ "step": 437
2631
+ },
2632
+ {
2633
+ "epoch": 1.08,
2634
+ "learning_rate": 1.3953488372093025e-05,
2635
+ "loss": 0.7725,
2636
+ "step": 438
2637
+ },
2638
+ {
2639
+ "epoch": 1.08,
2640
+ "learning_rate": 1.3935599284436494e-05,
2641
+ "loss": 0.8013,
2642
+ "step": 439
2643
+ },
2644
+ {
2645
+ "epoch": 1.08,
2646
+ "learning_rate": 1.3917710196779965e-05,
2647
+ "loss": 0.8475,
2648
+ "step": 440
2649
+ },
2650
+ {
2651
+ "epoch": 1.09,
2652
+ "learning_rate": 1.3899821109123435e-05,
2653
+ "loss": 0.8314,
2654
+ "step": 441
2655
+ },
2656
+ {
2657
+ "epoch": 1.09,
2658
+ "learning_rate": 1.3881932021466906e-05,
2659
+ "loss": 0.8102,
2660
+ "step": 442
2661
+ },
2662
+ {
2663
+ "epoch": 1.09,
2664
+ "learning_rate": 1.3864042933810378e-05,
2665
+ "loss": 0.8208,
2666
+ "step": 443
2667
+ },
2668
+ {
2669
+ "epoch": 1.09,
2670
+ "learning_rate": 1.3846153846153847e-05,
2671
+ "loss": 0.8327,
2672
+ "step": 444
2673
+ },
2674
+ {
2675
+ "epoch": 1.1,
2676
+ "learning_rate": 1.3828264758497319e-05,
2677
+ "loss": 0.8524,
2678
+ "step": 445
2679
+ },
2680
+ {
2681
+ "epoch": 1.1,
2682
+ "learning_rate": 1.3810375670840788e-05,
2683
+ "loss": 0.8201,
2684
+ "step": 446
2685
+ },
2686
+ {
2687
+ "epoch": 1.1,
2688
+ "learning_rate": 1.379248658318426e-05,
2689
+ "loss": 0.8399,
2690
+ "step": 447
2691
+ },
2692
+ {
2693
+ "epoch": 1.1,
2694
+ "learning_rate": 1.3774597495527729e-05,
2695
+ "loss": 0.8894,
2696
+ "step": 448
2697
+ },
2698
+ {
2699
+ "epoch": 1.11,
2700
+ "learning_rate": 1.37567084078712e-05,
2701
+ "loss": 0.8275,
2702
+ "step": 449
2703
+ },
2704
+ {
2705
+ "epoch": 1.11,
2706
+ "learning_rate": 1.373881932021467e-05,
2707
+ "loss": 0.8194,
2708
+ "step": 450
2709
+ },
2710
+ {
2711
+ "epoch": 1.11,
2712
+ "learning_rate": 1.372093023255814e-05,
2713
+ "loss": 0.7857,
2714
+ "step": 451
2715
+ },
2716
+ {
2717
+ "epoch": 1.11,
2718
+ "learning_rate": 1.3703041144901612e-05,
2719
+ "loss": 0.7948,
2720
+ "step": 452
2721
+ },
2722
+ {
2723
+ "epoch": 1.12,
2724
+ "learning_rate": 1.3685152057245081e-05,
2725
+ "loss": 0.832,
2726
+ "step": 453
2727
+ },
2728
+ {
2729
+ "epoch": 1.12,
2730
+ "learning_rate": 1.3667262969588553e-05,
2731
+ "loss": 0.7991,
2732
+ "step": 454
2733
+ },
2734
+ {
2735
+ "epoch": 1.12,
2736
+ "learning_rate": 1.3649373881932022e-05,
2737
+ "loss": 0.8105,
2738
+ "step": 455
2739
+ },
2740
+ {
2741
+ "epoch": 1.12,
2742
+ "learning_rate": 1.3631484794275493e-05,
2743
+ "loss": 0.8577,
2744
+ "step": 456
2745
+ },
2746
+ {
2747
+ "epoch": 1.13,
2748
+ "learning_rate": 1.3613595706618963e-05,
2749
+ "loss": 0.7955,
2750
+ "step": 457
2751
+ },
2752
+ {
2753
+ "epoch": 1.13,
2754
+ "learning_rate": 1.3595706618962434e-05,
2755
+ "loss": 0.827,
2756
+ "step": 458
2757
+ },
2758
+ {
2759
+ "epoch": 1.13,
2760
+ "learning_rate": 1.3577817531305904e-05,
2761
+ "loss": 0.8027,
2762
+ "step": 459
2763
+ },
2764
+ {
2765
+ "epoch": 1.13,
2766
+ "learning_rate": 1.3559928443649375e-05,
2767
+ "loss": 0.8225,
2768
+ "step": 460
2769
+ },
2770
+ {
2771
+ "epoch": 1.14,
2772
+ "learning_rate": 1.3542039355992847e-05,
2773
+ "loss": 0.8167,
2774
+ "step": 461
2775
+ },
2776
+ {
2777
+ "epoch": 1.14,
2778
+ "learning_rate": 1.3524150268336316e-05,
2779
+ "loss": 0.8369,
2780
+ "step": 462
2781
+ },
2782
+ {
2783
+ "epoch": 1.14,
2784
+ "learning_rate": 1.3506261180679788e-05,
2785
+ "loss": 0.7537,
2786
+ "step": 463
2787
+ },
2788
+ {
2789
+ "epoch": 1.14,
2790
+ "learning_rate": 1.3488372093023257e-05,
2791
+ "loss": 0.817,
2792
+ "step": 464
2793
+ },
2794
+ {
2795
+ "epoch": 1.15,
2796
+ "learning_rate": 1.3470483005366727e-05,
2797
+ "loss": 0.7976,
2798
+ "step": 465
2799
+ },
2800
+ {
2801
+ "epoch": 1.15,
2802
+ "learning_rate": 1.3452593917710198e-05,
2803
+ "loss": 0.8305,
2804
+ "step": 466
2805
+ },
2806
+ {
2807
+ "epoch": 1.15,
2808
+ "learning_rate": 1.3434704830053668e-05,
2809
+ "loss": 0.7939,
2810
+ "step": 467
2811
+ },
2812
+ {
2813
+ "epoch": 1.15,
2814
+ "learning_rate": 1.3416815742397137e-05,
2815
+ "loss": 0.8733,
2816
+ "step": 468
2817
+ },
2818
+ {
2819
+ "epoch": 1.16,
2820
+ "learning_rate": 1.339892665474061e-05,
2821
+ "loss": 0.8397,
2822
+ "step": 469
2823
+ },
2824
+ {
2825
+ "epoch": 1.16,
2826
+ "learning_rate": 1.3381037567084081e-05,
2827
+ "loss": 0.7967,
2828
+ "step": 470
2829
+ },
2830
+ {
2831
+ "epoch": 1.16,
2832
+ "learning_rate": 1.336314847942755e-05,
2833
+ "loss": 0.8074,
2834
+ "step": 471
2835
+ },
2836
+ {
2837
+ "epoch": 1.16,
2838
+ "learning_rate": 1.334525939177102e-05,
2839
+ "loss": 0.8251,
2840
+ "step": 472
2841
+ },
2842
+ {
2843
+ "epoch": 1.16,
2844
+ "learning_rate": 1.3327370304114491e-05,
2845
+ "loss": 0.8205,
2846
+ "step": 473
2847
+ },
2848
+ {
2849
+ "epoch": 1.17,
2850
+ "learning_rate": 1.3309481216457962e-05,
2851
+ "loss": 0.8295,
2852
+ "step": 474
2853
+ },
2854
+ {
2855
+ "epoch": 1.17,
2856
+ "learning_rate": 1.3291592128801432e-05,
2857
+ "loss": 0.7885,
2858
+ "step": 475
2859
+ },
2860
+ {
2861
+ "epoch": 1.17,
2862
+ "learning_rate": 1.3273703041144903e-05,
2863
+ "loss": 0.7987,
2864
+ "step": 476
2865
+ },
2866
+ {
2867
+ "epoch": 1.17,
2868
+ "learning_rate": 1.3255813953488372e-05,
2869
+ "loss": 0.7914,
2870
+ "step": 477
2871
+ },
2872
+ {
2873
+ "epoch": 1.18,
2874
+ "learning_rate": 1.3237924865831844e-05,
2875
+ "loss": 0.8191,
2876
+ "step": 478
2877
+ },
2878
+ {
2879
+ "epoch": 1.18,
2880
+ "learning_rate": 1.3220035778175316e-05,
2881
+ "loss": 0.8328,
2882
+ "step": 479
2883
+ },
2884
+ {
2885
+ "epoch": 1.18,
2886
+ "learning_rate": 1.3202146690518785e-05,
2887
+ "loss": 0.7961,
2888
+ "step": 480
2889
+ },
2890
+ {
2891
+ "epoch": 1.18,
2892
+ "learning_rate": 1.3184257602862255e-05,
2893
+ "loss": 0.8636,
2894
+ "step": 481
2895
+ },
2896
+ {
2897
+ "epoch": 1.19,
2898
+ "learning_rate": 1.3166368515205726e-05,
2899
+ "loss": 0.8132,
2900
+ "step": 482
2901
+ },
2902
+ {
2903
+ "epoch": 1.19,
2904
+ "learning_rate": 1.3148479427549196e-05,
2905
+ "loss": 0.7955,
2906
+ "step": 483
2907
+ },
2908
+ {
2909
+ "epoch": 1.19,
2910
+ "learning_rate": 1.3130590339892667e-05,
2911
+ "loss": 0.7777,
2912
+ "step": 484
2913
+ },
2914
+ {
2915
+ "epoch": 1.19,
2916
+ "learning_rate": 1.3112701252236137e-05,
2917
+ "loss": 0.797,
2918
+ "step": 485
2919
+ },
2920
+ {
2921
+ "epoch": 1.2,
2922
+ "learning_rate": 1.3094812164579606e-05,
2923
+ "loss": 0.7756,
2924
+ "step": 486
2925
+ },
2926
+ {
2927
+ "epoch": 1.2,
2928
+ "learning_rate": 1.3076923076923078e-05,
2929
+ "loss": 0.7983,
2930
+ "step": 487
2931
+ },
2932
+ {
2933
+ "epoch": 1.2,
2934
+ "learning_rate": 1.3059033989266547e-05,
2935
+ "loss": 0.8013,
2936
+ "step": 488
2937
+ },
2938
+ {
2939
+ "epoch": 1.2,
2940
+ "learning_rate": 1.304114490161002e-05,
2941
+ "loss": 0.8122,
2942
+ "step": 489
2943
+ },
2944
+ {
2945
+ "epoch": 1.21,
2946
+ "learning_rate": 1.302325581395349e-05,
2947
+ "loss": 0.8039,
2948
+ "step": 490
2949
+ },
2950
+ {
2951
+ "epoch": 1.21,
2952
+ "learning_rate": 1.300536672629696e-05,
2953
+ "loss": 0.791,
2954
+ "step": 491
2955
+ },
2956
+ {
2957
+ "epoch": 1.21,
2958
+ "learning_rate": 1.298747763864043e-05,
2959
+ "loss": 0.7874,
2960
+ "step": 492
2961
+ },
2962
+ {
2963
+ "epoch": 1.21,
2964
+ "learning_rate": 1.29695885509839e-05,
2965
+ "loss": 0.8043,
2966
+ "step": 493
2967
+ },
2968
+ {
2969
+ "epoch": 1.22,
2970
+ "learning_rate": 1.2951699463327372e-05,
2971
+ "loss": 0.848,
2972
+ "step": 494
2973
+ },
2974
+ {
2975
+ "epoch": 1.22,
2976
+ "learning_rate": 1.293381037567084e-05,
2977
+ "loss": 0.8273,
2978
+ "step": 495
2979
+ },
2980
+ {
2981
+ "epoch": 1.22,
2982
+ "learning_rate": 1.2915921288014313e-05,
2983
+ "loss": 0.8292,
2984
+ "step": 496
2985
+ },
2986
+ {
2987
+ "epoch": 1.22,
2988
+ "learning_rate": 1.2898032200357782e-05,
2989
+ "loss": 0.8332,
2990
+ "step": 497
2991
+ },
2992
+ {
2993
+ "epoch": 1.23,
2994
+ "learning_rate": 1.2880143112701254e-05,
2995
+ "loss": 0.8476,
2996
+ "step": 498
2997
+ },
2998
+ {
2999
+ "epoch": 1.23,
3000
+ "learning_rate": 1.2862254025044724e-05,
3001
+ "loss": 0.7687,
3002
+ "step": 499
3003
+ },
3004
+ {
3005
+ "epoch": 1.23,
3006
+ "learning_rate": 1.2844364937388195e-05,
3007
+ "loss": 0.8023,
3008
+ "step": 500
3009
+ }
3010
+ ],
3011
+ "max_steps": 1218,
3012
+ "num_train_epochs": 3,
3013
+ "total_flos": 6.5007963669214e+17,
3014
+ "trial_name": null,
3015
+ "trial_params": null
3016
+ }