chavinlo commited on
Commit
53ff912
1 Parent(s): 63b7843

Upload checkpoint-400/trainer_state.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. checkpoint-400/trainer_state.json +2416 -0
checkpoint-400/trainer_state.json ADDED
@@ -0,0 +1,2416 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9843125192248539,
5
+ "global_step": 400,
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": 5.405405405405406e-07,
13
+ "loss": 1.5039,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 1.0810810810810812e-06,
19
+ "loss": 1.5785,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.01,
24
+ "learning_rate": 1.6216216216216219e-06,
25
+ "loss": 1.2981,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.01,
30
+ "learning_rate": 2.1621621621621623e-06,
31
+ "loss": 1.1631,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.01,
36
+ "learning_rate": 2.702702702702703e-06,
37
+ "loss": 1.2493,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.01,
42
+ "learning_rate": 3.2432432432432437e-06,
43
+ "loss": 1.2413,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.02,
48
+ "learning_rate": 3.7837837837837844e-06,
49
+ "loss": 1.2368,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.02,
54
+ "learning_rate": 4.324324324324325e-06,
55
+ "loss": 1.2152,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.02,
60
+ "learning_rate": 4.864864864864866e-06,
61
+ "loss": 1.1487,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.02,
66
+ "learning_rate": 5.405405405405406e-06,
67
+ "loss": 1.116,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.03,
72
+ "learning_rate": 5.945945945945947e-06,
73
+ "loss": 1.179,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.03,
78
+ "learning_rate": 6.486486486486487e-06,
79
+ "loss": 1.2134,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.03,
84
+ "learning_rate": 7.027027027027028e-06,
85
+ "loss": 1.1641,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.03,
90
+ "learning_rate": 7.567567567567569e-06,
91
+ "loss": 1.0623,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.04,
96
+ "learning_rate": 8.108108108108109e-06,
97
+ "loss": 1.0922,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.04,
102
+ "learning_rate": 8.64864864864865e-06,
103
+ "loss": 1.0979,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.04,
108
+ "learning_rate": 9.189189189189191e-06,
109
+ "loss": 1.1572,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.04,
114
+ "learning_rate": 9.729729729729732e-06,
115
+ "loss": 1.0788,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.05,
120
+ "learning_rate": 1.027027027027027e-05,
121
+ "loss": 1.1467,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.05,
126
+ "learning_rate": 1.0810810810810812e-05,
127
+ "loss": 1.0852,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.05,
132
+ "learning_rate": 1.1351351351351352e-05,
133
+ "loss": 1.1176,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.05,
138
+ "learning_rate": 1.1891891891891894e-05,
139
+ "loss": 1.1049,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.06,
144
+ "learning_rate": 1.2432432432432433e-05,
145
+ "loss": 1.0972,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.06,
150
+ "learning_rate": 1.2972972972972975e-05,
151
+ "loss": 1.1404,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.06,
156
+ "learning_rate": 1.3513513513513515e-05,
157
+ "loss": 1.1416,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.06,
162
+ "learning_rate": 1.4054054054054055e-05,
163
+ "loss": 1.1458,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.07,
168
+ "learning_rate": 1.4594594594594596e-05,
169
+ "loss": 1.1497,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.07,
174
+ "learning_rate": 1.5135135135135138e-05,
175
+ "loss": 1.14,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.07,
180
+ "learning_rate": 1.5675675675675676e-05,
181
+ "loss": 1.0808,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.07,
186
+ "learning_rate": 1.6216216216216218e-05,
187
+ "loss": 1.0996,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.08,
192
+ "learning_rate": 1.6756756756756757e-05,
193
+ "loss": 1.1064,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.08,
198
+ "learning_rate": 1.72972972972973e-05,
199
+ "loss": 1.1823,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.08,
204
+ "learning_rate": 1.783783783783784e-05,
205
+ "loss": 1.0483,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.08,
210
+ "learning_rate": 1.8378378378378383e-05,
211
+ "loss": 1.1019,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.09,
216
+ "learning_rate": 1.891891891891892e-05,
217
+ "loss": 1.0325,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.09,
222
+ "learning_rate": 1.9459459459459463e-05,
223
+ "loss": 1.1006,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.09,
228
+ "learning_rate": 2e-05,
229
+ "loss": 1.0715,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.09,
234
+ "learning_rate": 1.999996461903301e-05,
235
+ "loss": 1.1542,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.1,
240
+ "learning_rate": 1.9999858476382388e-05,
241
+ "loss": 1.1402,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.1,
246
+ "learning_rate": 1.9999681572799226e-05,
247
+ "loss": 1.0163,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.1,
252
+ "learning_rate": 1.9999433909535333e-05,
253
+ "loss": 1.0851,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.1,
258
+ "learning_rate": 1.9999115488343213e-05,
259
+ "loss": 1.1119,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.11,
264
+ "learning_rate": 1.999872631147608e-05,
265
+ "loss": 1.0726,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.11,
270
+ "learning_rate": 1.999826638168783e-05,
271
+ "loss": 1.0927,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.11,
276
+ "learning_rate": 1.9997735702233006e-05,
277
+ "loss": 1.0697,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 0.11,
282
+ "learning_rate": 1.99971342768668e-05,
283
+ "loss": 1.0363,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 0.12,
288
+ "learning_rate": 1.999646210984502e-05,
289
+ "loss": 1.1251,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 0.12,
294
+ "learning_rate": 1.999571920592405e-05,
295
+ "loss": 1.127,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 0.12,
300
+ "learning_rate": 1.9994905570360817e-05,
301
+ "loss": 1.0994,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 0.12,
306
+ "learning_rate": 1.999402120891276e-05,
307
+ "loss": 1.0688,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 0.13,
312
+ "learning_rate": 1.99930661278378e-05,
313
+ "loss": 1.0544,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 0.13,
318
+ "learning_rate": 1.9992040333894273e-05,
319
+ "loss": 1.0908,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 0.13,
324
+ "learning_rate": 1.9990943834340893e-05,
325
+ "loss": 1.1394,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 0.13,
330
+ "learning_rate": 1.9989776636936705e-05,
331
+ "loss": 1.1065,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 0.14,
336
+ "learning_rate": 1.9988538749941024e-05,
337
+ "loss": 1.0752,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 0.14,
342
+ "learning_rate": 1.9987230182113374e-05,
343
+ "loss": 1.076,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 0.14,
348
+ "learning_rate": 1.998585094271344e-05,
349
+ "loss": 1.0504,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 0.14,
354
+ "learning_rate": 1.998440104150098e-05,
355
+ "loss": 1.0934,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 0.15,
360
+ "learning_rate": 1.998288048873578e-05,
361
+ "loss": 1.1355,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 0.15,
366
+ "learning_rate": 1.9981289295177566e-05,
367
+ "loss": 1.1693,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 0.15,
372
+ "learning_rate": 1.9979627472085927e-05,
373
+ "loss": 1.0514,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 0.15,
378
+ "learning_rate": 1.997789503122025e-05,
379
+ "loss": 1.119,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 0.16,
384
+ "learning_rate": 1.9976091984839616e-05,
385
+ "loss": 1.0635,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 0.16,
390
+ "learning_rate": 1.9974218345702733e-05,
391
+ "loss": 1.1111,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 0.16,
396
+ "learning_rate": 1.9972274127067838e-05,
397
+ "loss": 1.1732,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 0.16,
402
+ "learning_rate": 1.997025934269259e-05,
403
+ "loss": 1.149,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 0.16,
408
+ "learning_rate": 1.9968174006833996e-05,
409
+ "loss": 1.1011,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 0.17,
414
+ "learning_rate": 1.9966018134248296e-05,
415
+ "loss": 1.1367,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 0.17,
420
+ "learning_rate": 1.9963791740190863e-05,
421
+ "loss": 1.1354,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 0.17,
426
+ "learning_rate": 1.996149484041609e-05,
427
+ "loss": 1.0343,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 0.17,
432
+ "learning_rate": 1.9959127451177287e-05,
433
+ "loss": 1.1997,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 0.18,
438
+ "learning_rate": 1.9956689589226555e-05,
439
+ "loss": 1.1131,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 0.18,
444
+ "learning_rate": 1.9954181271814673e-05,
445
+ "loss": 1.0335,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 0.18,
450
+ "learning_rate": 1.9951602516690988e-05,
451
+ "loss": 1.0786,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 0.18,
456
+ "learning_rate": 1.9948953342103268e-05,
457
+ "loss": 1.0926,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 0.19,
462
+ "learning_rate": 1.994623376679758e-05,
463
+ "loss": 1.1353,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 0.19,
468
+ "learning_rate": 1.9943443810018174e-05,
469
+ "loss": 1.0605,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 0.19,
474
+ "learning_rate": 1.9940583491507314e-05,
475
+ "loss": 1.1255,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 0.19,
480
+ "learning_rate": 1.993765283150517e-05,
481
+ "loss": 1.0541,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 0.2,
486
+ "learning_rate": 1.9934651850749663e-05,
487
+ "loss": 1.1197,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 0.2,
492
+ "learning_rate": 1.9931580570476306e-05,
493
+ "loss": 1.1297,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 0.2,
498
+ "learning_rate": 1.9928439012418076e-05,
499
+ "loss": 1.1324,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 0.2,
504
+ "learning_rate": 1.9925227198805247e-05,
505
+ "loss": 1.0846,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 0.21,
510
+ "learning_rate": 1.9921945152365235e-05,
511
+ "loss": 1.1831,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 0.21,
516
+ "learning_rate": 1.9918592896322432e-05,
517
+ "loss": 1.0921,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 0.21,
522
+ "learning_rate": 1.9915170454398045e-05,
523
+ "loss": 1.1037,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 0.21,
528
+ "learning_rate": 1.9911677850809943e-05,
529
+ "loss": 1.125,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 0.22,
534
+ "learning_rate": 1.9908115110272463e-05,
535
+ "loss": 1.1279,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 0.22,
540
+ "learning_rate": 1.9904482257996244e-05,
541
+ "loss": 1.1427,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 0.22,
546
+ "learning_rate": 1.990077931968805e-05,
547
+ "loss": 1.1475,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 0.22,
552
+ "learning_rate": 1.9897006321550592e-05,
553
+ "loss": 1.136,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 0.23,
558
+ "learning_rate": 1.9893163290282335e-05,
559
+ "loss": 1.0469,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 0.23,
564
+ "learning_rate": 1.9889250253077306e-05,
565
+ "loss": 1.1028,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 0.23,
570
+ "learning_rate": 1.9885267237624923e-05,
571
+ "loss": 1.1775,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 0.23,
576
+ "learning_rate": 1.988121427210976e-05,
577
+ "loss": 1.1744,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 0.24,
582
+ "learning_rate": 1.98770913852114e-05,
583
+ "loss": 1.1049,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 0.24,
588
+ "learning_rate": 1.9872898606104175e-05,
589
+ "loss": 1.1085,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 0.24,
594
+ "learning_rate": 1.9868635964457007e-05,
595
+ "loss": 1.1153,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 0.24,
600
+ "learning_rate": 1.986430349043317e-05,
601
+ "loss": 1.1637,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 0.25,
606
+ "learning_rate": 1.9859901214690094e-05,
607
+ "loss": 1.0971,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 0.25,
612
+ "learning_rate": 1.9855429168379127e-05,
613
+ "loss": 1.1537,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 0.25,
618
+ "learning_rate": 1.985088738314534e-05,
619
+ "loss": 1.0882,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 0.25,
624
+ "learning_rate": 1.9846275891127275e-05,
625
+ "loss": 1.128,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 0.26,
630
+ "learning_rate": 1.9841594724956746e-05,
631
+ "loss": 1.14,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 0.26,
636
+ "learning_rate": 1.9836843917758593e-05,
637
+ "loss": 1.0273,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 0.26,
642
+ "learning_rate": 1.983202350315044e-05,
643
+ "loss": 1.0891,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 0.26,
648
+ "learning_rate": 1.982713351524248e-05,
649
+ "loss": 1.106,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 0.27,
654
+ "learning_rate": 1.982217398863721e-05,
655
+ "loss": 1.0599,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 0.27,
660
+ "learning_rate": 1.98171449584292e-05,
661
+ "loss": 1.0396,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 0.27,
666
+ "learning_rate": 1.9812046460204837e-05,
667
+ "loss": 1.0767,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 0.27,
672
+ "learning_rate": 1.9806878530042083e-05,
673
+ "loss": 1.1568,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 0.28,
678
+ "learning_rate": 1.9801641204510216e-05,
679
+ "loss": 1.1534,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 0.28,
684
+ "learning_rate": 1.9796334520669555e-05,
685
+ "loss": 1.0332,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 0.28,
690
+ "learning_rate": 1.9790958516071228e-05,
691
+ "loss": 1.1018,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 0.28,
696
+ "learning_rate": 1.978551322875688e-05,
697
+ "loss": 1.0634,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 0.29,
702
+ "learning_rate": 1.977999869725842e-05,
703
+ "loss": 1.0766,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 0.29,
708
+ "learning_rate": 1.977441496059774e-05,
709
+ "loss": 1.0456,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 0.29,
714
+ "learning_rate": 1.9768762058286433e-05,
715
+ "loss": 1.0047,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 0.29,
720
+ "learning_rate": 1.976304003032554e-05,
721
+ "loss": 1.0798,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 0.3,
726
+ "learning_rate": 1.9757248917205228e-05,
727
+ "loss": 1.1055,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 0.3,
732
+ "learning_rate": 1.975138875990454e-05,
733
+ "loss": 1.1921,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 0.3,
738
+ "learning_rate": 1.974545959989108e-05,
739
+ "loss": 1.0465,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 0.3,
744
+ "learning_rate": 1.9739461479120727e-05,
745
+ "loss": 1.129,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 0.31,
750
+ "learning_rate": 1.973339444003735e-05,
751
+ "loss": 1.0773,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 0.31,
756
+ "learning_rate": 1.9727258525572487e-05,
757
+ "loss": 1.0955,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 0.31,
762
+ "learning_rate": 1.9721053779145057e-05,
763
+ "loss": 1.0941,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 0.31,
768
+ "learning_rate": 1.9714780244661044e-05,
769
+ "loss": 1.1672,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 0.31,
774
+ "learning_rate": 1.9708437966513196e-05,
775
+ "loss": 1.1014,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 0.32,
780
+ "learning_rate": 1.9702026989580694e-05,
781
+ "loss": 1.1272,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 0.32,
786
+ "learning_rate": 1.969554735922885e-05,
787
+ "loss": 1.0398,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 0.32,
792
+ "learning_rate": 1.968899912130879e-05,
793
+ "loss": 1.106,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 0.32,
798
+ "learning_rate": 1.9682382322157103e-05,
799
+ "loss": 1.1033,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 0.33,
804
+ "learning_rate": 1.9675697008595545e-05,
805
+ "loss": 1.0154,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 0.33,
810
+ "learning_rate": 1.9668943227930686e-05,
811
+ "loss": 1.1219,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 0.33,
816
+ "learning_rate": 1.966212102795358e-05,
817
+ "loss": 1.0824,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 0.33,
822
+ "learning_rate": 1.965523045693944e-05,
823
+ "loss": 1.0581,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 0.34,
828
+ "learning_rate": 1.964827156364728e-05,
829
+ "loss": 1.1327,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 0.34,
834
+ "learning_rate": 1.964124439731957e-05,
835
+ "loss": 1.1144,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 0.34,
840
+ "learning_rate": 1.9634149007681894e-05,
841
+ "loss": 1.1436,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 0.34,
846
+ "learning_rate": 1.962698544494261e-05,
847
+ "loss": 1.0457,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 0.35,
852
+ "learning_rate": 1.9619753759792466e-05,
853
+ "loss": 1.1053,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 0.35,
858
+ "learning_rate": 1.961245400340427e-05,
859
+ "loss": 1.0881,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 0.35,
864
+ "learning_rate": 1.9605086227432512e-05,
865
+ "loss": 1.1058,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 0.35,
870
+ "learning_rate": 1.9597650484012997e-05,
871
+ "loss": 1.1418,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 0.36,
876
+ "learning_rate": 1.9590146825762476e-05,
877
+ "loss": 1.0623,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 0.36,
882
+ "learning_rate": 1.9582575305778297e-05,
883
+ "loss": 1.0987,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 0.36,
888
+ "learning_rate": 1.9574935977637994e-05,
889
+ "loss": 1.0764,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 0.36,
894
+ "learning_rate": 1.9567228895398936e-05,
895
+ "loss": 1.153,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 0.37,
900
+ "learning_rate": 1.955945411359792e-05,
901
+ "loss": 1.1233,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 0.37,
906
+ "learning_rate": 1.9551611687250808e-05,
907
+ "loss": 1.0827,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 0.37,
912
+ "learning_rate": 1.9543701671852127e-05,
913
+ "loss": 1.0487,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 0.37,
918
+ "learning_rate": 1.9535724123374674e-05,
919
+ "loss": 1.0795,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 0.38,
924
+ "learning_rate": 1.952767909826913e-05,
925
+ "loss": 1.15,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 0.38,
930
+ "learning_rate": 1.951956665346364e-05,
931
+ "loss": 1.0645,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 0.38,
936
+ "learning_rate": 1.951138684636344e-05,
937
+ "loss": 1.0764,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 0.38,
942
+ "learning_rate": 1.9503139734850426e-05,
943
+ "loss": 1.1185,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 0.39,
948
+ "learning_rate": 1.9494825377282746e-05,
949
+ "loss": 1.0879,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 0.39,
954
+ "learning_rate": 1.9486443832494414e-05,
955
+ "loss": 1.0558,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 0.39,
960
+ "learning_rate": 1.9477995159794854e-05,
961
+ "loss": 1.1042,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 0.39,
966
+ "learning_rate": 1.9469479418968506e-05,
967
+ "loss": 1.1029,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 0.4,
972
+ "learning_rate": 1.9460896670274408e-05,
973
+ "loss": 0.9836,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 0.4,
978
+ "learning_rate": 1.9452246974445743e-05,
979
+ "loss": 1.1325,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 0.4,
984
+ "learning_rate": 1.9443530392689434e-05,
985
+ "loss": 1.1018,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 0.4,
990
+ "learning_rate": 1.94347469866857e-05,
991
+ "loss": 1.1525,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 0.41,
996
+ "learning_rate": 1.9425896818587615e-05,
997
+ "loss": 1.0719,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 0.41,
1002
+ "learning_rate": 1.941697995102069e-05,
1003
+ "loss": 1.1595,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 0.41,
1008
+ "learning_rate": 1.9407996447082394e-05,
1009
+ "loss": 1.1698,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 0.41,
1014
+ "learning_rate": 1.939894637034174e-05,
1015
+ "loss": 1.131,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 0.42,
1020
+ "learning_rate": 1.9389829784838833e-05,
1021
+ "loss": 1.0604,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 0.42,
1026
+ "learning_rate": 1.938064675508438e-05,
1027
+ "loss": 1.0422,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 0.42,
1032
+ "learning_rate": 1.9371397346059286e-05,
1033
+ "loss": 1.1098,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 0.42,
1038
+ "learning_rate": 1.936208162321415e-05,
1039
+ "loss": 1.0682,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 0.43,
1044
+ "learning_rate": 1.9352699652468835e-05,
1045
+ "loss": 1.0763,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 0.43,
1050
+ "learning_rate": 1.9343251500211977e-05,
1051
+ "loss": 1.1038,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 0.43,
1056
+ "learning_rate": 1.933373723330053e-05,
1057
+ "loss": 1.0613,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 0.43,
1062
+ "learning_rate": 1.9324156919059286e-05,
1063
+ "loss": 1.1157,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 0.44,
1068
+ "learning_rate": 1.93145106252804e-05,
1069
+ "loss": 1.1264,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 0.44,
1074
+ "learning_rate": 1.9304798420222918e-05,
1075
+ "loss": 1.0787,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 0.44,
1080
+ "learning_rate": 1.9295020372612276e-05,
1081
+ "loss": 1.1298,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 0.44,
1086
+ "learning_rate": 1.9285176551639826e-05,
1087
+ "loss": 1.1175,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 0.45,
1092
+ "learning_rate": 1.9275267026962358e-05,
1093
+ "loss": 1.1686,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 0.45,
1098
+ "learning_rate": 1.9265291868701584e-05,
1099
+ "loss": 1.0939,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 0.45,
1104
+ "learning_rate": 1.9255251147443646e-05,
1105
+ "loss": 1.1674,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 0.45,
1110
+ "learning_rate": 1.924514493423864e-05,
1111
+ "loss": 1.0747,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 0.46,
1116
+ "learning_rate": 1.9234973300600074e-05,
1117
+ "loss": 1.1085,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 0.46,
1122
+ "learning_rate": 1.92247363185044e-05,
1123
+ "loss": 1.1157,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 0.46,
1128
+ "learning_rate": 1.9214434060390484e-05,
1129
+ "loss": 1.1827,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 0.46,
1134
+ "learning_rate": 1.9204066599159094e-05,
1135
+ "loss": 1.1126,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 0.47,
1140
+ "learning_rate": 1.9193634008172396e-05,
1141
+ "loss": 1.1158,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 0.47,
1146
+ "learning_rate": 1.9183136361253417e-05,
1147
+ "loss": 1.0609,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 0.47,
1152
+ "learning_rate": 1.917257373268554e-05,
1153
+ "loss": 1.0617,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 0.47,
1158
+ "learning_rate": 1.916194619721196e-05,
1159
+ "loss": 1.0724,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 0.47,
1164
+ "learning_rate": 1.915125383003518e-05,
1165
+ "loss": 1.0784,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 0.48,
1170
+ "learning_rate": 1.914049670681646e-05,
1171
+ "loss": 1.1141,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 0.48,
1176
+ "learning_rate": 1.912967490367528e-05,
1177
+ "loss": 1.0305,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 0.48,
1182
+ "learning_rate": 1.9118788497188815e-05,
1183
+ "loss": 1.026,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 0.48,
1188
+ "learning_rate": 1.9107837564391376e-05,
1189
+ "loss": 1.0663,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 0.49,
1194
+ "learning_rate": 1.9096822182773887e-05,
1195
+ "loss": 1.0368,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 0.49,
1200
+ "learning_rate": 1.9085742430283322e-05,
1201
+ "loss": 1.1052,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 0.49,
1206
+ "learning_rate": 1.907459838532215e-05,
1207
+ "loss": 1.0834,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 0.49,
1212
+ "learning_rate": 1.9063390126747778e-05,
1213
+ "loss": 1.1579,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 0.5,
1218
+ "learning_rate": 1.9052117733872025e-05,
1219
+ "loss": 1.1034,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 0.5,
1224
+ "learning_rate": 1.904078128646052e-05,
1225
+ "loss": 1.1173,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 0.5,
1230
+ "learning_rate": 1.902938086473215e-05,
1231
+ "loss": 1.1546,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 0.5,
1236
+ "learning_rate": 1.901791654935852e-05,
1237
+ "loss": 1.1274,
1238
+ "step": 205
1239
+ },
1240
+ {
1241
+ "epoch": 0.51,
1242
+ "learning_rate": 1.9006388421463322e-05,
1243
+ "loss": 1.0896,
1244
+ "step": 206
1245
+ },
1246
+ {
1247
+ "epoch": 0.51,
1248
+ "learning_rate": 1.899479656262183e-05,
1249
+ "loss": 1.1937,
1250
+ "step": 207
1251
+ },
1252
+ {
1253
+ "epoch": 0.51,
1254
+ "learning_rate": 1.898314105486028e-05,
1255
+ "loss": 1.0613,
1256
+ "step": 208
1257
+ },
1258
+ {
1259
+ "epoch": 0.51,
1260
+ "learning_rate": 1.8971421980655295e-05,
1261
+ "loss": 1.0896,
1262
+ "step": 209
1263
+ },
1264
+ {
1265
+ "epoch": 0.52,
1266
+ "learning_rate": 1.8959639422933316e-05,
1267
+ "loss": 1.0578,
1268
+ "step": 210
1269
+ },
1270
+ {
1271
+ "epoch": 0.52,
1272
+ "learning_rate": 1.894779346506999e-05,
1273
+ "loss": 1.0935,
1274
+ "step": 211
1275
+ },
1276
+ {
1277
+ "epoch": 0.52,
1278
+ "learning_rate": 1.893588419088962e-05,
1279
+ "loss": 1.0655,
1280
+ "step": 212
1281
+ },
1282
+ {
1283
+ "epoch": 0.52,
1284
+ "learning_rate": 1.892391168466452e-05,
1285
+ "loss": 1.1562,
1286
+ "step": 213
1287
+ },
1288
+ {
1289
+ "epoch": 0.53,
1290
+ "learning_rate": 1.891187603111447e-05,
1291
+ "loss": 1.1434,
1292
+ "step": 214
1293
+ },
1294
+ {
1295
+ "epoch": 0.53,
1296
+ "learning_rate": 1.8899777315406073e-05,
1297
+ "loss": 1.0462,
1298
+ "step": 215
1299
+ },
1300
+ {
1301
+ "epoch": 0.53,
1302
+ "learning_rate": 1.8887615623152188e-05,
1303
+ "loss": 1.1151,
1304
+ "step": 216
1305
+ },
1306
+ {
1307
+ "epoch": 0.53,
1308
+ "learning_rate": 1.88753910404113e-05,
1309
+ "loss": 1.1457,
1310
+ "step": 217
1311
+ },
1312
+ {
1313
+ "epoch": 0.54,
1314
+ "learning_rate": 1.8863103653686917e-05,
1315
+ "loss": 1.0401,
1316
+ "step": 218
1317
+ },
1318
+ {
1319
+ "epoch": 0.54,
1320
+ "learning_rate": 1.8850753549926967e-05,
1321
+ "loss": 1.074,
1322
+ "step": 219
1323
+ },
1324
+ {
1325
+ "epoch": 0.54,
1326
+ "learning_rate": 1.8838340816523175e-05,
1327
+ "loss": 1.14,
1328
+ "step": 220
1329
+ },
1330
+ {
1331
+ "epoch": 0.54,
1332
+ "learning_rate": 1.8825865541310438e-05,
1333
+ "loss": 1.1483,
1334
+ "step": 221
1335
+ },
1336
+ {
1337
+ "epoch": 0.55,
1338
+ "learning_rate": 1.8813327812566217e-05,
1339
+ "loss": 1.156,
1340
+ "step": 222
1341
+ },
1342
+ {
1343
+ "epoch": 0.55,
1344
+ "learning_rate": 1.880072771900991e-05,
1345
+ "loss": 1.0334,
1346
+ "step": 223
1347
+ },
1348
+ {
1349
+ "epoch": 0.55,
1350
+ "learning_rate": 1.878806534980221e-05,
1351
+ "loss": 1.1854,
1352
+ "step": 224
1353
+ },
1354
+ {
1355
+ "epoch": 0.55,
1356
+ "learning_rate": 1.8775340794544497e-05,
1357
+ "loss": 1.0593,
1358
+ "step": 225
1359
+ },
1360
+ {
1361
+ "epoch": 0.56,
1362
+ "learning_rate": 1.876255414327818e-05,
1363
+ "loss": 1.0863,
1364
+ "step": 226
1365
+ },
1366
+ {
1367
+ "epoch": 0.56,
1368
+ "learning_rate": 1.8749705486484074e-05,
1369
+ "loss": 1.2038,
1370
+ "step": 227
1371
+ },
1372
+ {
1373
+ "epoch": 0.56,
1374
+ "learning_rate": 1.8736794915081765e-05,
1375
+ "loss": 1.0991,
1376
+ "step": 228
1377
+ },
1378
+ {
1379
+ "epoch": 0.56,
1380
+ "learning_rate": 1.8723822520428954e-05,
1381
+ "loss": 1.0572,
1382
+ "step": 229
1383
+ },
1384
+ {
1385
+ "epoch": 0.57,
1386
+ "learning_rate": 1.8710788394320807e-05,
1387
+ "loss": 1.0798,
1388
+ "step": 230
1389
+ },
1390
+ {
1391
+ "epoch": 0.57,
1392
+ "learning_rate": 1.8697692628989327e-05,
1393
+ "loss": 1.082,
1394
+ "step": 231
1395
+ },
1396
+ {
1397
+ "epoch": 0.57,
1398
+ "learning_rate": 1.868453531710268e-05,
1399
+ "loss": 1.1121,
1400
+ "step": 232
1401
+ },
1402
+ {
1403
+ "epoch": 0.57,
1404
+ "learning_rate": 1.8671316551764552e-05,
1405
+ "loss": 1.089,
1406
+ "step": 233
1407
+ },
1408
+ {
1409
+ "epoch": 0.58,
1410
+ "learning_rate": 1.865803642651348e-05,
1411
+ "loss": 1.1132,
1412
+ "step": 234
1413
+ },
1414
+ {
1415
+ "epoch": 0.58,
1416
+ "learning_rate": 1.8644695035322203e-05,
1417
+ "loss": 1.1323,
1418
+ "step": 235
1419
+ },
1420
+ {
1421
+ "epoch": 0.58,
1422
+ "learning_rate": 1.8631292472596978e-05,
1423
+ "loss": 1.1352,
1424
+ "step": 236
1425
+ },
1426
+ {
1427
+ "epoch": 0.58,
1428
+ "learning_rate": 1.8617828833176935e-05,
1429
+ "loss": 1.0801,
1430
+ "step": 237
1431
+ },
1432
+ {
1433
+ "epoch": 0.59,
1434
+ "learning_rate": 1.860430421233339e-05,
1435
+ "loss": 1.1049,
1436
+ "step": 238
1437
+ },
1438
+ {
1439
+ "epoch": 0.59,
1440
+ "learning_rate": 1.859071870576918e-05,
1441
+ "loss": 1.0518,
1442
+ "step": 239
1443
+ },
1444
+ {
1445
+ "epoch": 0.59,
1446
+ "learning_rate": 1.857707240961797e-05,
1447
+ "loss": 1.0691,
1448
+ "step": 240
1449
+ },
1450
+ {
1451
+ "epoch": 0.59,
1452
+ "learning_rate": 1.8563365420443594e-05,
1453
+ "loss": 1.0941,
1454
+ "step": 241
1455
+ },
1456
+ {
1457
+ "epoch": 0.6,
1458
+ "learning_rate": 1.854959783523936e-05,
1459
+ "loss": 1.1016,
1460
+ "step": 242
1461
+ },
1462
+ {
1463
+ "epoch": 0.6,
1464
+ "learning_rate": 1.853576975142736e-05,
1465
+ "loss": 1.0896,
1466
+ "step": 243
1467
+ },
1468
+ {
1469
+ "epoch": 0.6,
1470
+ "learning_rate": 1.852188126685779e-05,
1471
+ "loss": 1.0822,
1472
+ "step": 244
1473
+ },
1474
+ {
1475
+ "epoch": 0.6,
1476
+ "learning_rate": 1.8507932479808254e-05,
1477
+ "loss": 0.9657,
1478
+ "step": 245
1479
+ },
1480
+ {
1481
+ "epoch": 0.61,
1482
+ "learning_rate": 1.8493923488983066e-05,
1483
+ "loss": 1.0548,
1484
+ "step": 246
1485
+ },
1486
+ {
1487
+ "epoch": 0.61,
1488
+ "learning_rate": 1.847985439351256e-05,
1489
+ "loss": 1.0908,
1490
+ "step": 247
1491
+ },
1492
+ {
1493
+ "epoch": 0.61,
1494
+ "learning_rate": 1.846572529295237e-05,
1495
+ "loss": 1.1709,
1496
+ "step": 248
1497
+ },
1498
+ {
1499
+ "epoch": 0.61,
1500
+ "learning_rate": 1.845153628728274e-05,
1501
+ "loss": 1.1686,
1502
+ "step": 249
1503
+ },
1504
+ {
1505
+ "epoch": 0.62,
1506
+ "learning_rate": 1.8437287476907828e-05,
1507
+ "loss": 1.1704,
1508
+ "step": 250
1509
+ },
1510
+ {
1511
+ "epoch": 0.62,
1512
+ "learning_rate": 1.842297896265497e-05,
1513
+ "loss": 1.054,
1514
+ "step": 251
1515
+ },
1516
+ {
1517
+ "epoch": 0.62,
1518
+ "learning_rate": 1.8408610845773974e-05,
1519
+ "loss": 1.1077,
1520
+ "step": 252
1521
+ },
1522
+ {
1523
+ "epoch": 0.62,
1524
+ "learning_rate": 1.8394183227936418e-05,
1525
+ "loss": 1.0797,
1526
+ "step": 253
1527
+ },
1528
+ {
1529
+ "epoch": 0.63,
1530
+ "learning_rate": 1.8379696211234918e-05,
1531
+ "loss": 1.1079,
1532
+ "step": 254
1533
+ },
1534
+ {
1535
+ "epoch": 0.63,
1536
+ "learning_rate": 1.8365149898182403e-05,
1537
+ "loss": 1.11,
1538
+ "step": 255
1539
+ },
1540
+ {
1541
+ "epoch": 0.63,
1542
+ "learning_rate": 1.8350544391711396e-05,
1543
+ "loss": 1.0777,
1544
+ "step": 256
1545
+ },
1546
+ {
1547
+ "epoch": 0.63,
1548
+ "learning_rate": 1.833587979517329e-05,
1549
+ "loss": 1.0761,
1550
+ "step": 257
1551
+ },
1552
+ {
1553
+ "epoch": 0.63,
1554
+ "learning_rate": 1.8321156212337604e-05,
1555
+ "loss": 1.0941,
1556
+ "step": 258
1557
+ },
1558
+ {
1559
+ "epoch": 0.64,
1560
+ "learning_rate": 1.830637374739126e-05,
1561
+ "loss": 1.128,
1562
+ "step": 259
1563
+ },
1564
+ {
1565
+ "epoch": 0.64,
1566
+ "learning_rate": 1.829153250493783e-05,
1567
+ "loss": 1.0112,
1568
+ "step": 260
1569
+ },
1570
+ {
1571
+ "epoch": 0.64,
1572
+ "learning_rate": 1.827663258999683e-05,
1573
+ "loss": 1.0975,
1574
+ "step": 261
1575
+ },
1576
+ {
1577
+ "epoch": 0.64,
1578
+ "learning_rate": 1.8261674108002925e-05,
1579
+ "loss": 1.0442,
1580
+ "step": 262
1581
+ },
1582
+ {
1583
+ "epoch": 0.65,
1584
+ "learning_rate": 1.824665716480524e-05,
1585
+ "loss": 1.0653,
1586
+ "step": 263
1587
+ },
1588
+ {
1589
+ "epoch": 0.65,
1590
+ "learning_rate": 1.823158186666656e-05,
1591
+ "loss": 1.0561,
1592
+ "step": 264
1593
+ },
1594
+ {
1595
+ "epoch": 0.65,
1596
+ "learning_rate": 1.821644832026261e-05,
1597
+ "loss": 1.1615,
1598
+ "step": 265
1599
+ },
1600
+ {
1601
+ "epoch": 0.65,
1602
+ "learning_rate": 1.82012566326813e-05,
1603
+ "loss": 1.1045,
1604
+ "step": 266
1605
+ },
1606
+ {
1607
+ "epoch": 0.66,
1608
+ "learning_rate": 1.8186006911421937e-05,
1609
+ "loss": 1.1486,
1610
+ "step": 267
1611
+ },
1612
+ {
1613
+ "epoch": 0.66,
1614
+ "learning_rate": 1.817069926439451e-05,
1615
+ "loss": 1.1622,
1616
+ "step": 268
1617
+ },
1618
+ {
1619
+ "epoch": 0.66,
1620
+ "learning_rate": 1.8155333799918883e-05,
1621
+ "loss": 1.0986,
1622
+ "step": 269
1623
+ },
1624
+ {
1625
+ "epoch": 0.66,
1626
+ "learning_rate": 1.8139910626724058e-05,
1627
+ "loss": 1.0416,
1628
+ "step": 270
1629
+ },
1630
+ {
1631
+ "epoch": 0.67,
1632
+ "learning_rate": 1.8124429853947387e-05,
1633
+ "loss": 1.1328,
1634
+ "step": 271
1635
+ },
1636
+ {
1637
+ "epoch": 0.67,
1638
+ "learning_rate": 1.8108891591133812e-05,
1639
+ "loss": 1.0921,
1640
+ "step": 272
1641
+ },
1642
+ {
1643
+ "epoch": 0.67,
1644
+ "learning_rate": 1.809329594823509e-05,
1645
+ "loss": 1.0795,
1646
+ "step": 273
1647
+ },
1648
+ {
1649
+ "epoch": 0.67,
1650
+ "learning_rate": 1.8077643035609006e-05,
1651
+ "loss": 1.1597,
1652
+ "step": 274
1653
+ },
1654
+ {
1655
+ "epoch": 0.68,
1656
+ "learning_rate": 1.806193296401859e-05,
1657
+ "loss": 1.0555,
1658
+ "step": 275
1659
+ },
1660
+ {
1661
+ "epoch": 0.68,
1662
+ "learning_rate": 1.804616584463136e-05,
1663
+ "loss": 1.1308,
1664
+ "step": 276
1665
+ },
1666
+ {
1667
+ "epoch": 0.68,
1668
+ "learning_rate": 1.803034178901849e-05,
1669
+ "loss": 1.0728,
1670
+ "step": 277
1671
+ },
1672
+ {
1673
+ "epoch": 0.68,
1674
+ "learning_rate": 1.8014460909154058e-05,
1675
+ "loss": 1.0919,
1676
+ "step": 278
1677
+ },
1678
+ {
1679
+ "epoch": 0.69,
1680
+ "learning_rate": 1.799852331741425e-05,
1681
+ "loss": 1.1573,
1682
+ "step": 279
1683
+ },
1684
+ {
1685
+ "epoch": 0.69,
1686
+ "learning_rate": 1.7982529126576543e-05,
1687
+ "loss": 1.0903,
1688
+ "step": 280
1689
+ },
1690
+ {
1691
+ "epoch": 0.69,
1692
+ "learning_rate": 1.7966478449818925e-05,
1693
+ "loss": 1.1511,
1694
+ "step": 281
1695
+ },
1696
+ {
1697
+ "epoch": 0.69,
1698
+ "learning_rate": 1.7950371400719087e-05,
1699
+ "loss": 1.1447,
1700
+ "step": 282
1701
+ },
1702
+ {
1703
+ "epoch": 0.7,
1704
+ "learning_rate": 1.7934208093253625e-05,
1705
+ "loss": 1.1149,
1706
+ "step": 283
1707
+ },
1708
+ {
1709
+ "epoch": 0.7,
1710
+ "learning_rate": 1.7917988641797227e-05,
1711
+ "loss": 1.1297,
1712
+ "step": 284
1713
+ },
1714
+ {
1715
+ "epoch": 0.7,
1716
+ "learning_rate": 1.7901713161121873e-05,
1717
+ "loss": 1.0903,
1718
+ "step": 285
1719
+ },
1720
+ {
1721
+ "epoch": 0.7,
1722
+ "learning_rate": 1.7885381766396008e-05,
1723
+ "loss": 1.0677,
1724
+ "step": 286
1725
+ },
1726
+ {
1727
+ "epoch": 0.71,
1728
+ "learning_rate": 1.786899457318374e-05,
1729
+ "loss": 1.0687,
1730
+ "step": 287
1731
+ },
1732
+ {
1733
+ "epoch": 0.71,
1734
+ "learning_rate": 1.7852551697444017e-05,
1735
+ "loss": 1.0536,
1736
+ "step": 288
1737
+ },
1738
+ {
1739
+ "epoch": 0.71,
1740
+ "learning_rate": 1.783605325552981e-05,
1741
+ "loss": 1.0364,
1742
+ "step": 289
1743
+ },
1744
+ {
1745
+ "epoch": 0.71,
1746
+ "learning_rate": 1.7819499364187282e-05,
1747
+ "loss": 1.1249,
1748
+ "step": 290
1749
+ },
1750
+ {
1751
+ "epoch": 0.72,
1752
+ "learning_rate": 1.780289014055497e-05,
1753
+ "loss": 1.0534,
1754
+ "step": 291
1755
+ },
1756
+ {
1757
+ "epoch": 0.72,
1758
+ "learning_rate": 1.7786225702162955e-05,
1759
+ "loss": 1.0512,
1760
+ "step": 292
1761
+ },
1762
+ {
1763
+ "epoch": 0.72,
1764
+ "learning_rate": 1.7769506166932026e-05,
1765
+ "loss": 1.1828,
1766
+ "step": 293
1767
+ },
1768
+ {
1769
+ "epoch": 0.72,
1770
+ "learning_rate": 1.7752731653172847e-05,
1771
+ "loss": 1.064,
1772
+ "step": 294
1773
+ },
1774
+ {
1775
+ "epoch": 0.73,
1776
+ "learning_rate": 1.7735902279585118e-05,
1777
+ "loss": 1.0973,
1778
+ "step": 295
1779
+ },
1780
+ {
1781
+ "epoch": 0.73,
1782
+ "learning_rate": 1.7719018165256745e-05,
1783
+ "loss": 1.0576,
1784
+ "step": 296
1785
+ },
1786
+ {
1787
+ "epoch": 0.73,
1788
+ "learning_rate": 1.7702079429662986e-05,
1789
+ "loss": 1.1355,
1790
+ "step": 297
1791
+ },
1792
+ {
1793
+ "epoch": 0.73,
1794
+ "learning_rate": 1.7685086192665605e-05,
1795
+ "loss": 0.9462,
1796
+ "step": 298
1797
+ },
1798
+ {
1799
+ "epoch": 0.74,
1800
+ "learning_rate": 1.7668038574512045e-05,
1801
+ "loss": 1.0778,
1802
+ "step": 299
1803
+ },
1804
+ {
1805
+ "epoch": 0.74,
1806
+ "learning_rate": 1.7650936695834536e-05,
1807
+ "loss": 1.1013,
1808
+ "step": 300
1809
+ },
1810
+ {
1811
+ "epoch": 0.74,
1812
+ "learning_rate": 1.763378067764929e-05,
1813
+ "loss": 1.0889,
1814
+ "step": 301
1815
+ },
1816
+ {
1817
+ "epoch": 0.74,
1818
+ "learning_rate": 1.7616570641355602e-05,
1819
+ "loss": 1.013,
1820
+ "step": 302
1821
+ },
1822
+ {
1823
+ "epoch": 0.75,
1824
+ "learning_rate": 1.759930670873502e-05,
1825
+ "loss": 1.0847,
1826
+ "step": 303
1827
+ },
1828
+ {
1829
+ "epoch": 0.75,
1830
+ "learning_rate": 1.758198900195047e-05,
1831
+ "loss": 1.1345,
1832
+ "step": 304
1833
+ },
1834
+ {
1835
+ "epoch": 0.75,
1836
+ "learning_rate": 1.7564617643545395e-05,
1837
+ "loss": 1.0481,
1838
+ "step": 305
1839
+ },
1840
+ {
1841
+ "epoch": 0.75,
1842
+ "learning_rate": 1.7547192756442887e-05,
1843
+ "loss": 1.0952,
1844
+ "step": 306
1845
+ },
1846
+ {
1847
+ "epoch": 0.76,
1848
+ "learning_rate": 1.7529714463944815e-05,
1849
+ "loss": 1.0869,
1850
+ "step": 307
1851
+ },
1852
+ {
1853
+ "epoch": 0.76,
1854
+ "learning_rate": 1.751218288973096e-05,
1855
+ "loss": 1.1197,
1856
+ "step": 308
1857
+ },
1858
+ {
1859
+ "epoch": 0.76,
1860
+ "learning_rate": 1.7494598157858127e-05,
1861
+ "loss": 1.1011,
1862
+ "step": 309
1863
+ },
1864
+ {
1865
+ "epoch": 0.76,
1866
+ "learning_rate": 1.7476960392759284e-05,
1867
+ "loss": 1.079,
1868
+ "step": 310
1869
+ },
1870
+ {
1871
+ "epoch": 0.77,
1872
+ "learning_rate": 1.7459269719242665e-05,
1873
+ "loss": 1.0569,
1874
+ "step": 311
1875
+ },
1876
+ {
1877
+ "epoch": 0.77,
1878
+ "learning_rate": 1.74415262624909e-05,
1879
+ "loss": 1.1322,
1880
+ "step": 312
1881
+ },
1882
+ {
1883
+ "epoch": 0.77,
1884
+ "learning_rate": 1.742373014806012e-05,
1885
+ "loss": 1.121,
1886
+ "step": 313
1887
+ },
1888
+ {
1889
+ "epoch": 0.77,
1890
+ "learning_rate": 1.740588150187907e-05,
1891
+ "loss": 1.1544,
1892
+ "step": 314
1893
+ },
1894
+ {
1895
+ "epoch": 0.78,
1896
+ "learning_rate": 1.7387980450248222e-05,
1897
+ "loss": 1.1591,
1898
+ "step": 315
1899
+ },
1900
+ {
1901
+ "epoch": 0.78,
1902
+ "learning_rate": 1.7370027119838884e-05,
1903
+ "loss": 1.0956,
1904
+ "step": 316
1905
+ },
1906
+ {
1907
+ "epoch": 0.78,
1908
+ "learning_rate": 1.735202163769229e-05,
1909
+ "loss": 1.1096,
1910
+ "step": 317
1911
+ },
1912
+ {
1913
+ "epoch": 0.78,
1914
+ "learning_rate": 1.7333964131218714e-05,
1915
+ "loss": 1.0962,
1916
+ "step": 318
1917
+ },
1918
+ {
1919
+ "epoch": 0.78,
1920
+ "learning_rate": 1.7315854728196568e-05,
1921
+ "loss": 1.1088,
1922
+ "step": 319
1923
+ },
1924
+ {
1925
+ "epoch": 0.79,
1926
+ "learning_rate": 1.729769355677149e-05,
1927
+ "loss": 1.0693,
1928
+ "step": 320
1929
+ },
1930
+ {
1931
+ "epoch": 0.79,
1932
+ "learning_rate": 1.7279480745455433e-05,
1933
+ "loss": 1.0909,
1934
+ "step": 321
1935
+ },
1936
+ {
1937
+ "epoch": 0.79,
1938
+ "learning_rate": 1.7261216423125782e-05,
1939
+ "loss": 1.0636,
1940
+ "step": 322
1941
+ },
1942
+ {
1943
+ "epoch": 0.79,
1944
+ "learning_rate": 1.724290071902441e-05,
1945
+ "loss": 1.131,
1946
+ "step": 323
1947
+ },
1948
+ {
1949
+ "epoch": 0.8,
1950
+ "learning_rate": 1.7224533762756775e-05,
1951
+ "loss": 1.1824,
1952
+ "step": 324
1953
+ },
1954
+ {
1955
+ "epoch": 0.8,
1956
+ "learning_rate": 1.720611568429103e-05,
1957
+ "loss": 1.002,
1958
+ "step": 325
1959
+ },
1960
+ {
1961
+ "epoch": 0.8,
1962
+ "learning_rate": 1.718764661395704e-05,
1963
+ "loss": 1.0908,
1964
+ "step": 326
1965
+ },
1966
+ {
1967
+ "epoch": 0.8,
1968
+ "learning_rate": 1.716912668244553e-05,
1969
+ "loss": 1.1213,
1970
+ "step": 327
1971
+ },
1972
+ {
1973
+ "epoch": 0.81,
1974
+ "learning_rate": 1.715055602080711e-05,
1975
+ "loss": 1.1056,
1976
+ "step": 328
1977
+ },
1978
+ {
1979
+ "epoch": 0.81,
1980
+ "learning_rate": 1.7131934760451385e-05,
1981
+ "loss": 1.073,
1982
+ "step": 329
1983
+ },
1984
+ {
1985
+ "epoch": 0.81,
1986
+ "learning_rate": 1.7113263033145985e-05,
1987
+ "loss": 1.1817,
1988
+ "step": 330
1989
+ },
1990
+ {
1991
+ "epoch": 0.81,
1992
+ "learning_rate": 1.7094540971015663e-05,
1993
+ "loss": 1.0157,
1994
+ "step": 331
1995
+ },
1996
+ {
1997
+ "epoch": 0.82,
1998
+ "learning_rate": 1.7075768706541355e-05,
1999
+ "loss": 1.0835,
2000
+ "step": 332
2001
+ },
2002
+ {
2003
+ "epoch": 0.82,
2004
+ "learning_rate": 1.7056946372559234e-05,
2005
+ "loss": 0.9943,
2006
+ "step": 333
2007
+ },
2008
+ {
2009
+ "epoch": 0.82,
2010
+ "learning_rate": 1.7038074102259775e-05,
2011
+ "loss": 1.1713,
2012
+ "step": 334
2013
+ },
2014
+ {
2015
+ "epoch": 0.82,
2016
+ "learning_rate": 1.7019152029186817e-05,
2017
+ "loss": 1.0421,
2018
+ "step": 335
2019
+ },
2020
+ {
2021
+ "epoch": 0.83,
2022
+ "learning_rate": 1.70001802872366e-05,
2023
+ "loss": 1.1,
2024
+ "step": 336
2025
+ },
2026
+ {
2027
+ "epoch": 0.83,
2028
+ "learning_rate": 1.6981159010656847e-05,
2029
+ "loss": 1.0711,
2030
+ "step": 337
2031
+ },
2032
+ {
2033
+ "epoch": 0.83,
2034
+ "learning_rate": 1.6962088334045785e-05,
2035
+ "loss": 1.1572,
2036
+ "step": 338
2037
+ },
2038
+ {
2039
+ "epoch": 0.83,
2040
+ "learning_rate": 1.694296839235121e-05,
2041
+ "loss": 1.0448,
2042
+ "step": 339
2043
+ },
2044
+ {
2045
+ "epoch": 0.84,
2046
+ "learning_rate": 1.692379932086953e-05,
2047
+ "loss": 1.0735,
2048
+ "step": 340
2049
+ },
2050
+ {
2051
+ "epoch": 0.84,
2052
+ "learning_rate": 1.6904581255244802e-05,
2053
+ "loss": 1.0977,
2054
+ "step": 341
2055
+ },
2056
+ {
2057
+ "epoch": 0.84,
2058
+ "learning_rate": 1.688531433146777e-05,
2059
+ "loss": 1.0777,
2060
+ "step": 342
2061
+ },
2062
+ {
2063
+ "epoch": 0.84,
2064
+ "learning_rate": 1.6865998685874923e-05,
2065
+ "loss": 1.077,
2066
+ "step": 343
2067
+ },
2068
+ {
2069
+ "epoch": 0.85,
2070
+ "learning_rate": 1.6846634455147498e-05,
2071
+ "loss": 1.0797,
2072
+ "step": 344
2073
+ },
2074
+ {
2075
+ "epoch": 0.85,
2076
+ "learning_rate": 1.6827221776310532e-05,
2077
+ "loss": 1.0048,
2078
+ "step": 345
2079
+ },
2080
+ {
2081
+ "epoch": 0.85,
2082
+ "learning_rate": 1.6807760786731905e-05,
2083
+ "loss": 1.0392,
2084
+ "step": 346
2085
+ },
2086
+ {
2087
+ "epoch": 0.85,
2088
+ "learning_rate": 1.6788251624121335e-05,
2089
+ "loss": 1.1693,
2090
+ "step": 347
2091
+ },
2092
+ {
2093
+ "epoch": 0.86,
2094
+ "learning_rate": 1.6768694426529432e-05,
2095
+ "loss": 1.1036,
2096
+ "step": 348
2097
+ },
2098
+ {
2099
+ "epoch": 0.86,
2100
+ "learning_rate": 1.6749089332346714e-05,
2101
+ "loss": 1.0874,
2102
+ "step": 349
2103
+ },
2104
+ {
2105
+ "epoch": 0.86,
2106
+ "learning_rate": 1.672943648030261e-05,
2107
+ "loss": 1.1224,
2108
+ "step": 350
2109
+ },
2110
+ {
2111
+ "epoch": 0.86,
2112
+ "learning_rate": 1.6709736009464504e-05,
2113
+ "loss": 1.0791,
2114
+ "step": 351
2115
+ },
2116
+ {
2117
+ "epoch": 0.87,
2118
+ "learning_rate": 1.668998805923675e-05,
2119
+ "loss": 1.0822,
2120
+ "step": 352
2121
+ },
2122
+ {
2123
+ "epoch": 0.87,
2124
+ "learning_rate": 1.6670192769359643e-05,
2125
+ "loss": 1.0818,
2126
+ "step": 353
2127
+ },
2128
+ {
2129
+ "epoch": 0.87,
2130
+ "learning_rate": 1.6650350279908497e-05,
2131
+ "loss": 1.1335,
2132
+ "step": 354
2133
+ },
2134
+ {
2135
+ "epoch": 0.87,
2136
+ "learning_rate": 1.6630460731292597e-05,
2137
+ "loss": 1.0699,
2138
+ "step": 355
2139
+ },
2140
+ {
2141
+ "epoch": 0.88,
2142
+ "learning_rate": 1.661052426425424e-05,
2143
+ "loss": 1.1318,
2144
+ "step": 356
2145
+ },
2146
+ {
2147
+ "epoch": 0.88,
2148
+ "learning_rate": 1.6590541019867722e-05,
2149
+ "loss": 1.1362,
2150
+ "step": 357
2151
+ },
2152
+ {
2153
+ "epoch": 0.88,
2154
+ "learning_rate": 1.6570511139538348e-05,
2155
+ "loss": 1.0223,
2156
+ "step": 358
2157
+ },
2158
+ {
2159
+ "epoch": 0.88,
2160
+ "learning_rate": 1.655043476500142e-05,
2161
+ "loss": 1.0471,
2162
+ "step": 359
2163
+ },
2164
+ {
2165
+ "epoch": 0.89,
2166
+ "learning_rate": 1.6530312038321247e-05,
2167
+ "loss": 1.1665,
2168
+ "step": 360
2169
+ },
2170
+ {
2171
+ "epoch": 0.89,
2172
+ "learning_rate": 1.6510143101890136e-05,
2173
+ "loss": 1.0409,
2174
+ "step": 361
2175
+ },
2176
+ {
2177
+ "epoch": 0.89,
2178
+ "learning_rate": 1.6489928098427383e-05,
2179
+ "loss": 1.1231,
2180
+ "step": 362
2181
+ },
2182
+ {
2183
+ "epoch": 0.89,
2184
+ "learning_rate": 1.6469667170978258e-05,
2185
+ "loss": 1.0062,
2186
+ "step": 363
2187
+ },
2188
+ {
2189
+ "epoch": 0.9,
2190
+ "learning_rate": 1.6449360462913005e-05,
2191
+ "loss": 1.0764,
2192
+ "step": 364
2193
+ },
2194
+ {
2195
+ "epoch": 0.9,
2196
+ "learning_rate": 1.642900811792582e-05,
2197
+ "loss": 1.1421,
2198
+ "step": 365
2199
+ },
2200
+ {
2201
+ "epoch": 0.9,
2202
+ "learning_rate": 1.640861028003383e-05,
2203
+ "loss": 1.0809,
2204
+ "step": 366
2205
+ },
2206
+ {
2207
+ "epoch": 0.9,
2208
+ "learning_rate": 1.6388167093576083e-05,
2209
+ "loss": 1.1295,
2210
+ "step": 367
2211
+ },
2212
+ {
2213
+ "epoch": 0.91,
2214
+ "learning_rate": 1.6367678703212515e-05,
2215
+ "loss": 1.0813,
2216
+ "step": 368
2217
+ },
2218
+ {
2219
+ "epoch": 0.91,
2220
+ "learning_rate": 1.6347145253922942e-05,
2221
+ "loss": 1.1542,
2222
+ "step": 369
2223
+ },
2224
+ {
2225
+ "epoch": 0.91,
2226
+ "learning_rate": 1.632656689100602e-05,
2227
+ "loss": 1.0917,
2228
+ "step": 370
2229
+ },
2230
+ {
2231
+ "epoch": 0.91,
2232
+ "learning_rate": 1.6305943760078226e-05,
2233
+ "loss": 1.1101,
2234
+ "step": 371
2235
+ },
2236
+ {
2237
+ "epoch": 0.92,
2238
+ "learning_rate": 1.628527600707283e-05,
2239
+ "loss": 1.0929,
2240
+ "step": 372
2241
+ },
2242
+ {
2243
+ "epoch": 0.92,
2244
+ "learning_rate": 1.6264563778238834e-05,
2245
+ "loss": 1.074,
2246
+ "step": 373
2247
+ },
2248
+ {
2249
+ "epoch": 0.92,
2250
+ "learning_rate": 1.6243807220139988e-05,
2251
+ "loss": 1.0869,
2252
+ "step": 374
2253
+ },
2254
+ {
2255
+ "epoch": 0.92,
2256
+ "learning_rate": 1.6223006479653708e-05,
2257
+ "loss": 1.0546,
2258
+ "step": 375
2259
+ },
2260
+ {
2261
+ "epoch": 0.93,
2262
+ "learning_rate": 1.6202161703970057e-05,
2263
+ "loss": 1.0218,
2264
+ "step": 376
2265
+ },
2266
+ {
2267
+ "epoch": 0.93,
2268
+ "learning_rate": 1.6181273040590696e-05,
2269
+ "loss": 1.054,
2270
+ "step": 377
2271
+ },
2272
+ {
2273
+ "epoch": 0.93,
2274
+ "learning_rate": 1.616034063732785e-05,
2275
+ "loss": 0.9993,
2276
+ "step": 378
2277
+ },
2278
+ {
2279
+ "epoch": 0.93,
2280
+ "learning_rate": 1.613936464230325e-05,
2281
+ "loss": 1.0481,
2282
+ "step": 379
2283
+ },
2284
+ {
2285
+ "epoch": 0.94,
2286
+ "learning_rate": 1.6118345203947093e-05,
2287
+ "loss": 1.1722,
2288
+ "step": 380
2289
+ },
2290
+ {
2291
+ "epoch": 0.94,
2292
+ "learning_rate": 1.6097282470996997e-05,
2293
+ "loss": 1.1137,
2294
+ "step": 381
2295
+ },
2296
+ {
2297
+ "epoch": 0.94,
2298
+ "learning_rate": 1.6076176592496926e-05,
2299
+ "loss": 1.0729,
2300
+ "step": 382
2301
+ },
2302
+ {
2303
+ "epoch": 0.94,
2304
+ "learning_rate": 1.605502771779616e-05,
2305
+ "loss": 1.0825,
2306
+ "step": 383
2307
+ },
2308
+ {
2309
+ "epoch": 0.94,
2310
+ "learning_rate": 1.603383599654823e-05,
2311
+ "loss": 1.0767,
2312
+ "step": 384
2313
+ },
2314
+ {
2315
+ "epoch": 0.95,
2316
+ "learning_rate": 1.601260157870985e-05,
2317
+ "loss": 1.1263,
2318
+ "step": 385
2319
+ },
2320
+ {
2321
+ "epoch": 0.95,
2322
+ "learning_rate": 1.599132461453987e-05,
2323
+ "loss": 1.0745,
2324
+ "step": 386
2325
+ },
2326
+ {
2327
+ "epoch": 0.95,
2328
+ "learning_rate": 1.5970005254598204e-05,
2329
+ "loss": 1.1024,
2330
+ "step": 387
2331
+ },
2332
+ {
2333
+ "epoch": 0.95,
2334
+ "learning_rate": 1.594864364974476e-05,
2335
+ "loss": 1.1268,
2336
+ "step": 388
2337
+ },
2338
+ {
2339
+ "epoch": 0.96,
2340
+ "learning_rate": 1.592723995113839e-05,
2341
+ "loss": 1.0998,
2342
+ "step": 389
2343
+ },
2344
+ {
2345
+ "epoch": 0.96,
2346
+ "learning_rate": 1.5905794310235808e-05,
2347
+ "loss": 1.0569,
2348
+ "step": 390
2349
+ },
2350
+ {
2351
+ "epoch": 0.96,
2352
+ "learning_rate": 1.5884306878790512e-05,
2353
+ "loss": 1.0528,
2354
+ "step": 391
2355
+ },
2356
+ {
2357
+ "epoch": 0.96,
2358
+ "learning_rate": 1.586277780885172e-05,
2359
+ "loss": 1.0959,
2360
+ "step": 392
2361
+ },
2362
+ {
2363
+ "epoch": 0.97,
2364
+ "learning_rate": 1.58412072527633e-05,
2365
+ "loss": 1.167,
2366
+ "step": 393
2367
+ },
2368
+ {
2369
+ "epoch": 0.97,
2370
+ "learning_rate": 1.5819595363162682e-05,
2371
+ "loss": 1.11,
2372
+ "step": 394
2373
+ },
2374
+ {
2375
+ "epoch": 0.97,
2376
+ "learning_rate": 1.5797942292979767e-05,
2377
+ "loss": 1.1125,
2378
+ "step": 395
2379
+ },
2380
+ {
2381
+ "epoch": 0.97,
2382
+ "learning_rate": 1.577624819543587e-05,
2383
+ "loss": 1.1226,
2384
+ "step": 396
2385
+ },
2386
+ {
2387
+ "epoch": 0.98,
2388
+ "learning_rate": 1.5754513224042625e-05,
2389
+ "loss": 1.134,
2390
+ "step": 397
2391
+ },
2392
+ {
2393
+ "epoch": 0.98,
2394
+ "learning_rate": 1.573273753260089e-05,
2395
+ "loss": 1.1455,
2396
+ "step": 398
2397
+ },
2398
+ {
2399
+ "epoch": 0.98,
2400
+ "learning_rate": 1.571092127519967e-05,
2401
+ "loss": 1.0762,
2402
+ "step": 399
2403
+ },
2404
+ {
2405
+ "epoch": 0.98,
2406
+ "learning_rate": 1.568906460621502e-05,
2407
+ "loss": 1.0632,
2408
+ "step": 400
2409
+ }
2410
+ ],
2411
+ "max_steps": 1218,
2412
+ "num_train_epochs": 3,
2413
+ "total_flos": 9.670402784414925e+16,
2414
+ "trial_name": null,
2415
+ "trial_params": null
2416
+ }