SushantGautam commited on
Commit
03eb613
·
1 Parent(s): c0bfeaf

End of training

Browse files
Files changed (3) hide show
  1. all_results.json +8 -0
  2. train_results.json +8 -0
  3. trainer_state.json +2023 -0
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "train_loss": 3.0590674640896083,
4
+ "train_runtime": 25.7247,
5
+ "train_samples": 3527,
6
+ "train_samples_per_second": 411.317,
7
+ "train_steps_per_second": 12.945
8
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "train_loss": 3.0590674640896083,
4
+ "train_runtime": 25.7247,
5
+ "train_samples": 3527,
6
+ "train_samples_per_second": 411.317,
7
+ "train_steps_per_second": 12.945
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,2023 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.0,
5
+ "global_step": 333,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 1.993993993993994e-05,
13
+ "loss": 7.239,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.02,
18
+ "learning_rate": 1.987987987987988e-05,
19
+ "loss": 6.4074,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.03,
24
+ "learning_rate": 1.981981981981982e-05,
25
+ "loss": 6.4224,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.04,
30
+ "learning_rate": 1.9759759759759763e-05,
31
+ "loss": 5.6383,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.05,
36
+ "learning_rate": 1.9699699699699702e-05,
37
+ "loss": 5.4807,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.05,
42
+ "learning_rate": 1.963963963963964e-05,
43
+ "loss": 5.4554,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.06,
48
+ "learning_rate": 1.957957957957958e-05,
49
+ "loss": 5.2014,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.07,
54
+ "learning_rate": 1.951951951951952e-05,
55
+ "loss": 5.3549,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.08,
60
+ "learning_rate": 1.9459459459459463e-05,
61
+ "loss": 5.0557,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.09,
66
+ "learning_rate": 1.9399399399399402e-05,
67
+ "loss": 4.7239,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.1,
72
+ "learning_rate": 1.9339339339339342e-05,
73
+ "loss": 4.5006,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.11,
78
+ "learning_rate": 1.927927927927928e-05,
79
+ "loss": 4.6071,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.12,
84
+ "learning_rate": 1.921921921921922e-05,
85
+ "loss": 4.6655,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.13,
90
+ "learning_rate": 1.915915915915916e-05,
91
+ "loss": 4.3521,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.14,
96
+ "learning_rate": 1.90990990990991e-05,
97
+ "loss": 4.8466,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.14,
102
+ "learning_rate": 1.903903903903904e-05,
103
+ "loss": 4.5095,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.15,
108
+ "learning_rate": 1.8978978978978982e-05,
109
+ "loss": 4.4407,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.16,
114
+ "learning_rate": 1.891891891891892e-05,
115
+ "loss": 4.4312,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.17,
120
+ "learning_rate": 1.885885885885886e-05,
121
+ "loss": 4.314,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.18,
126
+ "learning_rate": 1.87987987987988e-05,
127
+ "loss": 4.274,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.19,
132
+ "learning_rate": 1.873873873873874e-05,
133
+ "loss": 4.3241,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.2,
138
+ "learning_rate": 1.8678678678678682e-05,
139
+ "loss": 4.3871,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.21,
144
+ "learning_rate": 1.861861861861862e-05,
145
+ "loss": 4.0151,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.22,
150
+ "learning_rate": 1.855855855855856e-05,
151
+ "loss": 4.3852,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.23,
156
+ "learning_rate": 1.84984984984985e-05,
157
+ "loss": 4.2063,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.23,
162
+ "learning_rate": 1.843843843843844e-05,
163
+ "loss": 4.327,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.24,
168
+ "learning_rate": 1.8378378378378383e-05,
169
+ "loss": 3.8483,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.25,
174
+ "learning_rate": 1.831831831831832e-05,
175
+ "loss": 4.0045,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.26,
180
+ "learning_rate": 1.8258258258258258e-05,
181
+ "loss": 4.1035,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.27,
186
+ "learning_rate": 1.81981981981982e-05,
187
+ "loss": 4.0994,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.28,
192
+ "learning_rate": 1.813813813813814e-05,
193
+ "loss": 3.8518,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.29,
198
+ "learning_rate": 1.807807807807808e-05,
199
+ "loss": 3.8938,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.3,
204
+ "learning_rate": 1.801801801801802e-05,
205
+ "loss": 3.7891,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.31,
210
+ "learning_rate": 1.795795795795796e-05,
211
+ "loss": 3.978,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.32,
216
+ "learning_rate": 1.78978978978979e-05,
217
+ "loss": 3.9747,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.32,
222
+ "learning_rate": 1.783783783783784e-05,
223
+ "loss": 3.8688,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.33,
228
+ "learning_rate": 1.7777777777777777e-05,
229
+ "loss": 3.8895,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.34,
234
+ "learning_rate": 1.771771771771772e-05,
235
+ "loss": 3.7801,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.35,
240
+ "learning_rate": 1.765765765765766e-05,
241
+ "loss": 3.7968,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.36,
246
+ "learning_rate": 1.7597597597597598e-05,
247
+ "loss": 3.8245,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.37,
252
+ "learning_rate": 1.7537537537537538e-05,
253
+ "loss": 3.7415,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.38,
258
+ "learning_rate": 1.7477477477477477e-05,
259
+ "loss": 3.5251,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.39,
264
+ "learning_rate": 1.741741741741742e-05,
265
+ "loss": 3.667,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.4,
270
+ "learning_rate": 1.735735735735736e-05,
271
+ "loss": 3.6625,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.41,
276
+ "learning_rate": 1.72972972972973e-05,
277
+ "loss": 3.675,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 0.41,
282
+ "learning_rate": 1.7237237237237238e-05,
283
+ "loss": 3.8588,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 0.42,
288
+ "learning_rate": 1.7177177177177177e-05,
289
+ "loss": 3.5836,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 0.43,
294
+ "learning_rate": 1.711711711711712e-05,
295
+ "loss": 3.7319,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 0.44,
300
+ "learning_rate": 1.705705705705706e-05,
301
+ "loss": 3.5722,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 0.45,
306
+ "learning_rate": 1.6996996996997e-05,
307
+ "loss": 3.573,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 0.46,
312
+ "learning_rate": 1.693693693693694e-05,
313
+ "loss": 3.5678,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 0.47,
318
+ "learning_rate": 1.6876876876876878e-05,
319
+ "loss": 3.6492,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 0.48,
324
+ "learning_rate": 1.6816816816816817e-05,
325
+ "loss": 3.6264,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 0.49,
330
+ "learning_rate": 1.6756756756756757e-05,
331
+ "loss": 3.5554,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 0.5,
336
+ "learning_rate": 1.6696696696696696e-05,
337
+ "loss": 3.4947,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 0.5,
342
+ "learning_rate": 1.663663663663664e-05,
343
+ "loss": 3.2952,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 0.51,
348
+ "learning_rate": 1.6576576576576578e-05,
349
+ "loss": 3.4397,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 0.52,
354
+ "learning_rate": 1.6516516516516518e-05,
355
+ "loss": 3.6218,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 0.53,
360
+ "learning_rate": 1.6456456456456457e-05,
361
+ "loss": 3.6569,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 0.54,
366
+ "learning_rate": 1.6396396396396396e-05,
367
+ "loss": 3.2034,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 0.55,
372
+ "learning_rate": 1.633633633633634e-05,
373
+ "loss": 3.3751,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 0.56,
378
+ "learning_rate": 1.627627627627628e-05,
379
+ "loss": 3.4071,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 0.57,
384
+ "learning_rate": 1.6216216216216218e-05,
385
+ "loss": 3.3431,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 0.58,
390
+ "learning_rate": 1.6156156156156157e-05,
391
+ "loss": 3.2513,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 0.59,
396
+ "learning_rate": 1.6096096096096097e-05,
397
+ "loss": 3.4148,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 0.59,
402
+ "learning_rate": 1.6036036036036036e-05,
403
+ "loss": 3.3335,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 0.6,
408
+ "learning_rate": 1.5975975975975976e-05,
409
+ "loss": 3.3857,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 0.61,
414
+ "learning_rate": 1.591591591591592e-05,
415
+ "loss": 3.1585,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 0.62,
420
+ "learning_rate": 1.5855855855855858e-05,
421
+ "loss": 3.4215,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 0.63,
426
+ "learning_rate": 1.5795795795795797e-05,
427
+ "loss": 3.3944,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 0.64,
432
+ "learning_rate": 1.5735735735735737e-05,
433
+ "loss": 3.1508,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 0.65,
438
+ "learning_rate": 1.5675675675675676e-05,
439
+ "loss": 3.47,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 0.66,
444
+ "learning_rate": 1.5615615615615616e-05,
445
+ "loss": 3.1349,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 0.67,
450
+ "learning_rate": 1.555555555555556e-05,
451
+ "loss": 3.0709,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 0.68,
456
+ "learning_rate": 1.5495495495495498e-05,
457
+ "loss": 3.0129,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 0.68,
462
+ "learning_rate": 1.5435435435435437e-05,
463
+ "loss": 3.0602,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 0.69,
468
+ "learning_rate": 1.5375375375375377e-05,
469
+ "loss": 3.1955,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 0.7,
474
+ "learning_rate": 1.5315315315315316e-05,
475
+ "loss": 3.0284,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 0.71,
480
+ "learning_rate": 1.5255255255255257e-05,
481
+ "loss": 3.0934,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 0.72,
486
+ "learning_rate": 1.5195195195195196e-05,
487
+ "loss": 3.1247,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 0.73,
492
+ "learning_rate": 1.5135135135135138e-05,
493
+ "loss": 3.2246,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 0.74,
498
+ "learning_rate": 1.5075075075075077e-05,
499
+ "loss": 3.1494,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 0.75,
504
+ "learning_rate": 1.5015015015015015e-05,
505
+ "loss": 3.2155,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 0.76,
510
+ "learning_rate": 1.4954954954954957e-05,
511
+ "loss": 3.0095,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 0.77,
516
+ "learning_rate": 1.4894894894894895e-05,
517
+ "loss": 3.0313,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 0.77,
522
+ "learning_rate": 1.4834834834834836e-05,
523
+ "loss": 3.2005,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 0.78,
528
+ "learning_rate": 1.4774774774774776e-05,
529
+ "loss": 3.1127,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 0.79,
534
+ "learning_rate": 1.4714714714714715e-05,
535
+ "loss": 3.1082,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 0.8,
540
+ "learning_rate": 1.4654654654654656e-05,
541
+ "loss": 3.0101,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 0.81,
546
+ "learning_rate": 1.4594594594594596e-05,
547
+ "loss": 3.0918,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 0.82,
552
+ "learning_rate": 1.4534534534534537e-05,
553
+ "loss": 2.9955,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 0.83,
558
+ "learning_rate": 1.4474474474474476e-05,
559
+ "loss": 3.0981,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 0.84,
564
+ "learning_rate": 1.4414414414414416e-05,
565
+ "loss": 3.1136,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 0.85,
570
+ "learning_rate": 1.4354354354354357e-05,
571
+ "loss": 3.2654,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 0.86,
576
+ "learning_rate": 1.4294294294294296e-05,
577
+ "loss": 3.0342,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 0.86,
582
+ "learning_rate": 1.4234234234234234e-05,
583
+ "loss": 2.9786,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 0.87,
588
+ "learning_rate": 1.4174174174174175e-05,
589
+ "loss": 3.1309,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 0.88,
594
+ "learning_rate": 1.4114114114114114e-05,
595
+ "loss": 3.178,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 0.89,
600
+ "learning_rate": 1.4054054054054055e-05,
601
+ "loss": 2.9293,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 0.9,
606
+ "learning_rate": 1.3993993993993995e-05,
607
+ "loss": 3.1669,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 0.91,
612
+ "learning_rate": 1.3933933933933934e-05,
613
+ "loss": 2.8725,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 0.92,
618
+ "learning_rate": 1.3873873873873875e-05,
619
+ "loss": 3.0613,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 0.93,
624
+ "learning_rate": 1.3813813813813815e-05,
625
+ "loss": 2.8206,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 0.94,
630
+ "learning_rate": 1.3753753753753756e-05,
631
+ "loss": 2.9075,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 0.95,
636
+ "learning_rate": 1.3693693693693695e-05,
637
+ "loss": 2.9944,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 0.95,
642
+ "learning_rate": 1.3633633633633635e-05,
643
+ "loss": 2.955,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 0.96,
648
+ "learning_rate": 1.3573573573573576e-05,
649
+ "loss": 3.0356,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 0.97,
654
+ "learning_rate": 1.3513513513513515e-05,
655
+ "loss": 2.7716,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 0.98,
660
+ "learning_rate": 1.3453453453453456e-05,
661
+ "loss": 2.8735,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 0.99,
666
+ "learning_rate": 1.3393393393393394e-05,
667
+ "loss": 2.9966,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 1.0,
672
+ "learning_rate": 1.3333333333333333e-05,
673
+ "loss": 3.0701,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 1.01,
678
+ "learning_rate": 1.3273273273273274e-05,
679
+ "loss": 2.8179,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 1.02,
684
+ "learning_rate": 1.3213213213213214e-05,
685
+ "loss": 2.8501,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 1.03,
690
+ "learning_rate": 1.3153153153153155e-05,
691
+ "loss": 2.7673,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 1.04,
696
+ "learning_rate": 1.3093093093093094e-05,
697
+ "loss": 2.9732,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 1.05,
702
+ "learning_rate": 1.3033033033033034e-05,
703
+ "loss": 2.8297,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 1.05,
708
+ "learning_rate": 1.2972972972972975e-05,
709
+ "loss": 2.7962,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 1.06,
714
+ "learning_rate": 1.2912912912912914e-05,
715
+ "loss": 3.0255,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 1.07,
720
+ "learning_rate": 1.2852852852852854e-05,
721
+ "loss": 2.9122,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 1.08,
726
+ "learning_rate": 1.2792792792792795e-05,
727
+ "loss": 2.9703,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 1.09,
732
+ "learning_rate": 1.2732732732732732e-05,
733
+ "loss": 2.7686,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 1.1,
738
+ "learning_rate": 1.2672672672672675e-05,
739
+ "loss": 2.9087,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 1.11,
744
+ "learning_rate": 1.2612612612612613e-05,
745
+ "loss": 3.0105,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 1.12,
750
+ "learning_rate": 1.2552552552552552e-05,
751
+ "loss": 2.9715,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 1.13,
756
+ "learning_rate": 1.2492492492492493e-05,
757
+ "loss": 2.9058,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 1.14,
762
+ "learning_rate": 1.2432432432432433e-05,
763
+ "loss": 2.7668,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 1.14,
768
+ "learning_rate": 1.2372372372372374e-05,
769
+ "loss": 3.0009,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 1.15,
774
+ "learning_rate": 1.2312312312312313e-05,
775
+ "loss": 2.7362,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 1.16,
780
+ "learning_rate": 1.2252252252252253e-05,
781
+ "loss": 2.8804,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 1.17,
786
+ "learning_rate": 1.2192192192192194e-05,
787
+ "loss": 2.8841,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 1.18,
792
+ "learning_rate": 1.2132132132132133e-05,
793
+ "loss": 2.8075,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 1.19,
798
+ "learning_rate": 1.2072072072072074e-05,
799
+ "loss": 2.7388,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 1.2,
804
+ "learning_rate": 1.2012012012012014e-05,
805
+ "loss": 2.8681,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 1.21,
810
+ "learning_rate": 1.1951951951951951e-05,
811
+ "loss": 2.7851,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 1.22,
816
+ "learning_rate": 1.1891891891891894e-05,
817
+ "loss": 2.9474,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 1.23,
822
+ "learning_rate": 1.1831831831831832e-05,
823
+ "loss": 2.7493,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 1.23,
828
+ "learning_rate": 1.1771771771771771e-05,
829
+ "loss": 2.8871,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 1.24,
834
+ "learning_rate": 1.1711711711711713e-05,
835
+ "loss": 2.9732,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 1.25,
840
+ "learning_rate": 1.1651651651651652e-05,
841
+ "loss": 2.9281,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 1.26,
846
+ "learning_rate": 1.1591591591591593e-05,
847
+ "loss": 2.8014,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 1.27,
852
+ "learning_rate": 1.1531531531531532e-05,
853
+ "loss": 2.8444,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 1.28,
858
+ "learning_rate": 1.1471471471471472e-05,
859
+ "loss": 2.994,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 1.29,
864
+ "learning_rate": 1.1411411411411413e-05,
865
+ "loss": 2.9153,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 1.3,
870
+ "learning_rate": 1.1351351351351352e-05,
871
+ "loss": 2.7083,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 1.31,
876
+ "learning_rate": 1.1291291291291293e-05,
877
+ "loss": 2.8247,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 1.32,
882
+ "learning_rate": 1.1231231231231233e-05,
883
+ "loss": 2.6741,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 1.32,
888
+ "learning_rate": 1.117117117117117e-05,
889
+ "loss": 2.8198,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 1.33,
894
+ "learning_rate": 1.1111111111111113e-05,
895
+ "loss": 2.9513,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 1.34,
900
+ "learning_rate": 1.1051051051051051e-05,
901
+ "loss": 2.9997,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 1.35,
906
+ "learning_rate": 1.0990990990990992e-05,
907
+ "loss": 2.6673,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 1.36,
912
+ "learning_rate": 1.0930930930930932e-05,
913
+ "loss": 2.8524,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 1.37,
918
+ "learning_rate": 1.0870870870870871e-05,
919
+ "loss": 2.8598,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 1.38,
924
+ "learning_rate": 1.0810810810810812e-05,
925
+ "loss": 2.8718,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 1.39,
930
+ "learning_rate": 1.0750750750750751e-05,
931
+ "loss": 2.782,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 1.4,
936
+ "learning_rate": 1.0690690690690693e-05,
937
+ "loss": 2.8597,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 1.41,
942
+ "learning_rate": 1.0630630630630632e-05,
943
+ "loss": 2.842,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 1.41,
948
+ "learning_rate": 1.0570570570570571e-05,
949
+ "loss": 2.8293,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 1.42,
954
+ "learning_rate": 1.0510510510510512e-05,
955
+ "loss": 2.7173,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 1.43,
960
+ "learning_rate": 1.0450450450450452e-05,
961
+ "loss": 2.723,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 1.44,
966
+ "learning_rate": 1.039039039039039e-05,
967
+ "loss": 2.8006,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 1.45,
972
+ "learning_rate": 1.033033033033033e-05,
973
+ "loss": 2.75,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 1.46,
978
+ "learning_rate": 1.027027027027027e-05,
979
+ "loss": 2.5134,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 1.47,
984
+ "learning_rate": 1.0210210210210211e-05,
985
+ "loss": 2.8993,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 1.48,
990
+ "learning_rate": 1.015015015015015e-05,
991
+ "loss": 2.8422,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 1.49,
996
+ "learning_rate": 1.009009009009009e-05,
997
+ "loss": 2.513,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 1.5,
1002
+ "learning_rate": 1.0030030030030031e-05,
1003
+ "loss": 2.7341,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 1.5,
1008
+ "learning_rate": 9.96996996996997e-06,
1009
+ "loss": 2.9636,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 1.51,
1014
+ "learning_rate": 9.90990990990991e-06,
1015
+ "loss": 2.7541,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 1.52,
1020
+ "learning_rate": 9.849849849849851e-06,
1021
+ "loss": 3.0172,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 1.53,
1026
+ "learning_rate": 9.78978978978979e-06,
1027
+ "loss": 2.7282,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 1.54,
1032
+ "learning_rate": 9.729729729729732e-06,
1033
+ "loss": 2.7824,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 1.55,
1038
+ "learning_rate": 9.669669669669671e-06,
1039
+ "loss": 2.7859,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 1.56,
1044
+ "learning_rate": 9.60960960960961e-06,
1045
+ "loss": 2.6787,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 1.57,
1050
+ "learning_rate": 9.54954954954955e-06,
1051
+ "loss": 2.6765,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 1.58,
1056
+ "learning_rate": 9.489489489489491e-06,
1057
+ "loss": 2.6947,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 1.59,
1062
+ "learning_rate": 9.42942942942943e-06,
1063
+ "loss": 2.8773,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 1.59,
1068
+ "learning_rate": 9.36936936936937e-06,
1069
+ "loss": 2.7026,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 1.6,
1074
+ "learning_rate": 9.30930930930931e-06,
1075
+ "loss": 2.9494,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 1.61,
1080
+ "learning_rate": 9.24924924924925e-06,
1081
+ "loss": 2.6875,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 1.62,
1086
+ "learning_rate": 9.189189189189191e-06,
1087
+ "loss": 2.663,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 1.63,
1092
+ "learning_rate": 9.129129129129129e-06,
1093
+ "loss": 2.6553,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 1.64,
1098
+ "learning_rate": 9.06906906906907e-06,
1099
+ "loss": 2.8712,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 1.65,
1104
+ "learning_rate": 9.00900900900901e-06,
1105
+ "loss": 3.0305,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 1.66,
1110
+ "learning_rate": 8.94894894894895e-06,
1111
+ "loss": 3.1108,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 1.67,
1116
+ "learning_rate": 8.888888888888888e-06,
1117
+ "loss": 2.8379,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 1.68,
1122
+ "learning_rate": 8.82882882882883e-06,
1123
+ "loss": 2.7338,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 1.68,
1128
+ "learning_rate": 8.768768768768769e-06,
1129
+ "loss": 2.5357,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 1.69,
1134
+ "learning_rate": 8.70870870870871e-06,
1135
+ "loss": 2.6869,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 1.7,
1140
+ "learning_rate": 8.64864864864865e-06,
1141
+ "loss": 2.6221,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 1.71,
1146
+ "learning_rate": 8.588588588588589e-06,
1147
+ "loss": 2.7291,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 1.72,
1152
+ "learning_rate": 8.52852852852853e-06,
1153
+ "loss": 2.7433,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 1.73,
1158
+ "learning_rate": 8.46846846846847e-06,
1159
+ "loss": 2.742,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 1.74,
1164
+ "learning_rate": 8.408408408408409e-06,
1165
+ "loss": 2.7337,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 1.75,
1170
+ "learning_rate": 8.348348348348348e-06,
1171
+ "loss": 2.7747,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 1.76,
1176
+ "learning_rate": 8.288288288288289e-06,
1177
+ "loss": 2.708,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 1.77,
1182
+ "learning_rate": 8.228228228228229e-06,
1183
+ "loss": 3.1017,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 1.77,
1188
+ "learning_rate": 8.16816816816817e-06,
1189
+ "loss": 2.8754,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 1.78,
1194
+ "learning_rate": 8.108108108108109e-06,
1195
+ "loss": 2.8458,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 1.79,
1200
+ "learning_rate": 8.048048048048048e-06,
1201
+ "loss": 2.5941,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 1.8,
1206
+ "learning_rate": 7.987987987987988e-06,
1207
+ "loss": 2.5742,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 1.81,
1212
+ "learning_rate": 7.927927927927929e-06,
1213
+ "loss": 2.8937,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 1.82,
1218
+ "learning_rate": 7.867867867867868e-06,
1219
+ "loss": 2.9912,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 1.83,
1224
+ "learning_rate": 7.807807807807808e-06,
1225
+ "loss": 2.762,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 1.84,
1230
+ "learning_rate": 7.747747747747749e-06,
1231
+ "loss": 2.6746,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 1.85,
1236
+ "learning_rate": 7.687687687687688e-06,
1237
+ "loss": 2.6511,
1238
+ "step": 205
1239
+ },
1240
+ {
1241
+ "epoch": 1.86,
1242
+ "learning_rate": 7.6276276276276285e-06,
1243
+ "loss": 2.5135,
1244
+ "step": 206
1245
+ },
1246
+ {
1247
+ "epoch": 1.86,
1248
+ "learning_rate": 7.567567567567569e-06,
1249
+ "loss": 2.4414,
1250
+ "step": 207
1251
+ },
1252
+ {
1253
+ "epoch": 1.87,
1254
+ "learning_rate": 7.507507507507507e-06,
1255
+ "loss": 2.7932,
1256
+ "step": 208
1257
+ },
1258
+ {
1259
+ "epoch": 1.88,
1260
+ "learning_rate": 7.447447447447448e-06,
1261
+ "loss": 3.0052,
1262
+ "step": 209
1263
+ },
1264
+ {
1265
+ "epoch": 1.89,
1266
+ "learning_rate": 7.387387387387388e-06,
1267
+ "loss": 2.8108,
1268
+ "step": 210
1269
+ },
1270
+ {
1271
+ "epoch": 1.9,
1272
+ "learning_rate": 7.327327327327328e-06,
1273
+ "loss": 2.7098,
1274
+ "step": 211
1275
+ },
1276
+ {
1277
+ "epoch": 1.91,
1278
+ "learning_rate": 7.267267267267268e-06,
1279
+ "loss": 2.7694,
1280
+ "step": 212
1281
+ },
1282
+ {
1283
+ "epoch": 1.92,
1284
+ "learning_rate": 7.207207207207208e-06,
1285
+ "loss": 2.6369,
1286
+ "step": 213
1287
+ },
1288
+ {
1289
+ "epoch": 1.93,
1290
+ "learning_rate": 7.147147147147148e-06,
1291
+ "loss": 2.6996,
1292
+ "step": 214
1293
+ },
1294
+ {
1295
+ "epoch": 1.94,
1296
+ "learning_rate": 7.087087087087087e-06,
1297
+ "loss": 2.8253,
1298
+ "step": 215
1299
+ },
1300
+ {
1301
+ "epoch": 1.95,
1302
+ "learning_rate": 7.027027027027028e-06,
1303
+ "loss": 2.5087,
1304
+ "step": 216
1305
+ },
1306
+ {
1307
+ "epoch": 1.95,
1308
+ "learning_rate": 6.966966966966967e-06,
1309
+ "loss": 2.7107,
1310
+ "step": 217
1311
+ },
1312
+ {
1313
+ "epoch": 1.96,
1314
+ "learning_rate": 6.906906906906907e-06,
1315
+ "loss": 2.6624,
1316
+ "step": 218
1317
+ },
1318
+ {
1319
+ "epoch": 1.97,
1320
+ "learning_rate": 6.846846846846848e-06,
1321
+ "loss": 2.5835,
1322
+ "step": 219
1323
+ },
1324
+ {
1325
+ "epoch": 1.98,
1326
+ "learning_rate": 6.786786786786788e-06,
1327
+ "loss": 2.6851,
1328
+ "step": 220
1329
+ },
1330
+ {
1331
+ "epoch": 1.99,
1332
+ "learning_rate": 6.726726726726728e-06,
1333
+ "loss": 2.7938,
1334
+ "step": 221
1335
+ },
1336
+ {
1337
+ "epoch": 2.0,
1338
+ "learning_rate": 6.666666666666667e-06,
1339
+ "loss": 2.4996,
1340
+ "step": 222
1341
+ },
1342
+ {
1343
+ "epoch": 2.01,
1344
+ "learning_rate": 6.606606606606607e-06,
1345
+ "loss": 2.62,
1346
+ "step": 223
1347
+ },
1348
+ {
1349
+ "epoch": 2.02,
1350
+ "learning_rate": 6.546546546546547e-06,
1351
+ "loss": 2.7648,
1352
+ "step": 224
1353
+ },
1354
+ {
1355
+ "epoch": 2.03,
1356
+ "learning_rate": 6.486486486486487e-06,
1357
+ "loss": 2.8287,
1358
+ "step": 225
1359
+ },
1360
+ {
1361
+ "epoch": 2.04,
1362
+ "learning_rate": 6.426426426426427e-06,
1363
+ "loss": 2.6175,
1364
+ "step": 226
1365
+ },
1366
+ {
1367
+ "epoch": 2.05,
1368
+ "learning_rate": 6.366366366366366e-06,
1369
+ "loss": 2.4404,
1370
+ "step": 227
1371
+ },
1372
+ {
1373
+ "epoch": 2.05,
1374
+ "learning_rate": 6.3063063063063065e-06,
1375
+ "loss": 2.5263,
1376
+ "step": 228
1377
+ },
1378
+ {
1379
+ "epoch": 2.06,
1380
+ "learning_rate": 6.246246246246247e-06,
1381
+ "loss": 2.9257,
1382
+ "step": 229
1383
+ },
1384
+ {
1385
+ "epoch": 2.07,
1386
+ "learning_rate": 6.186186186186187e-06,
1387
+ "loss": 2.9413,
1388
+ "step": 230
1389
+ },
1390
+ {
1391
+ "epoch": 2.08,
1392
+ "learning_rate": 6.126126126126126e-06,
1393
+ "loss": 2.5352,
1394
+ "step": 231
1395
+ },
1396
+ {
1397
+ "epoch": 2.09,
1398
+ "learning_rate": 6.066066066066067e-06,
1399
+ "loss": 2.6219,
1400
+ "step": 232
1401
+ },
1402
+ {
1403
+ "epoch": 2.1,
1404
+ "learning_rate": 6.006006006006007e-06,
1405
+ "loss": 2.6584,
1406
+ "step": 233
1407
+ },
1408
+ {
1409
+ "epoch": 2.11,
1410
+ "learning_rate": 5.945945945945947e-06,
1411
+ "loss": 2.7308,
1412
+ "step": 234
1413
+ },
1414
+ {
1415
+ "epoch": 2.12,
1416
+ "learning_rate": 5.885885885885886e-06,
1417
+ "loss": 2.5438,
1418
+ "step": 235
1419
+ },
1420
+ {
1421
+ "epoch": 2.13,
1422
+ "learning_rate": 5.825825825825826e-06,
1423
+ "loss": 2.5946,
1424
+ "step": 236
1425
+ },
1426
+ {
1427
+ "epoch": 2.14,
1428
+ "learning_rate": 5.765765765765766e-06,
1429
+ "loss": 2.597,
1430
+ "step": 237
1431
+ },
1432
+ {
1433
+ "epoch": 2.14,
1434
+ "learning_rate": 5.7057057057057065e-06,
1435
+ "loss": 2.6751,
1436
+ "step": 238
1437
+ },
1438
+ {
1439
+ "epoch": 2.15,
1440
+ "learning_rate": 5.645645645645647e-06,
1441
+ "loss": 2.7518,
1442
+ "step": 239
1443
+ },
1444
+ {
1445
+ "epoch": 2.16,
1446
+ "learning_rate": 5.585585585585585e-06,
1447
+ "loss": 2.6249,
1448
+ "step": 240
1449
+ },
1450
+ {
1451
+ "epoch": 2.17,
1452
+ "learning_rate": 5.5255255255255255e-06,
1453
+ "loss": 2.7765,
1454
+ "step": 241
1455
+ },
1456
+ {
1457
+ "epoch": 2.18,
1458
+ "learning_rate": 5.465465465465466e-06,
1459
+ "loss": 2.7938,
1460
+ "step": 242
1461
+ },
1462
+ {
1463
+ "epoch": 2.19,
1464
+ "learning_rate": 5.405405405405406e-06,
1465
+ "loss": 2.3933,
1466
+ "step": 243
1467
+ },
1468
+ {
1469
+ "epoch": 2.2,
1470
+ "learning_rate": 5.345345345345346e-06,
1471
+ "loss": 2.8995,
1472
+ "step": 244
1473
+ },
1474
+ {
1475
+ "epoch": 2.21,
1476
+ "learning_rate": 5.285285285285286e-06,
1477
+ "loss": 2.7298,
1478
+ "step": 245
1479
+ },
1480
+ {
1481
+ "epoch": 2.22,
1482
+ "learning_rate": 5.225225225225226e-06,
1483
+ "loss": 2.5716,
1484
+ "step": 246
1485
+ },
1486
+ {
1487
+ "epoch": 2.23,
1488
+ "learning_rate": 5.165165165165165e-06,
1489
+ "loss": 2.8012,
1490
+ "step": 247
1491
+ },
1492
+ {
1493
+ "epoch": 2.23,
1494
+ "learning_rate": 5.105105105105106e-06,
1495
+ "loss": 2.5597,
1496
+ "step": 248
1497
+ },
1498
+ {
1499
+ "epoch": 2.24,
1500
+ "learning_rate": 5.045045045045045e-06,
1501
+ "loss": 2.5855,
1502
+ "step": 249
1503
+ },
1504
+ {
1505
+ "epoch": 2.25,
1506
+ "learning_rate": 4.984984984984985e-06,
1507
+ "loss": 2.6125,
1508
+ "step": 250
1509
+ },
1510
+ {
1511
+ "epoch": 2.26,
1512
+ "learning_rate": 4.9249249249249255e-06,
1513
+ "loss": 2.5589,
1514
+ "step": 251
1515
+ },
1516
+ {
1517
+ "epoch": 2.27,
1518
+ "learning_rate": 4.864864864864866e-06,
1519
+ "loss": 2.4177,
1520
+ "step": 252
1521
+ },
1522
+ {
1523
+ "epoch": 2.28,
1524
+ "learning_rate": 4.804804804804805e-06,
1525
+ "loss": 2.7882,
1526
+ "step": 253
1527
+ },
1528
+ {
1529
+ "epoch": 2.29,
1530
+ "learning_rate": 4.7447447447447454e-06,
1531
+ "loss": 2.6047,
1532
+ "step": 254
1533
+ },
1534
+ {
1535
+ "epoch": 2.3,
1536
+ "learning_rate": 4.684684684684685e-06,
1537
+ "loss": 2.6191,
1538
+ "step": 255
1539
+ },
1540
+ {
1541
+ "epoch": 2.31,
1542
+ "learning_rate": 4.624624624624625e-06,
1543
+ "loss": 2.5648,
1544
+ "step": 256
1545
+ },
1546
+ {
1547
+ "epoch": 2.32,
1548
+ "learning_rate": 4.5645645645645645e-06,
1549
+ "loss": 2.6584,
1550
+ "step": 257
1551
+ },
1552
+ {
1553
+ "epoch": 2.32,
1554
+ "learning_rate": 4.504504504504505e-06,
1555
+ "loss": 2.5868,
1556
+ "step": 258
1557
+ },
1558
+ {
1559
+ "epoch": 2.33,
1560
+ "learning_rate": 4.444444444444444e-06,
1561
+ "loss": 2.5987,
1562
+ "step": 259
1563
+ },
1564
+ {
1565
+ "epoch": 2.34,
1566
+ "learning_rate": 4.384384384384384e-06,
1567
+ "loss": 2.6715,
1568
+ "step": 260
1569
+ },
1570
+ {
1571
+ "epoch": 2.35,
1572
+ "learning_rate": 4.324324324324325e-06,
1573
+ "loss": 2.546,
1574
+ "step": 261
1575
+ },
1576
+ {
1577
+ "epoch": 2.36,
1578
+ "learning_rate": 4.264264264264265e-06,
1579
+ "loss": 2.7366,
1580
+ "step": 262
1581
+ },
1582
+ {
1583
+ "epoch": 2.37,
1584
+ "learning_rate": 4.204204204204204e-06,
1585
+ "loss": 2.7874,
1586
+ "step": 263
1587
+ },
1588
+ {
1589
+ "epoch": 2.38,
1590
+ "learning_rate": 4.1441441441441446e-06,
1591
+ "loss": 2.5763,
1592
+ "step": 264
1593
+ },
1594
+ {
1595
+ "epoch": 2.39,
1596
+ "learning_rate": 4.084084084084085e-06,
1597
+ "loss": 2.7863,
1598
+ "step": 265
1599
+ },
1600
+ {
1601
+ "epoch": 2.4,
1602
+ "learning_rate": 4.024024024024024e-06,
1603
+ "loss": 2.5058,
1604
+ "step": 266
1605
+ },
1606
+ {
1607
+ "epoch": 2.41,
1608
+ "learning_rate": 3.9639639639639645e-06,
1609
+ "loss": 2.623,
1610
+ "step": 267
1611
+ },
1612
+ {
1613
+ "epoch": 2.41,
1614
+ "learning_rate": 3.903903903903904e-06,
1615
+ "loss": 2.6655,
1616
+ "step": 268
1617
+ },
1618
+ {
1619
+ "epoch": 2.42,
1620
+ "learning_rate": 3.843843843843844e-06,
1621
+ "loss": 2.8604,
1622
+ "step": 269
1623
+ },
1624
+ {
1625
+ "epoch": 2.43,
1626
+ "learning_rate": 3.7837837837837844e-06,
1627
+ "loss": 2.8491,
1628
+ "step": 270
1629
+ },
1630
+ {
1631
+ "epoch": 2.44,
1632
+ "learning_rate": 3.723723723723724e-06,
1633
+ "loss": 2.6519,
1634
+ "step": 271
1635
+ },
1636
+ {
1637
+ "epoch": 2.45,
1638
+ "learning_rate": 3.663663663663664e-06,
1639
+ "loss": 2.3903,
1640
+ "step": 272
1641
+ },
1642
+ {
1643
+ "epoch": 2.46,
1644
+ "learning_rate": 3.603603603603604e-06,
1645
+ "loss": 2.5994,
1646
+ "step": 273
1647
+ },
1648
+ {
1649
+ "epoch": 2.47,
1650
+ "learning_rate": 3.5435435435435437e-06,
1651
+ "loss": 2.5526,
1652
+ "step": 274
1653
+ },
1654
+ {
1655
+ "epoch": 2.48,
1656
+ "learning_rate": 3.4834834834834835e-06,
1657
+ "loss": 2.5653,
1658
+ "step": 275
1659
+ },
1660
+ {
1661
+ "epoch": 2.49,
1662
+ "learning_rate": 3.423423423423424e-06,
1663
+ "loss": 2.5769,
1664
+ "step": 276
1665
+ },
1666
+ {
1667
+ "epoch": 2.5,
1668
+ "learning_rate": 3.363363363363364e-06,
1669
+ "loss": 2.7054,
1670
+ "step": 277
1671
+ },
1672
+ {
1673
+ "epoch": 2.5,
1674
+ "learning_rate": 3.3033033033033035e-06,
1675
+ "loss": 2.7291,
1676
+ "step": 278
1677
+ },
1678
+ {
1679
+ "epoch": 2.51,
1680
+ "learning_rate": 3.2432432432432437e-06,
1681
+ "loss": 2.5145,
1682
+ "step": 279
1683
+ },
1684
+ {
1685
+ "epoch": 2.52,
1686
+ "learning_rate": 3.183183183183183e-06,
1687
+ "loss": 2.7142,
1688
+ "step": 280
1689
+ },
1690
+ {
1691
+ "epoch": 2.53,
1692
+ "learning_rate": 3.1231231231231234e-06,
1693
+ "loss": 2.6612,
1694
+ "step": 281
1695
+ },
1696
+ {
1697
+ "epoch": 2.54,
1698
+ "learning_rate": 3.063063063063063e-06,
1699
+ "loss": 2.7143,
1700
+ "step": 282
1701
+ },
1702
+ {
1703
+ "epoch": 2.55,
1704
+ "learning_rate": 3.0030030030030034e-06,
1705
+ "loss": 2.6341,
1706
+ "step": 283
1707
+ },
1708
+ {
1709
+ "epoch": 2.56,
1710
+ "learning_rate": 2.942942942942943e-06,
1711
+ "loss": 2.5404,
1712
+ "step": 284
1713
+ },
1714
+ {
1715
+ "epoch": 2.57,
1716
+ "learning_rate": 2.882882882882883e-06,
1717
+ "loss": 2.7922,
1718
+ "step": 285
1719
+ },
1720
+ {
1721
+ "epoch": 2.58,
1722
+ "learning_rate": 2.8228228228228234e-06,
1723
+ "loss": 3.1301,
1724
+ "step": 286
1725
+ },
1726
+ {
1727
+ "epoch": 2.59,
1728
+ "learning_rate": 2.7627627627627628e-06,
1729
+ "loss": 2.8461,
1730
+ "step": 287
1731
+ },
1732
+ {
1733
+ "epoch": 2.59,
1734
+ "learning_rate": 2.702702702702703e-06,
1735
+ "loss": 2.6317,
1736
+ "step": 288
1737
+ },
1738
+ {
1739
+ "epoch": 2.6,
1740
+ "learning_rate": 2.642642642642643e-06,
1741
+ "loss": 2.7555,
1742
+ "step": 289
1743
+ },
1744
+ {
1745
+ "epoch": 2.61,
1746
+ "learning_rate": 2.5825825825825827e-06,
1747
+ "loss": 2.3511,
1748
+ "step": 290
1749
+ },
1750
+ {
1751
+ "epoch": 2.62,
1752
+ "learning_rate": 2.5225225225225225e-06,
1753
+ "loss": 2.6216,
1754
+ "step": 291
1755
+ },
1756
+ {
1757
+ "epoch": 2.63,
1758
+ "learning_rate": 2.4624624624624628e-06,
1759
+ "loss": 2.6718,
1760
+ "step": 292
1761
+ },
1762
+ {
1763
+ "epoch": 2.64,
1764
+ "learning_rate": 2.4024024024024026e-06,
1765
+ "loss": 2.6301,
1766
+ "step": 293
1767
+ },
1768
+ {
1769
+ "epoch": 2.65,
1770
+ "learning_rate": 2.3423423423423424e-06,
1771
+ "loss": 2.7771,
1772
+ "step": 294
1773
+ },
1774
+ {
1775
+ "epoch": 2.66,
1776
+ "learning_rate": 2.2822822822822822e-06,
1777
+ "loss": 2.5465,
1778
+ "step": 295
1779
+ },
1780
+ {
1781
+ "epoch": 2.67,
1782
+ "learning_rate": 2.222222222222222e-06,
1783
+ "loss": 2.5168,
1784
+ "step": 296
1785
+ },
1786
+ {
1787
+ "epoch": 2.68,
1788
+ "learning_rate": 2.1621621621621623e-06,
1789
+ "loss": 2.6376,
1790
+ "step": 297
1791
+ },
1792
+ {
1793
+ "epoch": 2.68,
1794
+ "learning_rate": 2.102102102102102e-06,
1795
+ "loss": 2.4557,
1796
+ "step": 298
1797
+ },
1798
+ {
1799
+ "epoch": 2.69,
1800
+ "learning_rate": 2.0420420420420424e-06,
1801
+ "loss": 2.5339,
1802
+ "step": 299
1803
+ },
1804
+ {
1805
+ "epoch": 2.7,
1806
+ "learning_rate": 1.9819819819819822e-06,
1807
+ "loss": 2.6562,
1808
+ "step": 300
1809
+ },
1810
+ {
1811
+ "epoch": 2.71,
1812
+ "learning_rate": 1.921921921921922e-06,
1813
+ "loss": 2.7499,
1814
+ "step": 301
1815
+ },
1816
+ {
1817
+ "epoch": 2.72,
1818
+ "learning_rate": 1.861861861861862e-06,
1819
+ "loss": 2.8368,
1820
+ "step": 302
1821
+ },
1822
+ {
1823
+ "epoch": 2.73,
1824
+ "learning_rate": 1.801801801801802e-06,
1825
+ "loss": 2.6257,
1826
+ "step": 303
1827
+ },
1828
+ {
1829
+ "epoch": 2.74,
1830
+ "learning_rate": 1.7417417417417418e-06,
1831
+ "loss": 2.6908,
1832
+ "step": 304
1833
+ },
1834
+ {
1835
+ "epoch": 2.75,
1836
+ "learning_rate": 1.681681681681682e-06,
1837
+ "loss": 2.8002,
1838
+ "step": 305
1839
+ },
1840
+ {
1841
+ "epoch": 2.76,
1842
+ "learning_rate": 1.6216216216216219e-06,
1843
+ "loss": 2.6201,
1844
+ "step": 306
1845
+ },
1846
+ {
1847
+ "epoch": 2.77,
1848
+ "learning_rate": 1.5615615615615617e-06,
1849
+ "loss": 2.6824,
1850
+ "step": 307
1851
+ },
1852
+ {
1853
+ "epoch": 2.77,
1854
+ "learning_rate": 1.5015015015015017e-06,
1855
+ "loss": 2.7193,
1856
+ "step": 308
1857
+ },
1858
+ {
1859
+ "epoch": 2.78,
1860
+ "learning_rate": 1.4414414414414416e-06,
1861
+ "loss": 2.7957,
1862
+ "step": 309
1863
+ },
1864
+ {
1865
+ "epoch": 2.79,
1866
+ "learning_rate": 1.3813813813813814e-06,
1867
+ "loss": 2.6111,
1868
+ "step": 310
1869
+ },
1870
+ {
1871
+ "epoch": 2.8,
1872
+ "learning_rate": 1.3213213213213214e-06,
1873
+ "loss": 2.7703,
1874
+ "step": 311
1875
+ },
1876
+ {
1877
+ "epoch": 2.81,
1878
+ "learning_rate": 1.2612612612612613e-06,
1879
+ "loss": 2.6248,
1880
+ "step": 312
1881
+ },
1882
+ {
1883
+ "epoch": 2.82,
1884
+ "learning_rate": 1.2012012012012013e-06,
1885
+ "loss": 2.6614,
1886
+ "step": 313
1887
+ },
1888
+ {
1889
+ "epoch": 2.83,
1890
+ "learning_rate": 1.1411411411411411e-06,
1891
+ "loss": 2.7111,
1892
+ "step": 314
1893
+ },
1894
+ {
1895
+ "epoch": 2.84,
1896
+ "learning_rate": 1.0810810810810812e-06,
1897
+ "loss": 2.3901,
1898
+ "step": 315
1899
+ },
1900
+ {
1901
+ "epoch": 2.85,
1902
+ "learning_rate": 1.0210210210210212e-06,
1903
+ "loss": 2.7881,
1904
+ "step": 316
1905
+ },
1906
+ {
1907
+ "epoch": 2.86,
1908
+ "learning_rate": 9.60960960960961e-07,
1909
+ "loss": 2.651,
1910
+ "step": 317
1911
+ },
1912
+ {
1913
+ "epoch": 2.86,
1914
+ "learning_rate": 9.00900900900901e-07,
1915
+ "loss": 2.5288,
1916
+ "step": 318
1917
+ },
1918
+ {
1919
+ "epoch": 2.87,
1920
+ "learning_rate": 8.40840840840841e-07,
1921
+ "loss": 2.8173,
1922
+ "step": 319
1923
+ },
1924
+ {
1925
+ "epoch": 2.88,
1926
+ "learning_rate": 7.807807807807808e-07,
1927
+ "loss": 2.5508,
1928
+ "step": 320
1929
+ },
1930
+ {
1931
+ "epoch": 2.89,
1932
+ "learning_rate": 7.207207207207208e-07,
1933
+ "loss": 2.6362,
1934
+ "step": 321
1935
+ },
1936
+ {
1937
+ "epoch": 2.9,
1938
+ "learning_rate": 6.606606606606607e-07,
1939
+ "loss": 2.8817,
1940
+ "step": 322
1941
+ },
1942
+ {
1943
+ "epoch": 2.91,
1944
+ "learning_rate": 6.006006006006006e-07,
1945
+ "loss": 2.7706,
1946
+ "step": 323
1947
+ },
1948
+ {
1949
+ "epoch": 2.92,
1950
+ "learning_rate": 5.405405405405406e-07,
1951
+ "loss": 2.7205,
1952
+ "step": 324
1953
+ },
1954
+ {
1955
+ "epoch": 2.93,
1956
+ "learning_rate": 4.804804804804805e-07,
1957
+ "loss": 2.8032,
1958
+ "step": 325
1959
+ },
1960
+ {
1961
+ "epoch": 2.94,
1962
+ "learning_rate": 4.204204204204205e-07,
1963
+ "loss": 2.6215,
1964
+ "step": 326
1965
+ },
1966
+ {
1967
+ "epoch": 2.95,
1968
+ "learning_rate": 3.603603603603604e-07,
1969
+ "loss": 2.831,
1970
+ "step": 327
1971
+ },
1972
+ {
1973
+ "epoch": 2.95,
1974
+ "learning_rate": 3.003003003003003e-07,
1975
+ "loss": 2.9592,
1976
+ "step": 328
1977
+ },
1978
+ {
1979
+ "epoch": 2.96,
1980
+ "learning_rate": 2.4024024024024026e-07,
1981
+ "loss": 2.4744,
1982
+ "step": 329
1983
+ },
1984
+ {
1985
+ "epoch": 2.97,
1986
+ "learning_rate": 1.801801801801802e-07,
1987
+ "loss": 2.7128,
1988
+ "step": 330
1989
+ },
1990
+ {
1991
+ "epoch": 2.98,
1992
+ "learning_rate": 1.2012012012012013e-07,
1993
+ "loss": 2.4625,
1994
+ "step": 331
1995
+ },
1996
+ {
1997
+ "epoch": 2.99,
1998
+ "learning_rate": 6.006006006006006e-08,
1999
+ "loss": 2.4875,
2000
+ "step": 332
2001
+ },
2002
+ {
2003
+ "epoch": 3.0,
2004
+ "learning_rate": 0.0,
2005
+ "loss": 2.5134,
2006
+ "step": 333
2007
+ },
2008
+ {
2009
+ "epoch": 3.0,
2010
+ "step": 333,
2011
+ "total_flos": 119723569053696.0,
2012
+ "train_loss": 3.0590674640896083,
2013
+ "train_runtime": 25.7247,
2014
+ "train_samples_per_second": 411.317,
2015
+ "train_steps_per_second": 12.945
2016
+ }
2017
+ ],
2018
+ "max_steps": 333,
2019
+ "num_train_epochs": 3,
2020
+ "total_flos": 119723569053696.0,
2021
+ "trial_name": null,
2022
+ "trial_params": null
2023
+ }