angie-chen55 commited on
Commit
f429a41
·
verified ·
1 Parent(s): 16fad27

Upload trainer_state.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. trainer_state.json +2416 -0
trainer_state.json ADDED
@@ -0,0 +1,2416 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 4.668374426299499,
5
+ "global_step": 200000,
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": 5e-06,
13
+ "loss": 9.1117,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.03,
18
+ "learning_rate": 1e-05,
19
+ "loss": 7.3389,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 0.04,
24
+ "learning_rate": 1.5e-05,
25
+ "loss": 6.8415,
26
+ "step": 1500
27
+ },
28
+ {
29
+ "epoch": 0.05,
30
+ "learning_rate": 2e-05,
31
+ "loss": 6.6819,
32
+ "step": 2000
33
+ },
34
+ {
35
+ "epoch": 0.07,
36
+ "learning_rate": 2.5e-05,
37
+ "loss": 6.5753,
38
+ "step": 2500
39
+ },
40
+ {
41
+ "epoch": 0.08,
42
+ "learning_rate": 3e-05,
43
+ "loss": 6.4883,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 0.09,
48
+ "learning_rate": 3.5e-05,
49
+ "loss": 6.421,
50
+ "step": 3500
51
+ },
52
+ {
53
+ "epoch": 0.11,
54
+ "learning_rate": 4e-05,
55
+ "loss": 6.3655,
56
+ "step": 4000
57
+ },
58
+ {
59
+ "epoch": 0.12,
60
+ "learning_rate": 4.5e-05,
61
+ "loss": 6.3167,
62
+ "step": 4500
63
+ },
64
+ {
65
+ "epoch": 0.13,
66
+ "learning_rate": 5e-05,
67
+ "loss": 6.2735,
68
+ "step": 5000
69
+ },
70
+ {
71
+ "epoch": 0.15,
72
+ "learning_rate": 5.500000000000001e-05,
73
+ "loss": 6.2371,
74
+ "step": 5500
75
+ },
76
+ {
77
+ "epoch": 0.16,
78
+ "learning_rate": 6e-05,
79
+ "loss": 6.2028,
80
+ "step": 6000
81
+ },
82
+ {
83
+ "epoch": 0.17,
84
+ "learning_rate": 6.500000000000001e-05,
85
+ "loss": 6.1733,
86
+ "step": 6500
87
+ },
88
+ {
89
+ "epoch": 0.19,
90
+ "learning_rate": 7e-05,
91
+ "loss": 6.1502,
92
+ "step": 7000
93
+ },
94
+ {
95
+ "epoch": 0.2,
96
+ "learning_rate": 7.500000000000001e-05,
97
+ "loss": 6.1308,
98
+ "step": 7500
99
+ },
100
+ {
101
+ "epoch": 0.21,
102
+ "learning_rate": 8e-05,
103
+ "loss": 6.1092,
104
+ "step": 8000
105
+ },
106
+ {
107
+ "epoch": 0.23,
108
+ "learning_rate": 8.5e-05,
109
+ "loss": 6.0945,
110
+ "step": 8500
111
+ },
112
+ {
113
+ "epoch": 0.24,
114
+ "learning_rate": 8.999000000000001e-05,
115
+ "loss": 6.0738,
116
+ "step": 9000
117
+ },
118
+ {
119
+ "epoch": 0.25,
120
+ "learning_rate": 9.499e-05,
121
+ "loss": 6.0627,
122
+ "step": 9500
123
+ },
124
+ {
125
+ "epoch": 0.27,
126
+ "learning_rate": 9.999000000000001e-05,
127
+ "loss": 6.0486,
128
+ "step": 10000
129
+ },
130
+ {
131
+ "epoch": 0.28,
132
+ "learning_rate": 9.994959595959596e-05,
133
+ "loss": 6.0388,
134
+ "step": 10500
135
+ },
136
+ {
137
+ "epoch": 0.29,
138
+ "learning_rate": 9.989919191919192e-05,
139
+ "loss": 6.0244,
140
+ "step": 11000
141
+ },
142
+ {
143
+ "epoch": 0.31,
144
+ "learning_rate": 9.984868686868687e-05,
145
+ "loss": 6.015,
146
+ "step": 11500
147
+ },
148
+ {
149
+ "epoch": 0.32,
150
+ "learning_rate": 9.979818181818183e-05,
151
+ "loss": 6.0047,
152
+ "step": 12000
153
+ },
154
+ {
155
+ "epoch": 0.33,
156
+ "learning_rate": 9.974767676767676e-05,
157
+ "loss": 5.9944,
158
+ "step": 12500
159
+ },
160
+ {
161
+ "epoch": 0.35,
162
+ "learning_rate": 9.969727272727272e-05,
163
+ "loss": 5.9915,
164
+ "step": 13000
165
+ },
166
+ {
167
+ "epoch": 0.36,
168
+ "learning_rate": 9.964676767676768e-05,
169
+ "loss": 5.9777,
170
+ "step": 13500
171
+ },
172
+ {
173
+ "epoch": 0.37,
174
+ "learning_rate": 9.959626262626263e-05,
175
+ "loss": 5.976,
176
+ "step": 14000
177
+ },
178
+ {
179
+ "epoch": 0.39,
180
+ "learning_rate": 9.954575757575758e-05,
181
+ "loss": 5.9659,
182
+ "step": 14500
183
+ },
184
+ {
185
+ "epoch": 0.4,
186
+ "learning_rate": 9.949535353535355e-05,
187
+ "loss": 5.9644,
188
+ "step": 15000
189
+ },
190
+ {
191
+ "epoch": 0.01,
192
+ "learning_rate": 9.944484848484848e-05,
193
+ "loss": 5.954,
194
+ "step": 15500
195
+ },
196
+ {
197
+ "epoch": 0.03,
198
+ "learning_rate": 9.939434343434344e-05,
199
+ "loss": 5.9498,
200
+ "step": 16000
201
+ },
202
+ {
203
+ "epoch": 0.04,
204
+ "learning_rate": 9.934383838383839e-05,
205
+ "loss": 5.946,
206
+ "step": 16500
207
+ },
208
+ {
209
+ "epoch": 0.05,
210
+ "learning_rate": 9.929343434343435e-05,
211
+ "loss": 5.941,
212
+ "step": 17000
213
+ },
214
+ {
215
+ "epoch": 0.07,
216
+ "learning_rate": 9.924292929292929e-05,
217
+ "loss": 5.9365,
218
+ "step": 17500
219
+ },
220
+ {
221
+ "epoch": 0.08,
222
+ "learning_rate": 9.919242424242424e-05,
223
+ "loss": 5.9317,
224
+ "step": 18000
225
+ },
226
+ {
227
+ "epoch": 0.09,
228
+ "learning_rate": 9.91420202020202e-05,
229
+ "loss": 5.9262,
230
+ "step": 18500
231
+ },
232
+ {
233
+ "epoch": 0.11,
234
+ "learning_rate": 9.909151515151516e-05,
235
+ "loss": 5.9223,
236
+ "step": 19000
237
+ },
238
+ {
239
+ "epoch": 0.12,
240
+ "learning_rate": 9.90410101010101e-05,
241
+ "loss": 5.918,
242
+ "step": 19500
243
+ },
244
+ {
245
+ "epoch": 0.13,
246
+ "learning_rate": 9.899050505050505e-05,
247
+ "loss": 5.9124,
248
+ "step": 20000
249
+ },
250
+ {
251
+ "epoch": 0.15,
252
+ "learning_rate": 9.894e-05,
253
+ "loss": 5.9087,
254
+ "step": 20500
255
+ },
256
+ {
257
+ "epoch": 0.16,
258
+ "learning_rate": 9.888959595959596e-05,
259
+ "loss": 5.9039,
260
+ "step": 21000
261
+ },
262
+ {
263
+ "epoch": 0.17,
264
+ "learning_rate": 9.883909090909092e-05,
265
+ "loss": 5.8978,
266
+ "step": 21500
267
+ },
268
+ {
269
+ "epoch": 0.19,
270
+ "learning_rate": 9.878858585858586e-05,
271
+ "loss": 5.8957,
272
+ "step": 22000
273
+ },
274
+ {
275
+ "epoch": 0.2,
276
+ "learning_rate": 9.873808080808081e-05,
277
+ "loss": 5.7647,
278
+ "step": 22500
279
+ },
280
+ {
281
+ "epoch": 0.21,
282
+ "learning_rate": 9.868757575757577e-05,
283
+ "loss": 5.5977,
284
+ "step": 23000
285
+ },
286
+ {
287
+ "epoch": 0.23,
288
+ "learning_rate": 9.863717171717172e-05,
289
+ "loss": 5.5037,
290
+ "step": 23500
291
+ },
292
+ {
293
+ "epoch": 0.24,
294
+ "learning_rate": 9.858666666666668e-05,
295
+ "loss": 5.2742,
296
+ "step": 24000
297
+ },
298
+ {
299
+ "epoch": 0.25,
300
+ "learning_rate": 9.853616161616162e-05,
301
+ "loss": 4.8529,
302
+ "step": 24500
303
+ },
304
+ {
305
+ "epoch": 0.27,
306
+ "learning_rate": 9.848565656565657e-05,
307
+ "loss": 4.462,
308
+ "step": 25000
309
+ },
310
+ {
311
+ "epoch": 0.28,
312
+ "learning_rate": 9.843515151515153e-05,
313
+ "loss": 3.8976,
314
+ "step": 25500
315
+ },
316
+ {
317
+ "epoch": 0.29,
318
+ "learning_rate": 9.838464646464647e-05,
319
+ "loss": 3.5281,
320
+ "step": 26000
321
+ },
322
+ {
323
+ "epoch": 0.31,
324
+ "learning_rate": 9.833424242424243e-05,
325
+ "loss": 3.3119,
326
+ "step": 26500
327
+ },
328
+ {
329
+ "epoch": 0.32,
330
+ "learning_rate": 9.82838383838384e-05,
331
+ "loss": 3.1643,
332
+ "step": 27000
333
+ },
334
+ {
335
+ "epoch": 0.33,
336
+ "learning_rate": 9.823333333333333e-05,
337
+ "loss": 3.0554,
338
+ "step": 27500
339
+ },
340
+ {
341
+ "epoch": 0.35,
342
+ "learning_rate": 9.81829292929293e-05,
343
+ "loss": 2.9745,
344
+ "step": 28000
345
+ },
346
+ {
347
+ "epoch": 0.36,
348
+ "learning_rate": 9.813242424242425e-05,
349
+ "loss": 2.9042,
350
+ "step": 28500
351
+ },
352
+ {
353
+ "epoch": 0.37,
354
+ "learning_rate": 9.808191919191919e-05,
355
+ "loss": 2.8514,
356
+ "step": 29000
357
+ },
358
+ {
359
+ "epoch": 0.39,
360
+ "learning_rate": 9.803141414141415e-05,
361
+ "loss": 2.7973,
362
+ "step": 29500
363
+ },
364
+ {
365
+ "epoch": 0.4,
366
+ "learning_rate": 9.79809090909091e-05,
367
+ "loss": 2.7575,
368
+ "step": 30000
369
+ },
370
+ {
371
+ "epoch": 0.41,
372
+ "learning_rate": 9.793040404040405e-05,
373
+ "loss": 2.7213,
374
+ "step": 30500
375
+ },
376
+ {
377
+ "epoch": 0.43,
378
+ "learning_rate": 9.7879898989899e-05,
379
+ "loss": 2.6879,
380
+ "step": 31000
381
+ },
382
+ {
383
+ "epoch": 0.44,
384
+ "learning_rate": 9.782939393939394e-05,
385
+ "loss": 2.6595,
386
+ "step": 31500
387
+ },
388
+ {
389
+ "epoch": 0.45,
390
+ "learning_rate": 9.77788888888889e-05,
391
+ "loss": 2.6326,
392
+ "step": 32000
393
+ },
394
+ {
395
+ "epoch": 0.47,
396
+ "learning_rate": 9.772848484848486e-05,
397
+ "loss": 2.6054,
398
+ "step": 32500
399
+ },
400
+ {
401
+ "epoch": 0.48,
402
+ "learning_rate": 9.76779797979798e-05,
403
+ "loss": 2.5839,
404
+ "step": 33000
405
+ },
406
+ {
407
+ "epoch": 0.49,
408
+ "learning_rate": 9.762747474747475e-05,
409
+ "loss": 2.5596,
410
+ "step": 33500
411
+ },
412
+ {
413
+ "epoch": 0.51,
414
+ "learning_rate": 9.757707070707072e-05,
415
+ "loss": 2.5399,
416
+ "step": 34000
417
+ },
418
+ {
419
+ "epoch": 0.52,
420
+ "learning_rate": 9.752656565656566e-05,
421
+ "loss": 2.5139,
422
+ "step": 34500
423
+ },
424
+ {
425
+ "epoch": 0.53,
426
+ "learning_rate": 9.747606060606062e-05,
427
+ "loss": 2.4959,
428
+ "step": 35000
429
+ },
430
+ {
431
+ "epoch": 0.55,
432
+ "learning_rate": 9.742555555555556e-05,
433
+ "loss": 2.4758,
434
+ "step": 35500
435
+ },
436
+ {
437
+ "epoch": 0.56,
438
+ "learning_rate": 9.737505050505051e-05,
439
+ "loss": 2.4582,
440
+ "step": 36000
441
+ },
442
+ {
443
+ "epoch": 0.57,
444
+ "learning_rate": 9.732454545454545e-05,
445
+ "loss": 2.443,
446
+ "step": 36500
447
+ },
448
+ {
449
+ "epoch": 0.59,
450
+ "learning_rate": 9.727404040404041e-05,
451
+ "loss": 2.4262,
452
+ "step": 37000
453
+ },
454
+ {
455
+ "epoch": 0.6,
456
+ "learning_rate": 9.722353535353536e-05,
457
+ "loss": 2.4139,
458
+ "step": 37500
459
+ },
460
+ {
461
+ "epoch": 0.61,
462
+ "learning_rate": 9.717303030303031e-05,
463
+ "loss": 2.3976,
464
+ "step": 38000
465
+ },
466
+ {
467
+ "epoch": 0.63,
468
+ "learning_rate": 9.712252525252525e-05,
469
+ "loss": 2.3856,
470
+ "step": 38500
471
+ },
472
+ {
473
+ "epoch": 0.64,
474
+ "learning_rate": 9.707202020202021e-05,
475
+ "loss": 2.3699,
476
+ "step": 39000
477
+ },
478
+ {
479
+ "epoch": 0.65,
480
+ "learning_rate": 9.702151515151515e-05,
481
+ "loss": 2.3563,
482
+ "step": 39500
483
+ },
484
+ {
485
+ "epoch": 0.67,
486
+ "learning_rate": 9.69710101010101e-05,
487
+ "loss": 2.3455,
488
+ "step": 40000
489
+ },
490
+ {
491
+ "epoch": 0.68,
492
+ "learning_rate": 9.692050505050505e-05,
493
+ "loss": 2.3341,
494
+ "step": 40500
495
+ },
496
+ {
497
+ "epoch": 0.69,
498
+ "learning_rate": 9.687e-05,
499
+ "loss": 2.3209,
500
+ "step": 41000
501
+ },
502
+ {
503
+ "epoch": 0.71,
504
+ "learning_rate": 9.681949494949495e-05,
505
+ "loss": 2.3087,
506
+ "step": 41500
507
+ },
508
+ {
509
+ "epoch": 0.72,
510
+ "learning_rate": 9.67689898989899e-05,
511
+ "loss": 2.3016,
512
+ "step": 42000
513
+ },
514
+ {
515
+ "epoch": 0.73,
516
+ "learning_rate": 9.671848484848485e-05,
517
+ "loss": 2.2899,
518
+ "step": 42500
519
+ },
520
+ {
521
+ "epoch": 0.75,
522
+ "learning_rate": 9.66679797979798e-05,
523
+ "loss": 2.2808,
524
+ "step": 43000
525
+ },
526
+ {
527
+ "epoch": 0.76,
528
+ "learning_rate": 9.661757575757576e-05,
529
+ "loss": 2.2732,
530
+ "step": 43500
531
+ },
532
+ {
533
+ "epoch": 0.77,
534
+ "learning_rate": 9.656707070707072e-05,
535
+ "loss": 2.2635,
536
+ "step": 44000
537
+ },
538
+ {
539
+ "epoch": 0.79,
540
+ "learning_rate": 9.651656565656566e-05,
541
+ "loss": 2.2496,
542
+ "step": 44500
543
+ },
544
+ {
545
+ "epoch": 0.8,
546
+ "learning_rate": 9.646606060606061e-05,
547
+ "loss": 2.247,
548
+ "step": 45000
549
+ },
550
+ {
551
+ "epoch": 0.81,
552
+ "learning_rate": 9.641555555555556e-05,
553
+ "loss": 2.2398,
554
+ "step": 45500
555
+ },
556
+ {
557
+ "epoch": 0.83,
558
+ "learning_rate": 9.636505050505052e-05,
559
+ "loss": 2.2295,
560
+ "step": 46000
561
+ },
562
+ {
563
+ "epoch": 0.84,
564
+ "learning_rate": 9.631454545454546e-05,
565
+ "loss": 2.2218,
566
+ "step": 46500
567
+ },
568
+ {
569
+ "epoch": 0.85,
570
+ "learning_rate": 9.626404040404041e-05,
571
+ "loss": 2.2126,
572
+ "step": 47000
573
+ },
574
+ {
575
+ "epoch": 0.87,
576
+ "learning_rate": 9.621373737373738e-05,
577
+ "loss": 2.2047,
578
+ "step": 47500
579
+ },
580
+ {
581
+ "epoch": 0.88,
582
+ "learning_rate": 9.616323232323233e-05,
583
+ "loss": 2.1967,
584
+ "step": 48000
585
+ },
586
+ {
587
+ "epoch": 0.89,
588
+ "learning_rate": 9.611313131313133e-05,
589
+ "loss": 2.1927,
590
+ "step": 48500
591
+ },
592
+ {
593
+ "epoch": 0.91,
594
+ "learning_rate": 9.606262626262626e-05,
595
+ "loss": 2.187,
596
+ "step": 49000
597
+ },
598
+ {
599
+ "epoch": 0.92,
600
+ "learning_rate": 9.601212121212122e-05,
601
+ "loss": 2.1775,
602
+ "step": 49500
603
+ },
604
+ {
605
+ "epoch": 0.93,
606
+ "learning_rate": 9.596161616161617e-05,
607
+ "loss": 2.1728,
608
+ "step": 50000
609
+ },
610
+ {
611
+ "epoch": 0.95,
612
+ "learning_rate": 9.591111111111111e-05,
613
+ "loss": 2.1627,
614
+ "step": 50500
615
+ },
616
+ {
617
+ "epoch": 0.96,
618
+ "learning_rate": 9.586060606060606e-05,
619
+ "loss": 2.1593,
620
+ "step": 51000
621
+ },
622
+ {
623
+ "epoch": 0.97,
624
+ "learning_rate": 9.581010101010102e-05,
625
+ "loss": 2.1523,
626
+ "step": 51500
627
+ },
628
+ {
629
+ "epoch": 0.99,
630
+ "learning_rate": 9.575959595959597e-05,
631
+ "loss": 2.1481,
632
+ "step": 52000
633
+ },
634
+ {
635
+ "epoch": 1.0,
636
+ "learning_rate": 9.570909090909091e-05,
637
+ "loss": 2.1405,
638
+ "step": 52500
639
+ },
640
+ {
641
+ "epoch": 1.01,
642
+ "learning_rate": 9.565858585858586e-05,
643
+ "loss": 2.1305,
644
+ "step": 53000
645
+ },
646
+ {
647
+ "epoch": 1.03,
648
+ "learning_rate": 9.560808080808082e-05,
649
+ "loss": 2.1286,
650
+ "step": 53500
651
+ },
652
+ {
653
+ "epoch": 1.04,
654
+ "learning_rate": 9.555757575757575e-05,
655
+ "loss": 2.1233,
656
+ "step": 54000
657
+ },
658
+ {
659
+ "epoch": 1.05,
660
+ "learning_rate": 9.550707070707071e-05,
661
+ "loss": 2.1175,
662
+ "step": 54500
663
+ },
664
+ {
665
+ "epoch": 1.07,
666
+ "learning_rate": 9.545656565656567e-05,
667
+ "loss": 2.1127,
668
+ "step": 55000
669
+ },
670
+ {
671
+ "epoch": 1.08,
672
+ "learning_rate": 9.54060606060606e-05,
673
+ "loss": 2.1042,
674
+ "step": 55500
675
+ },
676
+ {
677
+ "epoch": 1.09,
678
+ "learning_rate": 9.535555555555556e-05,
679
+ "loss": 2.1054,
680
+ "step": 56000
681
+ },
682
+ {
683
+ "epoch": 1.11,
684
+ "learning_rate": 9.530505050505051e-05,
685
+ "loss": 2.099,
686
+ "step": 56500
687
+ },
688
+ {
689
+ "epoch": 1.12,
690
+ "learning_rate": 9.525454545454547e-05,
691
+ "loss": 2.096,
692
+ "step": 57000
693
+ },
694
+ {
695
+ "epoch": 1.13,
696
+ "learning_rate": 9.52040404040404e-05,
697
+ "loss": 2.0863,
698
+ "step": 57500
699
+ },
700
+ {
701
+ "epoch": 1.15,
702
+ "learning_rate": 9.515353535353536e-05,
703
+ "loss": 2.0835,
704
+ "step": 58000
705
+ },
706
+ {
707
+ "epoch": 1.16,
708
+ "learning_rate": 9.510303030303031e-05,
709
+ "loss": 2.0819,
710
+ "step": 58500
711
+ },
712
+ {
713
+ "epoch": 1.17,
714
+ "learning_rate": 9.505252525252526e-05,
715
+ "loss": 2.0762,
716
+ "step": 59000
717
+ },
718
+ {
719
+ "epoch": 1.19,
720
+ "learning_rate": 9.500212121212121e-05,
721
+ "loss": 2.0713,
722
+ "step": 59500
723
+ },
724
+ {
725
+ "epoch": 1.2,
726
+ "learning_rate": 9.495161616161616e-05,
727
+ "loss": 2.068,
728
+ "step": 60000
729
+ },
730
+ {
731
+ "epoch": 1.01,
732
+ "learning_rate": 9.490111111111112e-05,
733
+ "loss": 2.0617,
734
+ "step": 60500
735
+ },
736
+ {
737
+ "epoch": 1.03,
738
+ "learning_rate": 9.485060606060607e-05,
739
+ "loss": 2.062,
740
+ "step": 61000
741
+ },
742
+ {
743
+ "epoch": 1.04,
744
+ "learning_rate": 9.480010101010101e-05,
745
+ "loss": 2.0536,
746
+ "step": 61500
747
+ },
748
+ {
749
+ "epoch": 1.05,
750
+ "learning_rate": 9.474959595959596e-05,
751
+ "loss": 2.0479,
752
+ "step": 62000
753
+ },
754
+ {
755
+ "epoch": 1.07,
756
+ "learning_rate": 9.469909090909092e-05,
757
+ "loss": 2.0468,
758
+ "step": 62500
759
+ },
760
+ {
761
+ "epoch": 1.08,
762
+ "learning_rate": 9.464858585858585e-05,
763
+ "loss": 2.0424,
764
+ "step": 63000
765
+ },
766
+ {
767
+ "epoch": 1.09,
768
+ "learning_rate": 9.459818181818183e-05,
769
+ "loss": 2.0416,
770
+ "step": 63500
771
+ },
772
+ {
773
+ "epoch": 1.11,
774
+ "learning_rate": 9.454777777777779e-05,
775
+ "loss": 2.0356,
776
+ "step": 64000
777
+ },
778
+ {
779
+ "epoch": 1.12,
780
+ "learning_rate": 9.449727272727273e-05,
781
+ "loss": 2.0325,
782
+ "step": 64500
783
+ },
784
+ {
785
+ "epoch": 1.13,
786
+ "learning_rate": 9.444676767676768e-05,
787
+ "loss": 2.0273,
788
+ "step": 65000
789
+ },
790
+ {
791
+ "epoch": 1.15,
792
+ "learning_rate": 9.439626262626264e-05,
793
+ "loss": 2.0243,
794
+ "step": 65500
795
+ },
796
+ {
797
+ "epoch": 1.16,
798
+ "learning_rate": 9.434575757575757e-05,
799
+ "loss": 2.0211,
800
+ "step": 66000
801
+ },
802
+ {
803
+ "epoch": 1.17,
804
+ "learning_rate": 9.429525252525253e-05,
805
+ "loss": 2.0163,
806
+ "step": 66500
807
+ },
808
+ {
809
+ "epoch": 1.19,
810
+ "learning_rate": 9.424484848484849e-05,
811
+ "loss": 2.0159,
812
+ "step": 67000
813
+ },
814
+ {
815
+ "epoch": 1.2,
816
+ "learning_rate": 9.419434343434344e-05,
817
+ "loss": 2.0124,
818
+ "step": 67500
819
+ },
820
+ {
821
+ "epoch": 1.21,
822
+ "learning_rate": 9.41438383838384e-05,
823
+ "loss": 2.0099,
824
+ "step": 68000
825
+ },
826
+ {
827
+ "epoch": 1.23,
828
+ "learning_rate": 9.409333333333333e-05,
829
+ "loss": 2.0052,
830
+ "step": 68500
831
+ },
832
+ {
833
+ "epoch": 1.24,
834
+ "learning_rate": 9.404282828282829e-05,
835
+ "loss": 2.0026,
836
+ "step": 69000
837
+ },
838
+ {
839
+ "epoch": 1.25,
840
+ "learning_rate": 9.399232323232324e-05,
841
+ "loss": 1.9975,
842
+ "step": 69500
843
+ },
844
+ {
845
+ "epoch": 1.27,
846
+ "learning_rate": 9.394181818181818e-05,
847
+ "loss": 1.9983,
848
+ "step": 70000
849
+ },
850
+ {
851
+ "epoch": 1.28,
852
+ "learning_rate": 9.389131313131314e-05,
853
+ "loss": 1.9933,
854
+ "step": 70500
855
+ },
856
+ {
857
+ "epoch": 1.29,
858
+ "learning_rate": 9.384090909090909e-05,
859
+ "loss": 1.9904,
860
+ "step": 71000
861
+ },
862
+ {
863
+ "epoch": 1.31,
864
+ "learning_rate": 9.379040404040405e-05,
865
+ "loss": 1.9854,
866
+ "step": 71500
867
+ },
868
+ {
869
+ "epoch": 1.32,
870
+ "learning_rate": 9.3739898989899e-05,
871
+ "loss": 1.9836,
872
+ "step": 72000
873
+ },
874
+ {
875
+ "epoch": 1.33,
876
+ "learning_rate": 9.368939393939394e-05,
877
+ "loss": 1.9802,
878
+ "step": 72500
879
+ },
880
+ {
881
+ "epoch": 1.35,
882
+ "learning_rate": 9.36389898989899e-05,
883
+ "loss": 1.9806,
884
+ "step": 73000
885
+ },
886
+ {
887
+ "epoch": 1.36,
888
+ "learning_rate": 9.358858585858586e-05,
889
+ "loss": 1.9775,
890
+ "step": 73500
891
+ },
892
+ {
893
+ "epoch": 1.37,
894
+ "learning_rate": 9.35380808080808e-05,
895
+ "loss": 1.9737,
896
+ "step": 74000
897
+ },
898
+ {
899
+ "epoch": 1.39,
900
+ "learning_rate": 9.348757575757577e-05,
901
+ "loss": 1.9708,
902
+ "step": 74500
903
+ },
904
+ {
905
+ "epoch": 1.4,
906
+ "learning_rate": 9.343707070707071e-05,
907
+ "loss": 1.9689,
908
+ "step": 75000
909
+ },
910
+ {
911
+ "epoch": 1.41,
912
+ "learning_rate": 9.338656565656566e-05,
913
+ "loss": 1.9649,
914
+ "step": 75500
915
+ },
916
+ {
917
+ "epoch": 1.43,
918
+ "learning_rate": 9.333606060606062e-05,
919
+ "loss": 1.9629,
920
+ "step": 76000
921
+ },
922
+ {
923
+ "epoch": 1.44,
924
+ "learning_rate": 9.328555555555556e-05,
925
+ "loss": 1.9583,
926
+ "step": 76500
927
+ },
928
+ {
929
+ "epoch": 1.45,
930
+ "learning_rate": 9.323505050505051e-05,
931
+ "loss": 1.957,
932
+ "step": 77000
933
+ },
934
+ {
935
+ "epoch": 1.47,
936
+ "learning_rate": 9.318464646464647e-05,
937
+ "loss": 1.9549,
938
+ "step": 77500
939
+ },
940
+ {
941
+ "epoch": 1.48,
942
+ "learning_rate": 9.313414141414142e-05,
943
+ "loss": 1.9576,
944
+ "step": 78000
945
+ },
946
+ {
947
+ "epoch": 1.49,
948
+ "learning_rate": 9.308363636363638e-05,
949
+ "loss": 1.9471,
950
+ "step": 78500
951
+ },
952
+ {
953
+ "epoch": 1.51,
954
+ "learning_rate": 9.303313131313132e-05,
955
+ "loss": 1.9478,
956
+ "step": 79000
957
+ },
958
+ {
959
+ "epoch": 1.52,
960
+ "learning_rate": 9.298262626262627e-05,
961
+ "loss": 1.9437,
962
+ "step": 79500
963
+ },
964
+ {
965
+ "epoch": 1.53,
966
+ "learning_rate": 9.293212121212122e-05,
967
+ "loss": 1.9433,
968
+ "step": 80000
969
+ },
970
+ {
971
+ "epoch": 1.55,
972
+ "learning_rate": 9.288161616161618e-05,
973
+ "loss": 1.9401,
974
+ "step": 80500
975
+ },
976
+ {
977
+ "epoch": 1.56,
978
+ "learning_rate": 9.283111111111111e-05,
979
+ "loss": 1.9401,
980
+ "step": 81000
981
+ },
982
+ {
983
+ "epoch": 1.57,
984
+ "learning_rate": 9.278070707070707e-05,
985
+ "loss": 1.9385,
986
+ "step": 81500
987
+ },
988
+ {
989
+ "epoch": 1.59,
990
+ "learning_rate": 9.273020202020203e-05,
991
+ "loss": 1.9356,
992
+ "step": 82000
993
+ },
994
+ {
995
+ "epoch": 1.6,
996
+ "learning_rate": 9.267969696969697e-05,
997
+ "loss": 1.9305,
998
+ "step": 82500
999
+ },
1000
+ {
1001
+ "epoch": 1.61,
1002
+ "learning_rate": 9.262919191919192e-05,
1003
+ "loss": 1.9306,
1004
+ "step": 83000
1005
+ },
1006
+ {
1007
+ "epoch": 1.63,
1008
+ "learning_rate": 9.257868686868687e-05,
1009
+ "loss": 1.9265,
1010
+ "step": 83500
1011
+ },
1012
+ {
1013
+ "epoch": 1.64,
1014
+ "learning_rate": 9.252818181818183e-05,
1015
+ "loss": 1.9266,
1016
+ "step": 84000
1017
+ },
1018
+ {
1019
+ "epoch": 1.65,
1020
+ "learning_rate": 9.247777777777779e-05,
1021
+ "loss": 1.922,
1022
+ "step": 84500
1023
+ },
1024
+ {
1025
+ "epoch": 1.67,
1026
+ "learning_rate": 9.242727272727273e-05,
1027
+ "loss": 1.924,
1028
+ "step": 85000
1029
+ },
1030
+ {
1031
+ "epoch": 1.68,
1032
+ "learning_rate": 9.237676767676768e-05,
1033
+ "loss": 1.9211,
1034
+ "step": 85500
1035
+ },
1036
+ {
1037
+ "epoch": 1.69,
1038
+ "learning_rate": 9.232626262626263e-05,
1039
+ "loss": 1.9181,
1040
+ "step": 86000
1041
+ },
1042
+ {
1043
+ "epoch": 1.71,
1044
+ "learning_rate": 9.227585858585859e-05,
1045
+ "loss": 1.9191,
1046
+ "step": 86500
1047
+ },
1048
+ {
1049
+ "epoch": 1.72,
1050
+ "learning_rate": 9.222535353535354e-05,
1051
+ "loss": 1.9154,
1052
+ "step": 87000
1053
+ },
1054
+ {
1055
+ "epoch": 1.73,
1056
+ "learning_rate": 9.217484848484849e-05,
1057
+ "loss": 1.9104,
1058
+ "step": 87500
1059
+ },
1060
+ {
1061
+ "epoch": 1.75,
1062
+ "learning_rate": 9.212434343434344e-05,
1063
+ "loss": 1.9092,
1064
+ "step": 88000
1065
+ },
1066
+ {
1067
+ "epoch": 1.76,
1068
+ "learning_rate": 9.20739393939394e-05,
1069
+ "loss": 1.9071,
1070
+ "step": 88500
1071
+ },
1072
+ {
1073
+ "epoch": 1.77,
1074
+ "learning_rate": 9.202343434343434e-05,
1075
+ "loss": 1.9066,
1076
+ "step": 89000
1077
+ },
1078
+ {
1079
+ "epoch": 1.79,
1080
+ "learning_rate": 9.19729292929293e-05,
1081
+ "loss": 1.9048,
1082
+ "step": 89500
1083
+ },
1084
+ {
1085
+ "epoch": 1.8,
1086
+ "learning_rate": 9.192242424242425e-05,
1087
+ "loss": 1.9006,
1088
+ "step": 90000
1089
+ },
1090
+ {
1091
+ "epoch": 1.81,
1092
+ "learning_rate": 9.18719191919192e-05,
1093
+ "loss": 1.9012,
1094
+ "step": 90500
1095
+ },
1096
+ {
1097
+ "epoch": 1.83,
1098
+ "learning_rate": 9.182151515151516e-05,
1099
+ "loss": 1.8953,
1100
+ "step": 91000
1101
+ },
1102
+ {
1103
+ "epoch": 1.84,
1104
+ "learning_rate": 9.17710101010101e-05,
1105
+ "loss": 1.8962,
1106
+ "step": 91500
1107
+ },
1108
+ {
1109
+ "epoch": 1.85,
1110
+ "learning_rate": 9.172050505050506e-05,
1111
+ "loss": 1.8991,
1112
+ "step": 92000
1113
+ },
1114
+ {
1115
+ "epoch": 1.87,
1116
+ "learning_rate": 9.167e-05,
1117
+ "loss": 1.8911,
1118
+ "step": 92500
1119
+ },
1120
+ {
1121
+ "epoch": 1.88,
1122
+ "learning_rate": 9.161949494949495e-05,
1123
+ "loss": 1.8945,
1124
+ "step": 93000
1125
+ },
1126
+ {
1127
+ "epoch": 1.89,
1128
+ "learning_rate": 9.15689898989899e-05,
1129
+ "loss": 1.8904,
1130
+ "step": 93500
1131
+ },
1132
+ {
1133
+ "epoch": 1.91,
1134
+ "learning_rate": 9.151848484848485e-05,
1135
+ "loss": 1.8859,
1136
+ "step": 94000
1137
+ },
1138
+ {
1139
+ "epoch": 1.92,
1140
+ "learning_rate": 9.14679797979798e-05,
1141
+ "loss": 1.8836,
1142
+ "step": 94500
1143
+ },
1144
+ {
1145
+ "epoch": 1.93,
1146
+ "learning_rate": 9.141747474747475e-05,
1147
+ "loss": 1.88,
1148
+ "step": 95000
1149
+ },
1150
+ {
1151
+ "epoch": 1.95,
1152
+ "learning_rate": 9.13669696969697e-05,
1153
+ "loss": 1.8834,
1154
+ "step": 95500
1155
+ },
1156
+ {
1157
+ "epoch": 1.96,
1158
+ "learning_rate": 9.131656565656566e-05,
1159
+ "loss": 1.8813,
1160
+ "step": 96000
1161
+ },
1162
+ {
1163
+ "epoch": 1.97,
1164
+ "learning_rate": 9.126616161616162e-05,
1165
+ "loss": 1.881,
1166
+ "step": 96500
1167
+ },
1168
+ {
1169
+ "epoch": 1.99,
1170
+ "learning_rate": 9.121575757575758e-05,
1171
+ "loss": 1.8785,
1172
+ "step": 97000
1173
+ },
1174
+ {
1175
+ "epoch": 2.0,
1176
+ "learning_rate": 9.116525252525252e-05,
1177
+ "loss": 1.8735,
1178
+ "step": 97500
1179
+ },
1180
+ {
1181
+ "epoch": 2.01,
1182
+ "learning_rate": 9.111474747474747e-05,
1183
+ "loss": 1.8707,
1184
+ "step": 98000
1185
+ },
1186
+ {
1187
+ "epoch": 2.03,
1188
+ "learning_rate": 9.106424242424243e-05,
1189
+ "loss": 1.8702,
1190
+ "step": 98500
1191
+ },
1192
+ {
1193
+ "epoch": 2.04,
1194
+ "learning_rate": 9.101373737373738e-05,
1195
+ "loss": 1.8696,
1196
+ "step": 99000
1197
+ },
1198
+ {
1199
+ "epoch": 2.05,
1200
+ "learning_rate": 9.096323232323232e-05,
1201
+ "loss": 1.8667,
1202
+ "step": 99500
1203
+ },
1204
+ {
1205
+ "epoch": 2.07,
1206
+ "learning_rate": 9.091272727272727e-05,
1207
+ "loss": 1.8645,
1208
+ "step": 100000
1209
+ },
1210
+ {
1211
+ "epoch": 2.01,
1212
+ "learning_rate": 9.086222222222223e-05,
1213
+ "loss": 1.8615,
1214
+ "step": 100500
1215
+ },
1216
+ {
1217
+ "epoch": 2.03,
1218
+ "learning_rate": 9.081171717171718e-05,
1219
+ "loss": 1.8605,
1220
+ "step": 101000
1221
+ },
1222
+ {
1223
+ "epoch": 2.04,
1224
+ "learning_rate": 9.076131313131314e-05,
1225
+ "loss": 1.8579,
1226
+ "step": 101500
1227
+ },
1228
+ {
1229
+ "epoch": 2.05,
1230
+ "learning_rate": 9.071080808080808e-05,
1231
+ "loss": 1.8513,
1232
+ "step": 102000
1233
+ },
1234
+ {
1235
+ "epoch": 2.07,
1236
+ "learning_rate": 9.066030303030304e-05,
1237
+ "loss": 1.854,
1238
+ "step": 102500
1239
+ },
1240
+ {
1241
+ "epoch": 2.08,
1242
+ "learning_rate": 9.060979797979799e-05,
1243
+ "loss": 1.8577,
1244
+ "step": 103000
1245
+ },
1246
+ {
1247
+ "epoch": 2.09,
1248
+ "learning_rate": 9.055929292929294e-05,
1249
+ "loss": 1.8557,
1250
+ "step": 103500
1251
+ },
1252
+ {
1253
+ "epoch": 2.11,
1254
+ "learning_rate": 9.05088888888889e-05,
1255
+ "loss": 1.854,
1256
+ "step": 104000
1257
+ },
1258
+ {
1259
+ "epoch": 2.12,
1260
+ "learning_rate": 9.045838383838384e-05,
1261
+ "loss": 1.8532,
1262
+ "step": 104500
1263
+ },
1264
+ {
1265
+ "epoch": 2.13,
1266
+ "learning_rate": 9.04078787878788e-05,
1267
+ "loss": 1.8507,
1268
+ "step": 105000
1269
+ },
1270
+ {
1271
+ "epoch": 2.15,
1272
+ "learning_rate": 9.035737373737375e-05,
1273
+ "loss": 1.8485,
1274
+ "step": 105500
1275
+ },
1276
+ {
1277
+ "epoch": 2.16,
1278
+ "learning_rate": 9.03069696969697e-05,
1279
+ "loss": 1.8498,
1280
+ "step": 106000
1281
+ },
1282
+ {
1283
+ "epoch": 2.17,
1284
+ "learning_rate": 9.025646464646465e-05,
1285
+ "loss": 1.8447,
1286
+ "step": 106500
1287
+ },
1288
+ {
1289
+ "epoch": 2.19,
1290
+ "learning_rate": 9.02059595959596e-05,
1291
+ "loss": 1.8434,
1292
+ "step": 107000
1293
+ },
1294
+ {
1295
+ "epoch": 2.2,
1296
+ "learning_rate": 9.015545454545456e-05,
1297
+ "loss": 1.8453,
1298
+ "step": 107500
1299
+ },
1300
+ {
1301
+ "epoch": 2.21,
1302
+ "learning_rate": 9.010505050505052e-05,
1303
+ "loss": 1.8461,
1304
+ "step": 108000
1305
+ },
1306
+ {
1307
+ "epoch": 2.23,
1308
+ "learning_rate": 9.005454545454545e-05,
1309
+ "loss": 1.8435,
1310
+ "step": 108500
1311
+ },
1312
+ {
1313
+ "epoch": 2.24,
1314
+ "learning_rate": 9.000404040404041e-05,
1315
+ "loss": 1.8408,
1316
+ "step": 109000
1317
+ },
1318
+ {
1319
+ "epoch": 2.25,
1320
+ "learning_rate": 8.995353535353536e-05,
1321
+ "loss": 1.8387,
1322
+ "step": 109500
1323
+ },
1324
+ {
1325
+ "epoch": 2.27,
1326
+ "learning_rate": 8.990303030303032e-05,
1327
+ "loss": 1.8399,
1328
+ "step": 110000
1329
+ },
1330
+ {
1331
+ "epoch": 2.28,
1332
+ "learning_rate": 8.985252525252525e-05,
1333
+ "loss": 1.838,
1334
+ "step": 110500
1335
+ },
1336
+ {
1337
+ "epoch": 2.29,
1338
+ "learning_rate": 8.980212121212121e-05,
1339
+ "loss": 1.8371,
1340
+ "step": 111000
1341
+ },
1342
+ {
1343
+ "epoch": 2.31,
1344
+ "learning_rate": 8.975161616161617e-05,
1345
+ "loss": 1.8337,
1346
+ "step": 111500
1347
+ },
1348
+ {
1349
+ "epoch": 2.32,
1350
+ "learning_rate": 8.970111111111112e-05,
1351
+ "loss": 1.8354,
1352
+ "step": 112000
1353
+ },
1354
+ {
1355
+ "epoch": 2.33,
1356
+ "learning_rate": 8.965060606060606e-05,
1357
+ "loss": 1.8291,
1358
+ "step": 112500
1359
+ },
1360
+ {
1361
+ "epoch": 2.35,
1362
+ "learning_rate": 8.960010101010101e-05,
1363
+ "loss": 1.8313,
1364
+ "step": 113000
1365
+ },
1366
+ {
1367
+ "epoch": 2.36,
1368
+ "learning_rate": 8.954959595959597e-05,
1369
+ "loss": 1.8295,
1370
+ "step": 113500
1371
+ },
1372
+ {
1373
+ "epoch": 2.37,
1374
+ "learning_rate": 8.949909090909092e-05,
1375
+ "loss": 1.8289,
1376
+ "step": 114000
1377
+ },
1378
+ {
1379
+ "epoch": 2.39,
1380
+ "learning_rate": 8.944858585858586e-05,
1381
+ "loss": 1.8248,
1382
+ "step": 114500
1383
+ },
1384
+ {
1385
+ "epoch": 2.4,
1386
+ "learning_rate": 8.939818181818182e-05,
1387
+ "loss": 1.8248,
1388
+ "step": 115000
1389
+ },
1390
+ {
1391
+ "epoch": 2.41,
1392
+ "learning_rate": 8.934767676767677e-05,
1393
+ "loss": 1.8244,
1394
+ "step": 115500
1395
+ },
1396
+ {
1397
+ "epoch": 2.43,
1398
+ "learning_rate": 8.929717171717173e-05,
1399
+ "loss": 1.8228,
1400
+ "step": 116000
1401
+ },
1402
+ {
1403
+ "epoch": 2.44,
1404
+ "learning_rate": 8.924666666666667e-05,
1405
+ "loss": 1.8262,
1406
+ "step": 116500
1407
+ },
1408
+ {
1409
+ "epoch": 2.45,
1410
+ "learning_rate": 8.919616161616162e-05,
1411
+ "loss": 1.8237,
1412
+ "step": 117000
1413
+ },
1414
+ {
1415
+ "epoch": 2.47,
1416
+ "learning_rate": 8.914565656565657e-05,
1417
+ "loss": 1.8182,
1418
+ "step": 117500
1419
+ },
1420
+ {
1421
+ "epoch": 2.48,
1422
+ "learning_rate": 8.909515151515153e-05,
1423
+ "loss": 1.8198,
1424
+ "step": 118000
1425
+ },
1426
+ {
1427
+ "epoch": 2.49,
1428
+ "learning_rate": 8.904464646464646e-05,
1429
+ "loss": 1.8176,
1430
+ "step": 118500
1431
+ },
1432
+ {
1433
+ "epoch": 2.51,
1434
+ "learning_rate": 8.899424242424242e-05,
1435
+ "loss": 1.8169,
1436
+ "step": 119000
1437
+ },
1438
+ {
1439
+ "epoch": 2.52,
1440
+ "learning_rate": 8.894373737373738e-05,
1441
+ "loss": 1.8131,
1442
+ "step": 119500
1443
+ },
1444
+ {
1445
+ "epoch": 2.53,
1446
+ "learning_rate": 8.889323232323233e-05,
1447
+ "loss": 1.812,
1448
+ "step": 120000
1449
+ },
1450
+ {
1451
+ "epoch": 2.55,
1452
+ "learning_rate": 8.884272727272727e-05,
1453
+ "loss": 1.8113,
1454
+ "step": 120500
1455
+ },
1456
+ {
1457
+ "epoch": 2.56,
1458
+ "learning_rate": 8.879222222222222e-05,
1459
+ "loss": 1.8122,
1460
+ "step": 121000
1461
+ },
1462
+ {
1463
+ "epoch": 2.57,
1464
+ "learning_rate": 8.87420202020202e-05,
1465
+ "loss": 1.8105,
1466
+ "step": 121500
1467
+ },
1468
+ {
1469
+ "epoch": 2.59,
1470
+ "learning_rate": 8.869151515151516e-05,
1471
+ "loss": 1.8107,
1472
+ "step": 122000
1473
+ },
1474
+ {
1475
+ "epoch": 2.6,
1476
+ "learning_rate": 8.86410101010101e-05,
1477
+ "loss": 1.8084,
1478
+ "step": 122500
1479
+ },
1480
+ {
1481
+ "epoch": 2.61,
1482
+ "learning_rate": 8.859050505050506e-05,
1483
+ "loss": 1.8092,
1484
+ "step": 123000
1485
+ },
1486
+ {
1487
+ "epoch": 2.63,
1488
+ "learning_rate": 8.854e-05,
1489
+ "loss": 1.8055,
1490
+ "step": 123500
1491
+ },
1492
+ {
1493
+ "epoch": 2.64,
1494
+ "learning_rate": 8.848949494949495e-05,
1495
+ "loss": 1.8041,
1496
+ "step": 124000
1497
+ },
1498
+ {
1499
+ "epoch": 2.65,
1500
+ "learning_rate": 8.84389898989899e-05,
1501
+ "loss": 1.8036,
1502
+ "step": 124500
1503
+ },
1504
+ {
1505
+ "epoch": 2.67,
1506
+ "learning_rate": 8.838858585858585e-05,
1507
+ "loss": 1.8059,
1508
+ "step": 125000
1509
+ },
1510
+ {
1511
+ "epoch": 2.68,
1512
+ "learning_rate": 8.833808080808081e-05,
1513
+ "loss": 1.8004,
1514
+ "step": 125500
1515
+ },
1516
+ {
1517
+ "epoch": 2.69,
1518
+ "learning_rate": 8.828757575757576e-05,
1519
+ "loss": 1.8008,
1520
+ "step": 126000
1521
+ },
1522
+ {
1523
+ "epoch": 2.71,
1524
+ "learning_rate": 8.823707070707071e-05,
1525
+ "loss": 1.8048,
1526
+ "step": 126500
1527
+ },
1528
+ {
1529
+ "epoch": 2.72,
1530
+ "learning_rate": 8.818656565656565e-05,
1531
+ "loss": 1.7992,
1532
+ "step": 127000
1533
+ },
1534
+ {
1535
+ "epoch": 2.73,
1536
+ "learning_rate": 8.813606060606061e-05,
1537
+ "loss": 1.797,
1538
+ "step": 127500
1539
+ },
1540
+ {
1541
+ "epoch": 2.75,
1542
+ "learning_rate": 8.808555555555556e-05,
1543
+ "loss": 1.7949,
1544
+ "step": 128000
1545
+ },
1546
+ {
1547
+ "epoch": 2.76,
1548
+ "learning_rate": 8.80350505050505e-05,
1549
+ "loss": 1.7984,
1550
+ "step": 128500
1551
+ },
1552
+ {
1553
+ "epoch": 2.77,
1554
+ "learning_rate": 8.798454545454547e-05,
1555
+ "loss": 1.7954,
1556
+ "step": 129000
1557
+ },
1558
+ {
1559
+ "epoch": 2.79,
1560
+ "learning_rate": 8.793414141414141e-05,
1561
+ "loss": 1.7979,
1562
+ "step": 129500
1563
+ },
1564
+ {
1565
+ "epoch": 2.8,
1566
+ "learning_rate": 8.788363636363637e-05,
1567
+ "loss": 1.795,
1568
+ "step": 130000
1569
+ },
1570
+ {
1571
+ "epoch": 2.81,
1572
+ "learning_rate": 8.783313131313132e-05,
1573
+ "loss": 1.7944,
1574
+ "step": 130500
1575
+ },
1576
+ {
1577
+ "epoch": 2.83,
1578
+ "learning_rate": 8.778262626262627e-05,
1579
+ "loss": 1.7907,
1580
+ "step": 131000
1581
+ },
1582
+ {
1583
+ "epoch": 2.84,
1584
+ "learning_rate": 8.773212121212122e-05,
1585
+ "loss": 1.7912,
1586
+ "step": 131500
1587
+ },
1588
+ {
1589
+ "epoch": 2.85,
1590
+ "learning_rate": 8.768171717171717e-05,
1591
+ "loss": 1.7935,
1592
+ "step": 132000
1593
+ },
1594
+ {
1595
+ "epoch": 2.87,
1596
+ "learning_rate": 8.763121212121213e-05,
1597
+ "loss": 1.79,
1598
+ "step": 132500
1599
+ },
1600
+ {
1601
+ "epoch": 2.88,
1602
+ "learning_rate": 8.758070707070708e-05,
1603
+ "loss": 1.7892,
1604
+ "step": 133000
1605
+ },
1606
+ {
1607
+ "epoch": 2.89,
1608
+ "learning_rate": 8.753020202020202e-05,
1609
+ "loss": 1.787,
1610
+ "step": 133500
1611
+ },
1612
+ {
1613
+ "epoch": 2.91,
1614
+ "learning_rate": 8.747969696969698e-05,
1615
+ "loss": 1.7887,
1616
+ "step": 134000
1617
+ },
1618
+ {
1619
+ "epoch": 2.92,
1620
+ "learning_rate": 8.742919191919192e-05,
1621
+ "loss": 1.7821,
1622
+ "step": 134500
1623
+ },
1624
+ {
1625
+ "epoch": 2.93,
1626
+ "learning_rate": 8.737878787878788e-05,
1627
+ "loss": 1.7862,
1628
+ "step": 135000
1629
+ },
1630
+ {
1631
+ "epoch": 2.95,
1632
+ "learning_rate": 8.732838383838385e-05,
1633
+ "loss": 1.7833,
1634
+ "step": 135500
1635
+ },
1636
+ {
1637
+ "epoch": 2.96,
1638
+ "learning_rate": 8.72778787878788e-05,
1639
+ "loss": 1.7795,
1640
+ "step": 136000
1641
+ },
1642
+ {
1643
+ "epoch": 2.97,
1644
+ "learning_rate": 8.722737373737374e-05,
1645
+ "loss": 1.7825,
1646
+ "step": 136500
1647
+ },
1648
+ {
1649
+ "epoch": 2.99,
1650
+ "learning_rate": 8.71768686868687e-05,
1651
+ "loss": 1.7804,
1652
+ "step": 137000
1653
+ },
1654
+ {
1655
+ "epoch": 3.0,
1656
+ "learning_rate": 8.712636363636363e-05,
1657
+ "loss": 1.7818,
1658
+ "step": 137500
1659
+ },
1660
+ {
1661
+ "epoch": 3.01,
1662
+ "learning_rate": 8.70758585858586e-05,
1663
+ "loss": 1.7751,
1664
+ "step": 138000
1665
+ },
1666
+ {
1667
+ "epoch": 3.03,
1668
+ "learning_rate": 8.702535353535354e-05,
1669
+ "loss": 1.7759,
1670
+ "step": 138500
1671
+ },
1672
+ {
1673
+ "epoch": 3.04,
1674
+ "learning_rate": 8.697484848484849e-05,
1675
+ "loss": 1.7723,
1676
+ "step": 139000
1677
+ },
1678
+ {
1679
+ "epoch": 3.05,
1680
+ "learning_rate": 8.692434343434343e-05,
1681
+ "loss": 1.7776,
1682
+ "step": 139500
1683
+ },
1684
+ {
1685
+ "epoch": 3.07,
1686
+ "learning_rate": 8.687393939393939e-05,
1687
+ "loss": 1.772,
1688
+ "step": 140000
1689
+ },
1690
+ {
1691
+ "epoch": 3.08,
1692
+ "learning_rate": 8.682343434343435e-05,
1693
+ "loss": 1.7723,
1694
+ "step": 140500
1695
+ },
1696
+ {
1697
+ "epoch": 3.09,
1698
+ "learning_rate": 8.67729292929293e-05,
1699
+ "loss": 1.773,
1700
+ "step": 141000
1701
+ },
1702
+ {
1703
+ "epoch": 3.11,
1704
+ "learning_rate": 8.672242424242425e-05,
1705
+ "loss": 1.7728,
1706
+ "step": 141500
1707
+ },
1708
+ {
1709
+ "epoch": 3.12,
1710
+ "learning_rate": 8.667191919191919e-05,
1711
+ "loss": 1.7723,
1712
+ "step": 142000
1713
+ },
1714
+ {
1715
+ "epoch": 3.13,
1716
+ "learning_rate": 8.662151515151515e-05,
1717
+ "loss": 1.7704,
1718
+ "step": 142500
1719
+ },
1720
+ {
1721
+ "epoch": 3.15,
1722
+ "learning_rate": 8.657101010101011e-05,
1723
+ "loss": 1.7684,
1724
+ "step": 143000
1725
+ },
1726
+ {
1727
+ "epoch": 3.16,
1728
+ "learning_rate": 8.652050505050506e-05,
1729
+ "loss": 1.7666,
1730
+ "step": 143500
1731
+ },
1732
+ {
1733
+ "epoch": 3.17,
1734
+ "learning_rate": 8.647e-05,
1735
+ "loss": 1.7679,
1736
+ "step": 144000
1737
+ },
1738
+ {
1739
+ "epoch": 3.19,
1740
+ "learning_rate": 8.641949494949495e-05,
1741
+ "loss": 1.7665,
1742
+ "step": 144500
1743
+ },
1744
+ {
1745
+ "epoch": 3.2,
1746
+ "learning_rate": 8.636909090909091e-05,
1747
+ "loss": 1.7651,
1748
+ "step": 145000
1749
+ },
1750
+ {
1751
+ "epoch": 3.21,
1752
+ "learning_rate": 8.631858585858587e-05,
1753
+ "loss": 1.7629,
1754
+ "step": 145500
1755
+ },
1756
+ {
1757
+ "epoch": 3.23,
1758
+ "learning_rate": 8.62680808080808e-05,
1759
+ "loss": 1.7663,
1760
+ "step": 146000
1761
+ },
1762
+ {
1763
+ "epoch": 3.24,
1764
+ "learning_rate": 8.621757575757576e-05,
1765
+ "loss": 1.762,
1766
+ "step": 146500
1767
+ },
1768
+ {
1769
+ "epoch": 3.25,
1770
+ "learning_rate": 8.616717171717172e-05,
1771
+ "loss": 1.7659,
1772
+ "step": 147000
1773
+ },
1774
+ {
1775
+ "epoch": 3.27,
1776
+ "learning_rate": 8.611666666666667e-05,
1777
+ "loss": 1.763,
1778
+ "step": 147500
1779
+ },
1780
+ {
1781
+ "epoch": 3.28,
1782
+ "learning_rate": 8.606626262626263e-05,
1783
+ "loss": 1.7618,
1784
+ "step": 148000
1785
+ },
1786
+ {
1787
+ "epoch": 3.29,
1788
+ "learning_rate": 8.601575757575759e-05,
1789
+ "loss": 1.7611,
1790
+ "step": 148500
1791
+ },
1792
+ {
1793
+ "epoch": 3.31,
1794
+ "learning_rate": 8.596525252525252e-05,
1795
+ "loss": 1.7598,
1796
+ "step": 149000
1797
+ },
1798
+ {
1799
+ "epoch": 3.32,
1800
+ "learning_rate": 8.591474747474748e-05,
1801
+ "loss": 1.7596,
1802
+ "step": 149500
1803
+ },
1804
+ {
1805
+ "epoch": 3.33,
1806
+ "learning_rate": 8.586424242424243e-05,
1807
+ "loss": 1.7591,
1808
+ "step": 150000
1809
+ },
1810
+ {
1811
+ "epoch": 3.35,
1812
+ "learning_rate": 8.581373737373737e-05,
1813
+ "loss": 1.7579,
1814
+ "step": 150500
1815
+ },
1816
+ {
1817
+ "epoch": 3.36,
1818
+ "learning_rate": 8.576323232323232e-05,
1819
+ "loss": 1.7576,
1820
+ "step": 151000
1821
+ },
1822
+ {
1823
+ "epoch": 3.37,
1824
+ "learning_rate": 8.571272727272728e-05,
1825
+ "loss": 1.7565,
1826
+ "step": 151500
1827
+ },
1828
+ {
1829
+ "epoch": 3.39,
1830
+ "learning_rate": 8.566222222222223e-05,
1831
+ "loss": 1.7553,
1832
+ "step": 152000
1833
+ },
1834
+ {
1835
+ "epoch": 3.4,
1836
+ "learning_rate": 8.561171717171717e-05,
1837
+ "loss": 1.7523,
1838
+ "step": 152500
1839
+ },
1840
+ {
1841
+ "epoch": 3.41,
1842
+ "learning_rate": 8.556131313131313e-05,
1843
+ "loss": 1.7522,
1844
+ "step": 153000
1845
+ },
1846
+ {
1847
+ "epoch": 3.43,
1848
+ "learning_rate": 8.551080808080808e-05,
1849
+ "loss": 1.7551,
1850
+ "step": 153500
1851
+ },
1852
+ {
1853
+ "epoch": 3.44,
1854
+ "learning_rate": 8.546030303030304e-05,
1855
+ "loss": 1.7525,
1856
+ "step": 154000
1857
+ },
1858
+ {
1859
+ "epoch": 3.45,
1860
+ "learning_rate": 8.540979797979798e-05,
1861
+ "loss": 1.7523,
1862
+ "step": 154500
1863
+ },
1864
+ {
1865
+ "epoch": 3.47,
1866
+ "learning_rate": 8.535929292929293e-05,
1867
+ "loss": 1.7543,
1868
+ "step": 155000
1869
+ },
1870
+ {
1871
+ "epoch": 3.48,
1872
+ "learning_rate": 8.530888888888889e-05,
1873
+ "loss": 1.7491,
1874
+ "step": 155500
1875
+ },
1876
+ {
1877
+ "epoch": 3.49,
1878
+ "learning_rate": 8.525848484848485e-05,
1879
+ "loss": 1.7486,
1880
+ "step": 156000
1881
+ },
1882
+ {
1883
+ "epoch": 3.51,
1884
+ "learning_rate": 8.52079797979798e-05,
1885
+ "loss": 1.7521,
1886
+ "step": 156500
1887
+ },
1888
+ {
1889
+ "epoch": 3.52,
1890
+ "learning_rate": 8.515747474747476e-05,
1891
+ "loss": 1.7481,
1892
+ "step": 157000
1893
+ },
1894
+ {
1895
+ "epoch": 3.53,
1896
+ "learning_rate": 8.51069696969697e-05,
1897
+ "loss": 1.7482,
1898
+ "step": 157500
1899
+ },
1900
+ {
1901
+ "epoch": 3.55,
1902
+ "learning_rate": 8.505646464646465e-05,
1903
+ "loss": 1.7481,
1904
+ "step": 158000
1905
+ },
1906
+ {
1907
+ "epoch": 3.56,
1908
+ "learning_rate": 8.500595959595961e-05,
1909
+ "loss": 1.7466,
1910
+ "step": 158500
1911
+ },
1912
+ {
1913
+ "epoch": 3.57,
1914
+ "learning_rate": 8.495545454545455e-05,
1915
+ "loss": 1.7479,
1916
+ "step": 159000
1917
+ },
1918
+ {
1919
+ "epoch": 3.59,
1920
+ "learning_rate": 8.49049494949495e-05,
1921
+ "loss": 1.7431,
1922
+ "step": 159500
1923
+ },
1924
+ {
1925
+ "epoch": 3.6,
1926
+ "learning_rate": 8.485444444444445e-05,
1927
+ "loss": 1.7464,
1928
+ "step": 160000
1929
+ },
1930
+ {
1931
+ "epoch": 3.61,
1932
+ "learning_rate": 8.480404040404041e-05,
1933
+ "loss": 1.7433,
1934
+ "step": 160500
1935
+ },
1936
+ {
1937
+ "epoch": 3.63,
1938
+ "learning_rate": 8.475363636363637e-05,
1939
+ "loss": 1.7421,
1940
+ "step": 161000
1941
+ },
1942
+ {
1943
+ "epoch": 3.64,
1944
+ "learning_rate": 8.470313131313131e-05,
1945
+ "loss": 1.7432,
1946
+ "step": 161500
1947
+ },
1948
+ {
1949
+ "epoch": 3.65,
1950
+ "learning_rate": 8.465262626262627e-05,
1951
+ "loss": 1.7413,
1952
+ "step": 162000
1953
+ },
1954
+ {
1955
+ "epoch": 3.67,
1956
+ "learning_rate": 8.460212121212122e-05,
1957
+ "loss": 1.7415,
1958
+ "step": 162500
1959
+ },
1960
+ {
1961
+ "epoch": 3.68,
1962
+ "learning_rate": 8.455161616161617e-05,
1963
+ "loss": 1.7388,
1964
+ "step": 163000
1965
+ },
1966
+ {
1967
+ "epoch": 3.69,
1968
+ "learning_rate": 8.450111111111113e-05,
1969
+ "loss": 1.74,
1970
+ "step": 163500
1971
+ },
1972
+ {
1973
+ "epoch": 3.71,
1974
+ "learning_rate": 8.445070707070709e-05,
1975
+ "loss": 1.7387,
1976
+ "step": 164000
1977
+ },
1978
+ {
1979
+ "epoch": 3.72,
1980
+ "learning_rate": 8.440020202020202e-05,
1981
+ "loss": 1.7392,
1982
+ "step": 164500
1983
+ },
1984
+ {
1985
+ "epoch": 3.73,
1986
+ "learning_rate": 8.434969696969698e-05,
1987
+ "loss": 1.7391,
1988
+ "step": 165000
1989
+ },
1990
+ {
1991
+ "epoch": 3.75,
1992
+ "learning_rate": 8.429919191919192e-05,
1993
+ "loss": 1.7389,
1994
+ "step": 165500
1995
+ },
1996
+ {
1997
+ "epoch": 3.76,
1998
+ "learning_rate": 8.424868686868687e-05,
1999
+ "loss": 1.7355,
2000
+ "step": 166000
2001
+ },
2002
+ {
2003
+ "epoch": 3.77,
2004
+ "learning_rate": 8.419818181818182e-05,
2005
+ "loss": 1.736,
2006
+ "step": 166500
2007
+ },
2008
+ {
2009
+ "epoch": 3.79,
2010
+ "learning_rate": 8.414767676767678e-05,
2011
+ "loss": 1.734,
2012
+ "step": 167000
2013
+ },
2014
+ {
2015
+ "epoch": 3.8,
2016
+ "learning_rate": 8.409717171717172e-05,
2017
+ "loss": 1.7356,
2018
+ "step": 167500
2019
+ },
2020
+ {
2021
+ "epoch": 3.81,
2022
+ "learning_rate": 8.404676767676768e-05,
2023
+ "loss": 1.7363,
2024
+ "step": 168000
2025
+ },
2026
+ {
2027
+ "epoch": 3.83,
2028
+ "learning_rate": 8.399626262626263e-05,
2029
+ "loss": 1.7342,
2030
+ "step": 168500
2031
+ },
2032
+ {
2033
+ "epoch": 3.84,
2034
+ "learning_rate": 8.394575757575758e-05,
2035
+ "loss": 1.7305,
2036
+ "step": 169000
2037
+ },
2038
+ {
2039
+ "epoch": 3.85,
2040
+ "learning_rate": 8.389525252525254e-05,
2041
+ "loss": 1.7314,
2042
+ "step": 169500
2043
+ },
2044
+ {
2045
+ "epoch": 3.87,
2046
+ "learning_rate": 8.384474747474748e-05,
2047
+ "loss": 1.729,
2048
+ "step": 170000
2049
+ },
2050
+ {
2051
+ "epoch": 3.88,
2052
+ "learning_rate": 8.379424242424243e-05,
2053
+ "loss": 1.7323,
2054
+ "step": 170500
2055
+ },
2056
+ {
2057
+ "epoch": 3.89,
2058
+ "learning_rate": 8.374383838383839e-05,
2059
+ "loss": 1.7309,
2060
+ "step": 171000
2061
+ },
2062
+ {
2063
+ "epoch": 3.91,
2064
+ "learning_rate": 8.369333333333333e-05,
2065
+ "loss": 1.7317,
2066
+ "step": 171500
2067
+ },
2068
+ {
2069
+ "epoch": 3.92,
2070
+ "learning_rate": 8.36428282828283e-05,
2071
+ "loss": 1.729,
2072
+ "step": 172000
2073
+ },
2074
+ {
2075
+ "epoch": 3.93,
2076
+ "learning_rate": 8.359232323232323e-05,
2077
+ "loss": 1.7302,
2078
+ "step": 172500
2079
+ },
2080
+ {
2081
+ "epoch": 3.95,
2082
+ "learning_rate": 8.354181818181819e-05,
2083
+ "loss": 1.7301,
2084
+ "step": 173000
2085
+ },
2086
+ {
2087
+ "epoch": 3.96,
2088
+ "learning_rate": 8.349131313131313e-05,
2089
+ "loss": 1.7267,
2090
+ "step": 173500
2091
+ },
2092
+ {
2093
+ "epoch": 3.97,
2094
+ "learning_rate": 8.344080808080809e-05,
2095
+ "loss": 1.7279,
2096
+ "step": 174000
2097
+ },
2098
+ {
2099
+ "epoch": 3.99,
2100
+ "learning_rate": 8.339030303030303e-05,
2101
+ "loss": 1.7258,
2102
+ "step": 174500
2103
+ },
2104
+ {
2105
+ "epoch": 4.0,
2106
+ "learning_rate": 8.333979797979799e-05,
2107
+ "loss": 1.7249,
2108
+ "step": 175000
2109
+ },
2110
+ {
2111
+ "epoch": 4.01,
2112
+ "learning_rate": 8.328939393939394e-05,
2113
+ "loss": 1.7266,
2114
+ "step": 175500
2115
+ },
2116
+ {
2117
+ "epoch": 4.03,
2118
+ "learning_rate": 8.32389898989899e-05,
2119
+ "loss": 1.7213,
2120
+ "step": 176000
2121
+ },
2122
+ {
2123
+ "epoch": 4.04,
2124
+ "learning_rate": 8.318848484848485e-05,
2125
+ "loss": 1.7227,
2126
+ "step": 176500
2127
+ },
2128
+ {
2129
+ "epoch": 4.05,
2130
+ "learning_rate": 8.31379797979798e-05,
2131
+ "loss": 1.7208,
2132
+ "step": 177000
2133
+ },
2134
+ {
2135
+ "epoch": 4.07,
2136
+ "learning_rate": 8.308747474747474e-05,
2137
+ "loss": 1.7235,
2138
+ "step": 177500
2139
+ },
2140
+ {
2141
+ "epoch": 4.08,
2142
+ "learning_rate": 8.30370707070707e-05,
2143
+ "loss": 1.7204,
2144
+ "step": 178000
2145
+ },
2146
+ {
2147
+ "epoch": 4.09,
2148
+ "learning_rate": 8.298656565656566e-05,
2149
+ "loss": 1.7214,
2150
+ "step": 178500
2151
+ },
2152
+ {
2153
+ "epoch": 4.11,
2154
+ "learning_rate": 8.293606060606061e-05,
2155
+ "loss": 1.7182,
2156
+ "step": 179000
2157
+ },
2158
+ {
2159
+ "epoch": 4.12,
2160
+ "learning_rate": 8.288555555555556e-05,
2161
+ "loss": 1.7178,
2162
+ "step": 179500
2163
+ },
2164
+ {
2165
+ "epoch": 4.13,
2166
+ "learning_rate": 8.28350505050505e-05,
2167
+ "loss": 1.7176,
2168
+ "step": 180000
2169
+ },
2170
+ {
2171
+ "epoch": 4.15,
2172
+ "learning_rate": 8.278454545454546e-05,
2173
+ "loss": 1.7194,
2174
+ "step": 180500
2175
+ },
2176
+ {
2177
+ "epoch": 4.16,
2178
+ "learning_rate": 8.273414141414142e-05,
2179
+ "loss": 1.7195,
2180
+ "step": 181000
2181
+ },
2182
+ {
2183
+ "epoch": 4.17,
2184
+ "learning_rate": 8.268373737373738e-05,
2185
+ "loss": 1.7167,
2186
+ "step": 181500
2187
+ },
2188
+ {
2189
+ "epoch": 4.19,
2190
+ "learning_rate": 8.263323232323233e-05,
2191
+ "loss": 1.7161,
2192
+ "step": 182000
2193
+ },
2194
+ {
2195
+ "epoch": 4.2,
2196
+ "learning_rate": 8.258272727272727e-05,
2197
+ "loss": 1.7162,
2198
+ "step": 182500
2199
+ },
2200
+ {
2201
+ "epoch": 4.21,
2202
+ "learning_rate": 8.253222222222222e-05,
2203
+ "loss": 1.716,
2204
+ "step": 183000
2205
+ },
2206
+ {
2207
+ "epoch": 4.23,
2208
+ "learning_rate": 8.248171717171718e-05,
2209
+ "loss": 1.7145,
2210
+ "step": 183500
2211
+ },
2212
+ {
2213
+ "epoch": 4.24,
2214
+ "learning_rate": 8.243131313131314e-05,
2215
+ "loss": 1.7113,
2216
+ "step": 184000
2217
+ },
2218
+ {
2219
+ "epoch": 4.25,
2220
+ "learning_rate": 8.238080808080809e-05,
2221
+ "loss": 1.7151,
2222
+ "step": 184500
2223
+ },
2224
+ {
2225
+ "epoch": 4.27,
2226
+ "learning_rate": 8.233030303030303e-05,
2227
+ "loss": 1.7118,
2228
+ "step": 185000
2229
+ },
2230
+ {
2231
+ "epoch": 4.28,
2232
+ "learning_rate": 8.227979797979798e-05,
2233
+ "loss": 1.7147,
2234
+ "step": 185500
2235
+ },
2236
+ {
2237
+ "epoch": 4.29,
2238
+ "learning_rate": 8.222929292929294e-05,
2239
+ "loss": 1.7144,
2240
+ "step": 186000
2241
+ },
2242
+ {
2243
+ "epoch": 4.31,
2244
+ "learning_rate": 8.217878787878787e-05,
2245
+ "loss": 1.7127,
2246
+ "step": 186500
2247
+ },
2248
+ {
2249
+ "epoch": 4.32,
2250
+ "learning_rate": 8.212828282828283e-05,
2251
+ "loss": 1.7103,
2252
+ "step": 187000
2253
+ },
2254
+ {
2255
+ "epoch": 4.33,
2256
+ "learning_rate": 8.207777777777779e-05,
2257
+ "loss": 1.7132,
2258
+ "step": 187500
2259
+ },
2260
+ {
2261
+ "epoch": 4.35,
2262
+ "learning_rate": 8.202737373737374e-05,
2263
+ "loss": 1.7088,
2264
+ "step": 188000
2265
+ },
2266
+ {
2267
+ "epoch": 4.36,
2268
+ "learning_rate": 8.19768686868687e-05,
2269
+ "loss": 1.7088,
2270
+ "step": 188500
2271
+ },
2272
+ {
2273
+ "epoch": 4.37,
2274
+ "learning_rate": 8.192636363636364e-05,
2275
+ "loss": 1.7094,
2276
+ "step": 189000
2277
+ },
2278
+ {
2279
+ "epoch": 4.39,
2280
+ "learning_rate": 8.187585858585859e-05,
2281
+ "loss": 1.709,
2282
+ "step": 189500
2283
+ },
2284
+ {
2285
+ "epoch": 4.4,
2286
+ "learning_rate": 8.182535353535355e-05,
2287
+ "loss": 1.7098,
2288
+ "step": 190000
2289
+ },
2290
+ {
2291
+ "epoch": 4.41,
2292
+ "learning_rate": 8.177505050505051e-05,
2293
+ "loss": 1.7083,
2294
+ "step": 190500
2295
+ },
2296
+ {
2297
+ "epoch": 4.43,
2298
+ "learning_rate": 8.172454545454546e-05,
2299
+ "loss": 1.7063,
2300
+ "step": 191000
2301
+ },
2302
+ {
2303
+ "epoch": 4.44,
2304
+ "learning_rate": 8.167404040404042e-05,
2305
+ "loss": 1.7083,
2306
+ "step": 191500
2307
+ },
2308
+ {
2309
+ "epoch": 4.45,
2310
+ "learning_rate": 8.162353535353535e-05,
2311
+ "loss": 1.7079,
2312
+ "step": 192000
2313
+ },
2314
+ {
2315
+ "epoch": 4.47,
2316
+ "learning_rate": 8.157303030303031e-05,
2317
+ "loss": 1.7069,
2318
+ "step": 192500
2319
+ },
2320
+ {
2321
+ "epoch": 4.48,
2322
+ "learning_rate": 8.152252525252527e-05,
2323
+ "loss": 1.7064,
2324
+ "step": 193000
2325
+ },
2326
+ {
2327
+ "epoch": 4.49,
2328
+ "learning_rate": 8.14720202020202e-05,
2329
+ "loss": 1.7037,
2330
+ "step": 193500
2331
+ },
2332
+ {
2333
+ "epoch": 4.51,
2334
+ "learning_rate": 8.142151515151516e-05,
2335
+ "loss": 1.7043,
2336
+ "step": 194000
2337
+ },
2338
+ {
2339
+ "epoch": 4.52,
2340
+ "learning_rate": 8.13710101010101e-05,
2341
+ "loss": 1.7044,
2342
+ "step": 194500
2343
+ },
2344
+ {
2345
+ "epoch": 4.53,
2346
+ "learning_rate": 8.132050505050505e-05,
2347
+ "loss": 1.7019,
2348
+ "step": 195000
2349
+ },
2350
+ {
2351
+ "epoch": 4.55,
2352
+ "learning_rate": 8.127010101010101e-05,
2353
+ "loss": 1.7043,
2354
+ "step": 195500
2355
+ },
2356
+ {
2357
+ "epoch": 4.56,
2358
+ "learning_rate": 8.121959595959596e-05,
2359
+ "loss": 1.7025,
2360
+ "step": 196000
2361
+ },
2362
+ {
2363
+ "epoch": 4.57,
2364
+ "learning_rate": 8.116909090909092e-05,
2365
+ "loss": 1.7028,
2366
+ "step": 196500
2367
+ },
2368
+ {
2369
+ "epoch": 4.59,
2370
+ "learning_rate": 8.111858585858587e-05,
2371
+ "loss": 1.6991,
2372
+ "step": 197000
2373
+ },
2374
+ {
2375
+ "epoch": 4.6,
2376
+ "learning_rate": 8.106808080808081e-05,
2377
+ "loss": 1.7023,
2378
+ "step": 197500
2379
+ },
2380
+ {
2381
+ "epoch": 4.62,
2382
+ "learning_rate": 8.101757575757576e-05,
2383
+ "loss": 1.7004,
2384
+ "step": 198000
2385
+ },
2386
+ {
2387
+ "epoch": 4.63,
2388
+ "learning_rate": 8.096707070707072e-05,
2389
+ "loss": 1.7012,
2390
+ "step": 198500
2391
+ },
2392
+ {
2393
+ "epoch": 4.64,
2394
+ "learning_rate": 8.091656565656565e-05,
2395
+ "loss": 1.6983,
2396
+ "step": 199000
2397
+ },
2398
+ {
2399
+ "epoch": 4.66,
2400
+ "learning_rate": 8.086616161616162e-05,
2401
+ "loss": 1.7011,
2402
+ "step": 199500
2403
+ },
2404
+ {
2405
+ "epoch": 4.67,
2406
+ "learning_rate": 8.081565656565657e-05,
2407
+ "loss": 1.6972,
2408
+ "step": 200000
2409
+ }
2410
+ ],
2411
+ "max_steps": 200000,
2412
+ "num_train_epochs": 6,
2413
+ "total_flos": 1.3476043989572387e+19,
2414
+ "trial_name": null,
2415
+ "trial_params": null
2416
+ }