farleyknight commited on
Commit
098c00d
1 Parent(s): 509154b

End of training

Browse files
Files changed (4) hide show
  1. all_results.json +12 -0
  2. eval_results.json +8 -0
  3. train_results.json +7 -0
  4. trainer_state.json +1933 -0
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.9923333333333333,
4
+ "eval_loss": 0.031892526894807816,
5
+ "eval_runtime": 112.6018,
6
+ "eval_samples_per_second": 79.928,
7
+ "eval_steps_per_second": 9.991,
8
+ "train_loss": 0.3782404615215227,
9
+ "train_runtime": 8280.4212,
10
+ "train_samples_per_second": 30.796,
11
+ "train_steps_per_second": 3.849
12
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.9923333333333333,
4
+ "eval_loss": 0.031892526894807816,
5
+ "eval_runtime": 112.6018,
6
+ "eval_samples_per_second": 79.928,
7
+ "eval_steps_per_second": 9.991
8
+ }
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "train_loss": 0.3782404615215227,
4
+ "train_runtime": 8280.4212,
5
+ "train_samples_per_second": 30.796,
6
+ "train_steps_per_second": 3.849
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1933 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 5.0,
5
+ "global_step": 31875,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.02,
12
+ "learning_rate": 1.9937254901960785e-05,
13
+ "loss": 2.0694,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 0.03,
18
+ "learning_rate": 1.987450980392157e-05,
19
+ "loss": 1.6002,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 0.05,
24
+ "learning_rate": 1.9811764705882354e-05,
25
+ "loss": 1.3103,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 0.06,
30
+ "learning_rate": 1.9749019607843137e-05,
31
+ "loss": 1.0879,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 0.08,
36
+ "learning_rate": 1.9686274509803924e-05,
37
+ "loss": 0.9775,
38
+ "step": 500
39
+ },
40
+ {
41
+ "epoch": 0.09,
42
+ "learning_rate": 1.9623529411764707e-05,
43
+ "loss": 0.8361,
44
+ "step": 600
45
+ },
46
+ {
47
+ "epoch": 0.11,
48
+ "learning_rate": 1.9560784313725493e-05,
49
+ "loss": 0.8576,
50
+ "step": 700
51
+ },
52
+ {
53
+ "epoch": 0.13,
54
+ "learning_rate": 1.9498039215686276e-05,
55
+ "loss": 0.7407,
56
+ "step": 800
57
+ },
58
+ {
59
+ "epoch": 0.14,
60
+ "learning_rate": 1.9435294117647063e-05,
61
+ "loss": 0.6984,
62
+ "step": 900
63
+ },
64
+ {
65
+ "epoch": 0.16,
66
+ "learning_rate": 1.9372549019607846e-05,
67
+ "loss": 0.6481,
68
+ "step": 1000
69
+ },
70
+ {
71
+ "epoch": 0.17,
72
+ "learning_rate": 1.930980392156863e-05,
73
+ "loss": 0.6464,
74
+ "step": 1100
75
+ },
76
+ {
77
+ "epoch": 0.19,
78
+ "learning_rate": 1.9247058823529415e-05,
79
+ "loss": 0.6199,
80
+ "step": 1200
81
+ },
82
+ {
83
+ "epoch": 0.2,
84
+ "learning_rate": 1.9184313725490198e-05,
85
+ "loss": 0.6207,
86
+ "step": 1300
87
+ },
88
+ {
89
+ "epoch": 0.22,
90
+ "learning_rate": 1.912156862745098e-05,
91
+ "loss": 0.5565,
92
+ "step": 1400
93
+ },
94
+ {
95
+ "epoch": 0.24,
96
+ "learning_rate": 1.9058823529411764e-05,
97
+ "loss": 0.5897,
98
+ "step": 1500
99
+ },
100
+ {
101
+ "epoch": 0.25,
102
+ "learning_rate": 1.899607843137255e-05,
103
+ "loss": 0.6142,
104
+ "step": 1600
105
+ },
106
+ {
107
+ "epoch": 0.27,
108
+ "learning_rate": 1.8933333333333334e-05,
109
+ "loss": 0.4906,
110
+ "step": 1700
111
+ },
112
+ {
113
+ "epoch": 0.28,
114
+ "learning_rate": 1.8870588235294117e-05,
115
+ "loss": 0.5228,
116
+ "step": 1800
117
+ },
118
+ {
119
+ "epoch": 0.3,
120
+ "learning_rate": 1.8807843137254903e-05,
121
+ "loss": 0.4749,
122
+ "step": 1900
123
+ },
124
+ {
125
+ "epoch": 0.31,
126
+ "learning_rate": 1.8745098039215686e-05,
127
+ "loss": 0.5168,
128
+ "step": 2000
129
+ },
130
+ {
131
+ "epoch": 0.33,
132
+ "learning_rate": 1.8682352941176473e-05,
133
+ "loss": 0.4819,
134
+ "step": 2100
135
+ },
136
+ {
137
+ "epoch": 0.35,
138
+ "learning_rate": 1.8619607843137256e-05,
139
+ "loss": 0.4997,
140
+ "step": 2200
141
+ },
142
+ {
143
+ "epoch": 0.36,
144
+ "learning_rate": 1.8556862745098042e-05,
145
+ "loss": 0.4594,
146
+ "step": 2300
147
+ },
148
+ {
149
+ "epoch": 0.38,
150
+ "learning_rate": 1.8494117647058825e-05,
151
+ "loss": 0.4877,
152
+ "step": 2400
153
+ },
154
+ {
155
+ "epoch": 0.39,
156
+ "learning_rate": 1.843137254901961e-05,
157
+ "loss": 0.4424,
158
+ "step": 2500
159
+ },
160
+ {
161
+ "epoch": 0.41,
162
+ "learning_rate": 1.8368627450980395e-05,
163
+ "loss": 0.488,
164
+ "step": 2600
165
+ },
166
+ {
167
+ "epoch": 0.42,
168
+ "learning_rate": 1.8305882352941178e-05,
169
+ "loss": 0.4658,
170
+ "step": 2700
171
+ },
172
+ {
173
+ "epoch": 0.44,
174
+ "learning_rate": 1.824313725490196e-05,
175
+ "loss": 0.48,
176
+ "step": 2800
177
+ },
178
+ {
179
+ "epoch": 0.45,
180
+ "learning_rate": 1.8180392156862747e-05,
181
+ "loss": 0.4641,
182
+ "step": 2900
183
+ },
184
+ {
185
+ "epoch": 0.47,
186
+ "learning_rate": 1.811764705882353e-05,
187
+ "loss": 0.4891,
188
+ "step": 3000
189
+ },
190
+ {
191
+ "epoch": 0.49,
192
+ "learning_rate": 1.8054901960784314e-05,
193
+ "loss": 0.4572,
194
+ "step": 3100
195
+ },
196
+ {
197
+ "epoch": 0.5,
198
+ "learning_rate": 1.79921568627451e-05,
199
+ "loss": 0.4948,
200
+ "step": 3200
201
+ },
202
+ {
203
+ "epoch": 0.52,
204
+ "learning_rate": 1.7929411764705883e-05,
205
+ "loss": 0.4909,
206
+ "step": 3300
207
+ },
208
+ {
209
+ "epoch": 0.53,
210
+ "learning_rate": 1.7866666666666666e-05,
211
+ "loss": 0.4564,
212
+ "step": 3400
213
+ },
214
+ {
215
+ "epoch": 0.55,
216
+ "learning_rate": 1.7803921568627453e-05,
217
+ "loss": 0.4435,
218
+ "step": 3500
219
+ },
220
+ {
221
+ "epoch": 0.56,
222
+ "learning_rate": 1.7741176470588236e-05,
223
+ "loss": 0.439,
224
+ "step": 3600
225
+ },
226
+ {
227
+ "epoch": 0.58,
228
+ "learning_rate": 1.7678431372549022e-05,
229
+ "loss": 0.4855,
230
+ "step": 3700
231
+ },
232
+ {
233
+ "epoch": 0.6,
234
+ "learning_rate": 1.7615686274509805e-05,
235
+ "loss": 0.5032,
236
+ "step": 3800
237
+ },
238
+ {
239
+ "epoch": 0.61,
240
+ "learning_rate": 1.755294117647059e-05,
241
+ "loss": 0.4067,
242
+ "step": 3900
243
+ },
244
+ {
245
+ "epoch": 0.63,
246
+ "learning_rate": 1.7490196078431375e-05,
247
+ "loss": 0.4537,
248
+ "step": 4000
249
+ },
250
+ {
251
+ "epoch": 0.64,
252
+ "learning_rate": 1.7427450980392158e-05,
253
+ "loss": 0.4495,
254
+ "step": 4100
255
+ },
256
+ {
257
+ "epoch": 0.66,
258
+ "learning_rate": 1.7364705882352944e-05,
259
+ "loss": 0.4586,
260
+ "step": 4200
261
+ },
262
+ {
263
+ "epoch": 0.67,
264
+ "learning_rate": 1.7301960784313727e-05,
265
+ "loss": 0.4648,
266
+ "step": 4300
267
+ },
268
+ {
269
+ "epoch": 0.69,
270
+ "learning_rate": 1.723921568627451e-05,
271
+ "loss": 0.4285,
272
+ "step": 4400
273
+ },
274
+ {
275
+ "epoch": 0.71,
276
+ "learning_rate": 1.7176470588235293e-05,
277
+ "loss": 0.4062,
278
+ "step": 4500
279
+ },
280
+ {
281
+ "epoch": 0.72,
282
+ "learning_rate": 1.711372549019608e-05,
283
+ "loss": 0.4103,
284
+ "step": 4600
285
+ },
286
+ {
287
+ "epoch": 0.74,
288
+ "learning_rate": 1.7050980392156863e-05,
289
+ "loss": 0.3933,
290
+ "step": 4700
291
+ },
292
+ {
293
+ "epoch": 0.75,
294
+ "learning_rate": 1.698823529411765e-05,
295
+ "loss": 0.3962,
296
+ "step": 4800
297
+ },
298
+ {
299
+ "epoch": 0.77,
300
+ "learning_rate": 1.6925490196078432e-05,
301
+ "loss": 0.3899,
302
+ "step": 4900
303
+ },
304
+ {
305
+ "epoch": 0.78,
306
+ "learning_rate": 1.686274509803922e-05,
307
+ "loss": 0.4506,
308
+ "step": 5000
309
+ },
310
+ {
311
+ "epoch": 0.8,
312
+ "learning_rate": 1.6800000000000002e-05,
313
+ "loss": 0.3806,
314
+ "step": 5100
315
+ },
316
+ {
317
+ "epoch": 0.82,
318
+ "learning_rate": 1.6737254901960788e-05,
319
+ "loss": 0.4416,
320
+ "step": 5200
321
+ },
322
+ {
323
+ "epoch": 0.83,
324
+ "learning_rate": 1.667450980392157e-05,
325
+ "loss": 0.397,
326
+ "step": 5300
327
+ },
328
+ {
329
+ "epoch": 0.85,
330
+ "learning_rate": 1.6611764705882354e-05,
331
+ "loss": 0.3912,
332
+ "step": 5400
333
+ },
334
+ {
335
+ "epoch": 0.86,
336
+ "learning_rate": 1.6549019607843137e-05,
337
+ "loss": 0.4465,
338
+ "step": 5500
339
+ },
340
+ {
341
+ "epoch": 0.88,
342
+ "learning_rate": 1.6486274509803924e-05,
343
+ "loss": 0.4295,
344
+ "step": 5600
345
+ },
346
+ {
347
+ "epoch": 0.89,
348
+ "learning_rate": 1.6423529411764707e-05,
349
+ "loss": 0.4045,
350
+ "step": 5700
351
+ },
352
+ {
353
+ "epoch": 0.91,
354
+ "learning_rate": 1.636078431372549e-05,
355
+ "loss": 0.4475,
356
+ "step": 5800
357
+ },
358
+ {
359
+ "epoch": 0.93,
360
+ "learning_rate": 1.6298039215686276e-05,
361
+ "loss": 0.4162,
362
+ "step": 5900
363
+ },
364
+ {
365
+ "epoch": 0.94,
366
+ "learning_rate": 1.623529411764706e-05,
367
+ "loss": 0.4437,
368
+ "step": 6000
369
+ },
370
+ {
371
+ "epoch": 0.96,
372
+ "learning_rate": 1.6172549019607842e-05,
373
+ "loss": 0.4248,
374
+ "step": 6100
375
+ },
376
+ {
377
+ "epoch": 0.97,
378
+ "learning_rate": 1.610980392156863e-05,
379
+ "loss": 0.4062,
380
+ "step": 6200
381
+ },
382
+ {
383
+ "epoch": 0.99,
384
+ "learning_rate": 1.6047058823529412e-05,
385
+ "loss": 0.3908,
386
+ "step": 6300
387
+ },
388
+ {
389
+ "epoch": 1.0,
390
+ "learning_rate": 1.59843137254902e-05,
391
+ "loss": 0.3811,
392
+ "step": 6400
393
+ },
394
+ {
395
+ "epoch": 1.02,
396
+ "learning_rate": 1.592156862745098e-05,
397
+ "loss": 0.3818,
398
+ "step": 6500
399
+ },
400
+ {
401
+ "epoch": 1.04,
402
+ "learning_rate": 1.5858823529411768e-05,
403
+ "loss": 0.3628,
404
+ "step": 6600
405
+ },
406
+ {
407
+ "epoch": 1.05,
408
+ "learning_rate": 1.579607843137255e-05,
409
+ "loss": 0.3784,
410
+ "step": 6700
411
+ },
412
+ {
413
+ "epoch": 1.07,
414
+ "learning_rate": 1.5733333333333334e-05,
415
+ "loss": 0.411,
416
+ "step": 6800
417
+ },
418
+ {
419
+ "epoch": 1.08,
420
+ "learning_rate": 1.567058823529412e-05,
421
+ "loss": 0.3647,
422
+ "step": 6900
423
+ },
424
+ {
425
+ "epoch": 1.1,
426
+ "learning_rate": 1.5607843137254904e-05,
427
+ "loss": 0.4586,
428
+ "step": 7000
429
+ },
430
+ {
431
+ "epoch": 1.11,
432
+ "learning_rate": 1.5545098039215687e-05,
433
+ "loss": 0.387,
434
+ "step": 7100
435
+ },
436
+ {
437
+ "epoch": 1.13,
438
+ "learning_rate": 1.5482352941176473e-05,
439
+ "loss": 0.4048,
440
+ "step": 7200
441
+ },
442
+ {
443
+ "epoch": 1.15,
444
+ "learning_rate": 1.5419607843137256e-05,
445
+ "loss": 0.387,
446
+ "step": 7300
447
+ },
448
+ {
449
+ "epoch": 1.16,
450
+ "learning_rate": 1.535686274509804e-05,
451
+ "loss": 0.4028,
452
+ "step": 7400
453
+ },
454
+ {
455
+ "epoch": 1.18,
456
+ "learning_rate": 1.5294117647058822e-05,
457
+ "loss": 0.4103,
458
+ "step": 7500
459
+ },
460
+ {
461
+ "epoch": 1.19,
462
+ "learning_rate": 1.523137254901961e-05,
463
+ "loss": 0.4362,
464
+ "step": 7600
465
+ },
466
+ {
467
+ "epoch": 1.21,
468
+ "learning_rate": 1.5168627450980393e-05,
469
+ "loss": 0.3912,
470
+ "step": 7700
471
+ },
472
+ {
473
+ "epoch": 1.22,
474
+ "learning_rate": 1.5105882352941176e-05,
475
+ "loss": 0.3622,
476
+ "step": 7800
477
+ },
478
+ {
479
+ "epoch": 1.24,
480
+ "learning_rate": 1.5043137254901963e-05,
481
+ "loss": 0.3955,
482
+ "step": 7900
483
+ },
484
+ {
485
+ "epoch": 1.25,
486
+ "learning_rate": 1.4980392156862746e-05,
487
+ "loss": 0.4162,
488
+ "step": 8000
489
+ },
490
+ {
491
+ "epoch": 1.27,
492
+ "learning_rate": 1.491764705882353e-05,
493
+ "loss": 0.4388,
494
+ "step": 8100
495
+ },
496
+ {
497
+ "epoch": 1.29,
498
+ "learning_rate": 1.4854901960784315e-05,
499
+ "loss": 0.4293,
500
+ "step": 8200
501
+ },
502
+ {
503
+ "epoch": 1.3,
504
+ "learning_rate": 1.47921568627451e-05,
505
+ "loss": 0.3754,
506
+ "step": 8300
507
+ },
508
+ {
509
+ "epoch": 1.32,
510
+ "learning_rate": 1.4729411764705883e-05,
511
+ "loss": 0.4118,
512
+ "step": 8400
513
+ },
514
+ {
515
+ "epoch": 1.33,
516
+ "learning_rate": 1.4666666666666666e-05,
517
+ "loss": 0.3456,
518
+ "step": 8500
519
+ },
520
+ {
521
+ "epoch": 1.35,
522
+ "learning_rate": 1.4603921568627453e-05,
523
+ "loss": 0.3869,
524
+ "step": 8600
525
+ },
526
+ {
527
+ "epoch": 1.36,
528
+ "learning_rate": 1.4541176470588236e-05,
529
+ "loss": 0.3624,
530
+ "step": 8700
531
+ },
532
+ {
533
+ "epoch": 1.38,
534
+ "learning_rate": 1.447843137254902e-05,
535
+ "loss": 0.3944,
536
+ "step": 8800
537
+ },
538
+ {
539
+ "epoch": 1.4,
540
+ "learning_rate": 1.4415686274509805e-05,
541
+ "loss": 0.3635,
542
+ "step": 8900
543
+ },
544
+ {
545
+ "epoch": 1.41,
546
+ "learning_rate": 1.435294117647059e-05,
547
+ "loss": 0.3683,
548
+ "step": 9000
549
+ },
550
+ {
551
+ "epoch": 1.43,
552
+ "learning_rate": 1.4290196078431373e-05,
553
+ "loss": 0.3839,
554
+ "step": 9100
555
+ },
556
+ {
557
+ "epoch": 1.44,
558
+ "learning_rate": 1.422745098039216e-05,
559
+ "loss": 0.3285,
560
+ "step": 9200
561
+ },
562
+ {
563
+ "epoch": 1.46,
564
+ "learning_rate": 1.4164705882352943e-05,
565
+ "loss": 0.3834,
566
+ "step": 9300
567
+ },
568
+ {
569
+ "epoch": 1.47,
570
+ "learning_rate": 1.4101960784313726e-05,
571
+ "loss": 0.3725,
572
+ "step": 9400
573
+ },
574
+ {
575
+ "epoch": 1.49,
576
+ "learning_rate": 1.403921568627451e-05,
577
+ "loss": 0.3469,
578
+ "step": 9500
579
+ },
580
+ {
581
+ "epoch": 1.51,
582
+ "learning_rate": 1.3976470588235295e-05,
583
+ "loss": 0.3885,
584
+ "step": 9600
585
+ },
586
+ {
587
+ "epoch": 1.52,
588
+ "learning_rate": 1.391372549019608e-05,
589
+ "loss": 0.3915,
590
+ "step": 9700
591
+ },
592
+ {
593
+ "epoch": 1.54,
594
+ "learning_rate": 1.3850980392156863e-05,
595
+ "loss": 0.3262,
596
+ "step": 9800
597
+ },
598
+ {
599
+ "epoch": 1.55,
600
+ "learning_rate": 1.378823529411765e-05,
601
+ "loss": 0.4212,
602
+ "step": 9900
603
+ },
604
+ {
605
+ "epoch": 1.57,
606
+ "learning_rate": 1.3725490196078432e-05,
607
+ "loss": 0.3592,
608
+ "step": 10000
609
+ },
610
+ {
611
+ "epoch": 1.58,
612
+ "learning_rate": 1.3662745098039215e-05,
613
+ "loss": 0.3903,
614
+ "step": 10100
615
+ },
616
+ {
617
+ "epoch": 1.6,
618
+ "learning_rate": 1.3600000000000002e-05,
619
+ "loss": 0.3342,
620
+ "step": 10200
621
+ },
622
+ {
623
+ "epoch": 1.62,
624
+ "learning_rate": 1.3537254901960785e-05,
625
+ "loss": 0.3085,
626
+ "step": 10300
627
+ },
628
+ {
629
+ "epoch": 1.63,
630
+ "learning_rate": 1.347450980392157e-05,
631
+ "loss": 0.44,
632
+ "step": 10400
633
+ },
634
+ {
635
+ "epoch": 1.65,
636
+ "learning_rate": 1.3411764705882353e-05,
637
+ "loss": 0.3954,
638
+ "step": 10500
639
+ },
640
+ {
641
+ "epoch": 1.66,
642
+ "learning_rate": 1.334901960784314e-05,
643
+ "loss": 0.4343,
644
+ "step": 10600
645
+ },
646
+ {
647
+ "epoch": 1.68,
648
+ "learning_rate": 1.3286274509803922e-05,
649
+ "loss": 0.3674,
650
+ "step": 10700
651
+ },
652
+ {
653
+ "epoch": 1.69,
654
+ "learning_rate": 1.3223529411764705e-05,
655
+ "loss": 0.4191,
656
+ "step": 10800
657
+ },
658
+ {
659
+ "epoch": 1.71,
660
+ "learning_rate": 1.3160784313725492e-05,
661
+ "loss": 0.3616,
662
+ "step": 10900
663
+ },
664
+ {
665
+ "epoch": 1.73,
666
+ "learning_rate": 1.3098039215686275e-05,
667
+ "loss": 0.3234,
668
+ "step": 11000
669
+ },
670
+ {
671
+ "epoch": 1.74,
672
+ "learning_rate": 1.303529411764706e-05,
673
+ "loss": 0.3509,
674
+ "step": 11100
675
+ },
676
+ {
677
+ "epoch": 1.76,
678
+ "learning_rate": 1.2972549019607846e-05,
679
+ "loss": 0.3745,
680
+ "step": 11200
681
+ },
682
+ {
683
+ "epoch": 1.77,
684
+ "learning_rate": 1.2909803921568629e-05,
685
+ "loss": 0.4175,
686
+ "step": 11300
687
+ },
688
+ {
689
+ "epoch": 1.79,
690
+ "learning_rate": 1.2847058823529412e-05,
691
+ "loss": 0.3901,
692
+ "step": 11400
693
+ },
694
+ {
695
+ "epoch": 1.8,
696
+ "learning_rate": 1.2784313725490197e-05,
697
+ "loss": 0.3952,
698
+ "step": 11500
699
+ },
700
+ {
701
+ "epoch": 1.82,
702
+ "learning_rate": 1.2721568627450982e-05,
703
+ "loss": 0.3517,
704
+ "step": 11600
705
+ },
706
+ {
707
+ "epoch": 1.84,
708
+ "learning_rate": 1.2658823529411766e-05,
709
+ "loss": 0.3546,
710
+ "step": 11700
711
+ },
712
+ {
713
+ "epoch": 1.85,
714
+ "learning_rate": 1.259607843137255e-05,
715
+ "loss": 0.3923,
716
+ "step": 11800
717
+ },
718
+ {
719
+ "epoch": 1.87,
720
+ "learning_rate": 1.2533333333333336e-05,
721
+ "loss": 0.3655,
722
+ "step": 11900
723
+ },
724
+ {
725
+ "epoch": 1.88,
726
+ "learning_rate": 1.2470588235294119e-05,
727
+ "loss": 0.3382,
728
+ "step": 12000
729
+ },
730
+ {
731
+ "epoch": 1.9,
732
+ "learning_rate": 1.2407843137254902e-05,
733
+ "loss": 0.2665,
734
+ "step": 12100
735
+ },
736
+ {
737
+ "epoch": 1.91,
738
+ "learning_rate": 1.2345098039215688e-05,
739
+ "loss": 0.3314,
740
+ "step": 12200
741
+ },
742
+ {
743
+ "epoch": 1.93,
744
+ "learning_rate": 1.2282352941176471e-05,
745
+ "loss": 0.3094,
746
+ "step": 12300
747
+ },
748
+ {
749
+ "epoch": 1.95,
750
+ "learning_rate": 1.2219607843137256e-05,
751
+ "loss": 0.4045,
752
+ "step": 12400
753
+ },
754
+ {
755
+ "epoch": 1.96,
756
+ "learning_rate": 1.215686274509804e-05,
757
+ "loss": 0.3164,
758
+ "step": 12500
759
+ },
760
+ {
761
+ "epoch": 1.98,
762
+ "learning_rate": 1.2094117647058826e-05,
763
+ "loss": 0.3195,
764
+ "step": 12600
765
+ },
766
+ {
767
+ "epoch": 1.99,
768
+ "learning_rate": 1.2031372549019609e-05,
769
+ "loss": 0.3494,
770
+ "step": 12700
771
+ },
772
+ {
773
+ "epoch": 2.01,
774
+ "learning_rate": 1.1968627450980392e-05,
775
+ "loss": 0.3757,
776
+ "step": 12800
777
+ },
778
+ {
779
+ "epoch": 2.02,
780
+ "learning_rate": 1.1905882352941178e-05,
781
+ "loss": 0.3308,
782
+ "step": 12900
783
+ },
784
+ {
785
+ "epoch": 2.04,
786
+ "learning_rate": 1.1843137254901961e-05,
787
+ "loss": 0.4156,
788
+ "step": 13000
789
+ },
790
+ {
791
+ "epoch": 2.05,
792
+ "learning_rate": 1.1780392156862746e-05,
793
+ "loss": 0.341,
794
+ "step": 13100
795
+ },
796
+ {
797
+ "epoch": 2.07,
798
+ "learning_rate": 1.171764705882353e-05,
799
+ "loss": 0.3565,
800
+ "step": 13200
801
+ },
802
+ {
803
+ "epoch": 2.09,
804
+ "learning_rate": 1.1654901960784316e-05,
805
+ "loss": 0.3887,
806
+ "step": 13300
807
+ },
808
+ {
809
+ "epoch": 2.1,
810
+ "learning_rate": 1.1592156862745099e-05,
811
+ "loss": 0.349,
812
+ "step": 13400
813
+ },
814
+ {
815
+ "epoch": 2.12,
816
+ "learning_rate": 1.1529411764705882e-05,
817
+ "loss": 0.2906,
818
+ "step": 13500
819
+ },
820
+ {
821
+ "epoch": 2.13,
822
+ "learning_rate": 1.1466666666666668e-05,
823
+ "loss": 0.3764,
824
+ "step": 13600
825
+ },
826
+ {
827
+ "epoch": 2.15,
828
+ "learning_rate": 1.1403921568627451e-05,
829
+ "loss": 0.3684,
830
+ "step": 13700
831
+ },
832
+ {
833
+ "epoch": 2.16,
834
+ "learning_rate": 1.1341176470588236e-05,
835
+ "loss": 0.3345,
836
+ "step": 13800
837
+ },
838
+ {
839
+ "epoch": 2.18,
840
+ "learning_rate": 1.127843137254902e-05,
841
+ "loss": 0.2847,
842
+ "step": 13900
843
+ },
844
+ {
845
+ "epoch": 2.2,
846
+ "learning_rate": 1.1215686274509805e-05,
847
+ "loss": 0.3356,
848
+ "step": 14000
849
+ },
850
+ {
851
+ "epoch": 2.21,
852
+ "learning_rate": 1.1152941176470588e-05,
853
+ "loss": 0.3043,
854
+ "step": 14100
855
+ },
856
+ {
857
+ "epoch": 2.23,
858
+ "learning_rate": 1.1090196078431375e-05,
859
+ "loss": 0.3741,
860
+ "step": 14200
861
+ },
862
+ {
863
+ "epoch": 2.24,
864
+ "learning_rate": 1.1027450980392158e-05,
865
+ "loss": 0.4068,
866
+ "step": 14300
867
+ },
868
+ {
869
+ "epoch": 2.26,
870
+ "learning_rate": 1.0964705882352941e-05,
871
+ "loss": 0.3698,
872
+ "step": 14400
873
+ },
874
+ {
875
+ "epoch": 2.27,
876
+ "learning_rate": 1.0901960784313726e-05,
877
+ "loss": 0.3253,
878
+ "step": 14500
879
+ },
880
+ {
881
+ "epoch": 2.29,
882
+ "learning_rate": 1.083921568627451e-05,
883
+ "loss": 0.3467,
884
+ "step": 14600
885
+ },
886
+ {
887
+ "epoch": 2.31,
888
+ "learning_rate": 1.0776470588235295e-05,
889
+ "loss": 0.3182,
890
+ "step": 14700
891
+ },
892
+ {
893
+ "epoch": 2.32,
894
+ "learning_rate": 1.0713725490196078e-05,
895
+ "loss": 0.4068,
896
+ "step": 14800
897
+ },
898
+ {
899
+ "epoch": 2.34,
900
+ "learning_rate": 1.0650980392156865e-05,
901
+ "loss": 0.3652,
902
+ "step": 14900
903
+ },
904
+ {
905
+ "epoch": 2.35,
906
+ "learning_rate": 1.0588235294117648e-05,
907
+ "loss": 0.3778,
908
+ "step": 15000
909
+ },
910
+ {
911
+ "epoch": 2.37,
912
+ "learning_rate": 1.0525490196078431e-05,
913
+ "loss": 0.3413,
914
+ "step": 15100
915
+ },
916
+ {
917
+ "epoch": 2.38,
918
+ "learning_rate": 1.0462745098039217e-05,
919
+ "loss": 0.3286,
920
+ "step": 15200
921
+ },
922
+ {
923
+ "epoch": 2.4,
924
+ "learning_rate": 1.04e-05,
925
+ "loss": 0.338,
926
+ "step": 15300
927
+ },
928
+ {
929
+ "epoch": 2.42,
930
+ "learning_rate": 1.0337254901960785e-05,
931
+ "loss": 0.2998,
932
+ "step": 15400
933
+ },
934
+ {
935
+ "epoch": 2.43,
936
+ "learning_rate": 1.0274509803921568e-05,
937
+ "loss": 0.3354,
938
+ "step": 15500
939
+ },
940
+ {
941
+ "epoch": 2.45,
942
+ "learning_rate": 1.0211764705882355e-05,
943
+ "loss": 0.3238,
944
+ "step": 15600
945
+ },
946
+ {
947
+ "epoch": 2.46,
948
+ "learning_rate": 1.0149019607843138e-05,
949
+ "loss": 0.3393,
950
+ "step": 15700
951
+ },
952
+ {
953
+ "epoch": 2.48,
954
+ "learning_rate": 1.0086274509803922e-05,
955
+ "loss": 0.3575,
956
+ "step": 15800
957
+ },
958
+ {
959
+ "epoch": 2.49,
960
+ "learning_rate": 1.0023529411764707e-05,
961
+ "loss": 0.3044,
962
+ "step": 15900
963
+ },
964
+ {
965
+ "epoch": 2.51,
966
+ "learning_rate": 9.960784313725492e-06,
967
+ "loss": 0.3337,
968
+ "step": 16000
969
+ },
970
+ {
971
+ "epoch": 2.53,
972
+ "learning_rate": 9.898039215686275e-06,
973
+ "loss": 0.2956,
974
+ "step": 16100
975
+ },
976
+ {
977
+ "epoch": 2.54,
978
+ "learning_rate": 9.83529411764706e-06,
979
+ "loss": 0.3222,
980
+ "step": 16200
981
+ },
982
+ {
983
+ "epoch": 2.56,
984
+ "learning_rate": 9.772549019607844e-06,
985
+ "loss": 0.3692,
986
+ "step": 16300
987
+ },
988
+ {
989
+ "epoch": 2.57,
990
+ "learning_rate": 9.709803921568628e-06,
991
+ "loss": 0.3526,
992
+ "step": 16400
993
+ },
994
+ {
995
+ "epoch": 2.59,
996
+ "learning_rate": 9.647058823529412e-06,
997
+ "loss": 0.2933,
998
+ "step": 16500
999
+ },
1000
+ {
1001
+ "epoch": 2.6,
1002
+ "learning_rate": 9.584313725490197e-06,
1003
+ "loss": 0.2853,
1004
+ "step": 16600
1005
+ },
1006
+ {
1007
+ "epoch": 2.62,
1008
+ "learning_rate": 9.521568627450982e-06,
1009
+ "loss": 0.3719,
1010
+ "step": 16700
1011
+ },
1012
+ {
1013
+ "epoch": 2.64,
1014
+ "learning_rate": 9.458823529411767e-06,
1015
+ "loss": 0.342,
1016
+ "step": 16800
1017
+ },
1018
+ {
1019
+ "epoch": 2.65,
1020
+ "learning_rate": 9.39607843137255e-06,
1021
+ "loss": 0.3587,
1022
+ "step": 16900
1023
+ },
1024
+ {
1025
+ "epoch": 2.67,
1026
+ "learning_rate": 9.333333333333334e-06,
1027
+ "loss": 0.3679,
1028
+ "step": 17000
1029
+ },
1030
+ {
1031
+ "epoch": 2.68,
1032
+ "learning_rate": 9.270588235294117e-06,
1033
+ "loss": 0.3299,
1034
+ "step": 17100
1035
+ },
1036
+ {
1037
+ "epoch": 2.7,
1038
+ "learning_rate": 9.207843137254902e-06,
1039
+ "loss": 0.2607,
1040
+ "step": 17200
1041
+ },
1042
+ {
1043
+ "epoch": 2.71,
1044
+ "learning_rate": 9.145098039215687e-06,
1045
+ "loss": 0.2938,
1046
+ "step": 17300
1047
+ },
1048
+ {
1049
+ "epoch": 2.73,
1050
+ "learning_rate": 9.082352941176472e-06,
1051
+ "loss": 0.338,
1052
+ "step": 17400
1053
+ },
1054
+ {
1055
+ "epoch": 2.75,
1056
+ "learning_rate": 9.019607843137256e-06,
1057
+ "loss": 0.3054,
1058
+ "step": 17500
1059
+ },
1060
+ {
1061
+ "epoch": 2.76,
1062
+ "learning_rate": 8.95686274509804e-06,
1063
+ "loss": 0.3437,
1064
+ "step": 17600
1065
+ },
1066
+ {
1067
+ "epoch": 2.78,
1068
+ "learning_rate": 8.894117647058824e-06,
1069
+ "loss": 0.2903,
1070
+ "step": 17700
1071
+ },
1072
+ {
1073
+ "epoch": 2.79,
1074
+ "learning_rate": 8.831372549019609e-06,
1075
+ "loss": 0.3338,
1076
+ "step": 17800
1077
+ },
1078
+ {
1079
+ "epoch": 2.81,
1080
+ "learning_rate": 8.768627450980392e-06,
1081
+ "loss": 0.3157,
1082
+ "step": 17900
1083
+ },
1084
+ {
1085
+ "epoch": 2.82,
1086
+ "learning_rate": 8.705882352941177e-06,
1087
+ "loss": 0.3438,
1088
+ "step": 18000
1089
+ },
1090
+ {
1091
+ "epoch": 2.84,
1092
+ "learning_rate": 8.643137254901961e-06,
1093
+ "loss": 0.3587,
1094
+ "step": 18100
1095
+ },
1096
+ {
1097
+ "epoch": 2.85,
1098
+ "learning_rate": 8.580392156862746e-06,
1099
+ "loss": 0.3633,
1100
+ "step": 18200
1101
+ },
1102
+ {
1103
+ "epoch": 2.87,
1104
+ "learning_rate": 8.517647058823531e-06,
1105
+ "loss": 0.3648,
1106
+ "step": 18300
1107
+ },
1108
+ {
1109
+ "epoch": 2.89,
1110
+ "learning_rate": 8.454901960784314e-06,
1111
+ "loss": 0.3227,
1112
+ "step": 18400
1113
+ },
1114
+ {
1115
+ "epoch": 2.9,
1116
+ "learning_rate": 8.392156862745099e-06,
1117
+ "loss": 0.3537,
1118
+ "step": 18500
1119
+ },
1120
+ {
1121
+ "epoch": 2.92,
1122
+ "learning_rate": 8.329411764705882e-06,
1123
+ "loss": 0.2904,
1124
+ "step": 18600
1125
+ },
1126
+ {
1127
+ "epoch": 2.93,
1128
+ "learning_rate": 8.266666666666667e-06,
1129
+ "loss": 0.3319,
1130
+ "step": 18700
1131
+ },
1132
+ {
1133
+ "epoch": 2.95,
1134
+ "learning_rate": 8.203921568627451e-06,
1135
+ "loss": 0.3882,
1136
+ "step": 18800
1137
+ },
1138
+ {
1139
+ "epoch": 2.96,
1140
+ "learning_rate": 8.141176470588236e-06,
1141
+ "loss": 0.3132,
1142
+ "step": 18900
1143
+ },
1144
+ {
1145
+ "epoch": 2.98,
1146
+ "learning_rate": 8.07843137254902e-06,
1147
+ "loss": 0.3319,
1148
+ "step": 19000
1149
+ },
1150
+ {
1151
+ "epoch": 3.0,
1152
+ "learning_rate": 8.015686274509804e-06,
1153
+ "loss": 0.3198,
1154
+ "step": 19100
1155
+ },
1156
+ {
1157
+ "epoch": 3.01,
1158
+ "learning_rate": 7.952941176470589e-06,
1159
+ "loss": 0.3628,
1160
+ "step": 19200
1161
+ },
1162
+ {
1163
+ "epoch": 3.03,
1164
+ "learning_rate": 7.890196078431373e-06,
1165
+ "loss": 0.3394,
1166
+ "step": 19300
1167
+ },
1168
+ {
1169
+ "epoch": 3.04,
1170
+ "learning_rate": 7.827450980392156e-06,
1171
+ "loss": 0.3532,
1172
+ "step": 19400
1173
+ },
1174
+ {
1175
+ "epoch": 3.06,
1176
+ "learning_rate": 7.764705882352941e-06,
1177
+ "loss": 0.2856,
1178
+ "step": 19500
1179
+ },
1180
+ {
1181
+ "epoch": 3.07,
1182
+ "learning_rate": 7.701960784313726e-06,
1183
+ "loss": 0.3187,
1184
+ "step": 19600
1185
+ },
1186
+ {
1187
+ "epoch": 3.09,
1188
+ "learning_rate": 7.63921568627451e-06,
1189
+ "loss": 0.3231,
1190
+ "step": 19700
1191
+ },
1192
+ {
1193
+ "epoch": 3.11,
1194
+ "learning_rate": 7.576470588235295e-06,
1195
+ "loss": 0.3511,
1196
+ "step": 19800
1197
+ },
1198
+ {
1199
+ "epoch": 3.12,
1200
+ "learning_rate": 7.5137254901960785e-06,
1201
+ "loss": 0.3409,
1202
+ "step": 19900
1203
+ },
1204
+ {
1205
+ "epoch": 3.14,
1206
+ "learning_rate": 7.450980392156863e-06,
1207
+ "loss": 0.2692,
1208
+ "step": 20000
1209
+ },
1210
+ {
1211
+ "epoch": 3.15,
1212
+ "learning_rate": 7.388235294117647e-06,
1213
+ "loss": 0.2757,
1214
+ "step": 20100
1215
+ },
1216
+ {
1217
+ "epoch": 3.17,
1218
+ "learning_rate": 7.325490196078432e-06,
1219
+ "loss": 0.3208,
1220
+ "step": 20200
1221
+ },
1222
+ {
1223
+ "epoch": 3.18,
1224
+ "learning_rate": 7.262745098039217e-06,
1225
+ "loss": 0.3469,
1226
+ "step": 20300
1227
+ },
1228
+ {
1229
+ "epoch": 3.2,
1230
+ "learning_rate": 7.2000000000000005e-06,
1231
+ "loss": 0.3503,
1232
+ "step": 20400
1233
+ },
1234
+ {
1235
+ "epoch": 3.22,
1236
+ "learning_rate": 7.137254901960785e-06,
1237
+ "loss": 0.2889,
1238
+ "step": 20500
1239
+ },
1240
+ {
1241
+ "epoch": 3.23,
1242
+ "learning_rate": 7.074509803921568e-06,
1243
+ "loss": 0.3715,
1244
+ "step": 20600
1245
+ },
1246
+ {
1247
+ "epoch": 3.25,
1248
+ "learning_rate": 7.011764705882353e-06,
1249
+ "loss": 0.3443,
1250
+ "step": 20700
1251
+ },
1252
+ {
1253
+ "epoch": 3.26,
1254
+ "learning_rate": 6.949019607843138e-06,
1255
+ "loss": 0.3416,
1256
+ "step": 20800
1257
+ },
1258
+ {
1259
+ "epoch": 3.28,
1260
+ "learning_rate": 6.886274509803922e-06,
1261
+ "loss": 0.2901,
1262
+ "step": 20900
1263
+ },
1264
+ {
1265
+ "epoch": 3.29,
1266
+ "learning_rate": 6.8235294117647065e-06,
1267
+ "loss": 0.3579,
1268
+ "step": 21000
1269
+ },
1270
+ {
1271
+ "epoch": 3.31,
1272
+ "learning_rate": 6.76078431372549e-06,
1273
+ "loss": 0.2908,
1274
+ "step": 21100
1275
+ },
1276
+ {
1277
+ "epoch": 3.33,
1278
+ "learning_rate": 6.698039215686275e-06,
1279
+ "loss": 0.3403,
1280
+ "step": 21200
1281
+ },
1282
+ {
1283
+ "epoch": 3.34,
1284
+ "learning_rate": 6.63529411764706e-06,
1285
+ "loss": 0.2988,
1286
+ "step": 21300
1287
+ },
1288
+ {
1289
+ "epoch": 3.36,
1290
+ "learning_rate": 6.572549019607844e-06,
1291
+ "loss": 0.2714,
1292
+ "step": 21400
1293
+ },
1294
+ {
1295
+ "epoch": 3.37,
1296
+ "learning_rate": 6.5098039215686285e-06,
1297
+ "loss": 0.2819,
1298
+ "step": 21500
1299
+ },
1300
+ {
1301
+ "epoch": 3.39,
1302
+ "learning_rate": 6.4470588235294116e-06,
1303
+ "loss": 0.3536,
1304
+ "step": 21600
1305
+ },
1306
+ {
1307
+ "epoch": 3.4,
1308
+ "learning_rate": 6.384313725490196e-06,
1309
+ "loss": 0.3073,
1310
+ "step": 21700
1311
+ },
1312
+ {
1313
+ "epoch": 3.42,
1314
+ "learning_rate": 6.321568627450981e-06,
1315
+ "loss": 0.3564,
1316
+ "step": 21800
1317
+ },
1318
+ {
1319
+ "epoch": 3.44,
1320
+ "learning_rate": 6.258823529411765e-06,
1321
+ "loss": 0.3839,
1322
+ "step": 21900
1323
+ },
1324
+ {
1325
+ "epoch": 3.45,
1326
+ "learning_rate": 6.19607843137255e-06,
1327
+ "loss": 0.277,
1328
+ "step": 22000
1329
+ },
1330
+ {
1331
+ "epoch": 3.47,
1332
+ "learning_rate": 6.133333333333334e-06,
1333
+ "loss": 0.3017,
1334
+ "step": 22100
1335
+ },
1336
+ {
1337
+ "epoch": 3.48,
1338
+ "learning_rate": 6.070588235294118e-06,
1339
+ "loss": 0.295,
1340
+ "step": 22200
1341
+ },
1342
+ {
1343
+ "epoch": 3.5,
1344
+ "learning_rate": 6.007843137254903e-06,
1345
+ "loss": 0.3126,
1346
+ "step": 22300
1347
+ },
1348
+ {
1349
+ "epoch": 3.51,
1350
+ "learning_rate": 5.945098039215686e-06,
1351
+ "loss": 0.3122,
1352
+ "step": 22400
1353
+ },
1354
+ {
1355
+ "epoch": 3.53,
1356
+ "learning_rate": 5.882352941176471e-06,
1357
+ "loss": 0.3131,
1358
+ "step": 22500
1359
+ },
1360
+ {
1361
+ "epoch": 3.55,
1362
+ "learning_rate": 5.819607843137255e-06,
1363
+ "loss": 0.3171,
1364
+ "step": 22600
1365
+ },
1366
+ {
1367
+ "epoch": 3.56,
1368
+ "learning_rate": 5.7568627450980396e-06,
1369
+ "loss": 0.2643,
1370
+ "step": 22700
1371
+ },
1372
+ {
1373
+ "epoch": 3.58,
1374
+ "learning_rate": 5.694117647058824e-06,
1375
+ "loss": 0.2745,
1376
+ "step": 22800
1377
+ },
1378
+ {
1379
+ "epoch": 3.59,
1380
+ "learning_rate": 5.631372549019608e-06,
1381
+ "loss": 0.2548,
1382
+ "step": 22900
1383
+ },
1384
+ {
1385
+ "epoch": 3.61,
1386
+ "learning_rate": 5.568627450980393e-06,
1387
+ "loss": 0.3525,
1388
+ "step": 23000
1389
+ },
1390
+ {
1391
+ "epoch": 3.62,
1392
+ "learning_rate": 5.505882352941177e-06,
1393
+ "loss": 0.2849,
1394
+ "step": 23100
1395
+ },
1396
+ {
1397
+ "epoch": 3.64,
1398
+ "learning_rate": 5.443137254901962e-06,
1399
+ "loss": 0.294,
1400
+ "step": 23200
1401
+ },
1402
+ {
1403
+ "epoch": 3.65,
1404
+ "learning_rate": 5.380392156862746e-06,
1405
+ "loss": 0.2729,
1406
+ "step": 23300
1407
+ },
1408
+ {
1409
+ "epoch": 3.67,
1410
+ "learning_rate": 5.317647058823529e-06,
1411
+ "loss": 0.2731,
1412
+ "step": 23400
1413
+ },
1414
+ {
1415
+ "epoch": 3.69,
1416
+ "learning_rate": 5.254901960784314e-06,
1417
+ "loss": 0.3299,
1418
+ "step": 23500
1419
+ },
1420
+ {
1421
+ "epoch": 3.7,
1422
+ "learning_rate": 5.192156862745098e-06,
1423
+ "loss": 0.2769,
1424
+ "step": 23600
1425
+ },
1426
+ {
1427
+ "epoch": 3.72,
1428
+ "learning_rate": 5.129411764705883e-06,
1429
+ "loss": 0.34,
1430
+ "step": 23700
1431
+ },
1432
+ {
1433
+ "epoch": 3.73,
1434
+ "learning_rate": 5.0666666666666676e-06,
1435
+ "loss": 0.338,
1436
+ "step": 23800
1437
+ },
1438
+ {
1439
+ "epoch": 3.75,
1440
+ "learning_rate": 5.0039215686274515e-06,
1441
+ "loss": 0.304,
1442
+ "step": 23900
1443
+ },
1444
+ {
1445
+ "epoch": 3.76,
1446
+ "learning_rate": 4.941176470588236e-06,
1447
+ "loss": 0.2602,
1448
+ "step": 24000
1449
+ },
1450
+ {
1451
+ "epoch": 3.78,
1452
+ "learning_rate": 4.87843137254902e-06,
1453
+ "loss": 0.3564,
1454
+ "step": 24100
1455
+ },
1456
+ {
1457
+ "epoch": 3.8,
1458
+ "learning_rate": 4.815686274509804e-06,
1459
+ "loss": 0.3455,
1460
+ "step": 24200
1461
+ },
1462
+ {
1463
+ "epoch": 3.81,
1464
+ "learning_rate": 4.752941176470589e-06,
1465
+ "loss": 0.3651,
1466
+ "step": 24300
1467
+ },
1468
+ {
1469
+ "epoch": 3.83,
1470
+ "learning_rate": 4.690196078431373e-06,
1471
+ "loss": 0.3341,
1472
+ "step": 24400
1473
+ },
1474
+ {
1475
+ "epoch": 3.84,
1476
+ "learning_rate": 4.627450980392157e-06,
1477
+ "loss": 0.2999,
1478
+ "step": 24500
1479
+ },
1480
+ {
1481
+ "epoch": 3.86,
1482
+ "learning_rate": 4.564705882352941e-06,
1483
+ "loss": 0.3324,
1484
+ "step": 24600
1485
+ },
1486
+ {
1487
+ "epoch": 3.87,
1488
+ "learning_rate": 4.501960784313726e-06,
1489
+ "loss": 0.3118,
1490
+ "step": 24700
1491
+ },
1492
+ {
1493
+ "epoch": 3.89,
1494
+ "learning_rate": 4.43921568627451e-06,
1495
+ "loss": 0.3328,
1496
+ "step": 24800
1497
+ },
1498
+ {
1499
+ "epoch": 3.91,
1500
+ "learning_rate": 4.376470588235294e-06,
1501
+ "loss": 0.2887,
1502
+ "step": 24900
1503
+ },
1504
+ {
1505
+ "epoch": 3.92,
1506
+ "learning_rate": 4.313725490196079e-06,
1507
+ "loss": 0.2785,
1508
+ "step": 25000
1509
+ },
1510
+ {
1511
+ "epoch": 3.94,
1512
+ "learning_rate": 4.250980392156863e-06,
1513
+ "loss": 0.3577,
1514
+ "step": 25100
1515
+ },
1516
+ {
1517
+ "epoch": 3.95,
1518
+ "learning_rate": 4.188235294117647e-06,
1519
+ "loss": 0.2535,
1520
+ "step": 25200
1521
+ },
1522
+ {
1523
+ "epoch": 3.97,
1524
+ "learning_rate": 4.125490196078432e-06,
1525
+ "loss": 0.3511,
1526
+ "step": 25300
1527
+ },
1528
+ {
1529
+ "epoch": 3.98,
1530
+ "learning_rate": 4.062745098039216e-06,
1531
+ "loss": 0.2844,
1532
+ "step": 25400
1533
+ },
1534
+ {
1535
+ "epoch": 4.0,
1536
+ "learning_rate": 4.000000000000001e-06,
1537
+ "loss": 0.3132,
1538
+ "step": 25500
1539
+ },
1540
+ {
1541
+ "epoch": 4.02,
1542
+ "learning_rate": 3.9372549019607846e-06,
1543
+ "loss": 0.321,
1544
+ "step": 25600
1545
+ },
1546
+ {
1547
+ "epoch": 4.03,
1548
+ "learning_rate": 3.874509803921569e-06,
1549
+ "loss": 0.3647,
1550
+ "step": 25700
1551
+ },
1552
+ {
1553
+ "epoch": 4.05,
1554
+ "learning_rate": 3.8117647058823532e-06,
1555
+ "loss": 0.2704,
1556
+ "step": 25800
1557
+ },
1558
+ {
1559
+ "epoch": 4.06,
1560
+ "learning_rate": 3.7490196078431375e-06,
1561
+ "loss": 0.2823,
1562
+ "step": 25900
1563
+ },
1564
+ {
1565
+ "epoch": 4.08,
1566
+ "learning_rate": 3.6862745098039223e-06,
1567
+ "loss": 0.3143,
1568
+ "step": 26000
1569
+ },
1570
+ {
1571
+ "epoch": 4.09,
1572
+ "learning_rate": 3.623529411764706e-06,
1573
+ "loss": 0.3528,
1574
+ "step": 26100
1575
+ },
1576
+ {
1577
+ "epoch": 4.11,
1578
+ "learning_rate": 3.5607843137254905e-06,
1579
+ "loss": 0.2732,
1580
+ "step": 26200
1581
+ },
1582
+ {
1583
+ "epoch": 4.13,
1584
+ "learning_rate": 3.498039215686275e-06,
1585
+ "loss": 0.337,
1586
+ "step": 26300
1587
+ },
1588
+ {
1589
+ "epoch": 4.14,
1590
+ "learning_rate": 3.4352941176470587e-06,
1591
+ "loss": 0.239,
1592
+ "step": 26400
1593
+ },
1594
+ {
1595
+ "epoch": 4.16,
1596
+ "learning_rate": 3.3725490196078435e-06,
1597
+ "loss": 0.3503,
1598
+ "step": 26500
1599
+ },
1600
+ {
1601
+ "epoch": 4.17,
1602
+ "learning_rate": 3.309803921568628e-06,
1603
+ "loss": 0.3103,
1604
+ "step": 26600
1605
+ },
1606
+ {
1607
+ "epoch": 4.19,
1608
+ "learning_rate": 3.247058823529412e-06,
1609
+ "loss": 0.2942,
1610
+ "step": 26700
1611
+ },
1612
+ {
1613
+ "epoch": 4.2,
1614
+ "learning_rate": 3.184313725490196e-06,
1615
+ "loss": 0.3624,
1616
+ "step": 26800
1617
+ },
1618
+ {
1619
+ "epoch": 4.22,
1620
+ "learning_rate": 3.1215686274509804e-06,
1621
+ "loss": 0.2862,
1622
+ "step": 26900
1623
+ },
1624
+ {
1625
+ "epoch": 4.24,
1626
+ "learning_rate": 3.058823529411765e-06,
1627
+ "loss": 0.2289,
1628
+ "step": 27000
1629
+ },
1630
+ {
1631
+ "epoch": 4.25,
1632
+ "learning_rate": 2.9960784313725494e-06,
1633
+ "loss": 0.3352,
1634
+ "step": 27100
1635
+ },
1636
+ {
1637
+ "epoch": 4.27,
1638
+ "learning_rate": 2.9333333333333338e-06,
1639
+ "loss": 0.3705,
1640
+ "step": 27200
1641
+ },
1642
+ {
1643
+ "epoch": 4.28,
1644
+ "learning_rate": 2.8705882352941177e-06,
1645
+ "loss": 0.2936,
1646
+ "step": 27300
1647
+ },
1648
+ {
1649
+ "epoch": 4.3,
1650
+ "learning_rate": 2.807843137254902e-06,
1651
+ "loss": 0.2536,
1652
+ "step": 27400
1653
+ },
1654
+ {
1655
+ "epoch": 4.31,
1656
+ "learning_rate": 2.7450980392156867e-06,
1657
+ "loss": 0.2545,
1658
+ "step": 27500
1659
+ },
1660
+ {
1661
+ "epoch": 4.33,
1662
+ "learning_rate": 2.682352941176471e-06,
1663
+ "loss": 0.2769,
1664
+ "step": 27600
1665
+ },
1666
+ {
1667
+ "epoch": 4.35,
1668
+ "learning_rate": 2.619607843137255e-06,
1669
+ "loss": 0.2874,
1670
+ "step": 27700
1671
+ },
1672
+ {
1673
+ "epoch": 4.36,
1674
+ "learning_rate": 2.5568627450980393e-06,
1675
+ "loss": 0.3563,
1676
+ "step": 27800
1677
+ },
1678
+ {
1679
+ "epoch": 4.38,
1680
+ "learning_rate": 2.4941176470588236e-06,
1681
+ "loss": 0.2348,
1682
+ "step": 27900
1683
+ },
1684
+ {
1685
+ "epoch": 4.39,
1686
+ "learning_rate": 2.431372549019608e-06,
1687
+ "loss": 0.3834,
1688
+ "step": 28000
1689
+ },
1690
+ {
1691
+ "epoch": 4.41,
1692
+ "learning_rate": 2.3686274509803927e-06,
1693
+ "loss": 0.3343,
1694
+ "step": 28100
1695
+ },
1696
+ {
1697
+ "epoch": 4.42,
1698
+ "learning_rate": 2.3058823529411766e-06,
1699
+ "loss": 0.3367,
1700
+ "step": 28200
1701
+ },
1702
+ {
1703
+ "epoch": 4.44,
1704
+ "learning_rate": 2.243137254901961e-06,
1705
+ "loss": 0.3008,
1706
+ "step": 28300
1707
+ },
1708
+ {
1709
+ "epoch": 4.45,
1710
+ "learning_rate": 2.1803921568627452e-06,
1711
+ "loss": 0.2658,
1712
+ "step": 28400
1713
+ },
1714
+ {
1715
+ "epoch": 4.47,
1716
+ "learning_rate": 2.1176470588235296e-06,
1717
+ "loss": 0.2769,
1718
+ "step": 28500
1719
+ },
1720
+ {
1721
+ "epoch": 4.49,
1722
+ "learning_rate": 2.054901960784314e-06,
1723
+ "loss": 0.3286,
1724
+ "step": 28600
1725
+ },
1726
+ {
1727
+ "epoch": 4.5,
1728
+ "learning_rate": 1.9921568627450982e-06,
1729
+ "loss": 0.281,
1730
+ "step": 28700
1731
+ },
1732
+ {
1733
+ "epoch": 4.52,
1734
+ "learning_rate": 1.9294117647058825e-06,
1735
+ "loss": 0.3042,
1736
+ "step": 28800
1737
+ },
1738
+ {
1739
+ "epoch": 4.53,
1740
+ "learning_rate": 1.8666666666666669e-06,
1741
+ "loss": 0.3194,
1742
+ "step": 28900
1743
+ },
1744
+ {
1745
+ "epoch": 4.55,
1746
+ "learning_rate": 1.8039215686274512e-06,
1747
+ "loss": 0.2526,
1748
+ "step": 29000
1749
+ },
1750
+ {
1751
+ "epoch": 4.56,
1752
+ "learning_rate": 1.7411764705882353e-06,
1753
+ "loss": 0.3558,
1754
+ "step": 29100
1755
+ },
1756
+ {
1757
+ "epoch": 4.58,
1758
+ "learning_rate": 1.6784313725490198e-06,
1759
+ "loss": 0.2889,
1760
+ "step": 29200
1761
+ },
1762
+ {
1763
+ "epoch": 4.6,
1764
+ "learning_rate": 1.615686274509804e-06,
1765
+ "loss": 0.3368,
1766
+ "step": 29300
1767
+ },
1768
+ {
1769
+ "epoch": 4.61,
1770
+ "learning_rate": 1.5529411764705885e-06,
1771
+ "loss": 0.2736,
1772
+ "step": 29400
1773
+ },
1774
+ {
1775
+ "epoch": 4.63,
1776
+ "learning_rate": 1.4901960784313726e-06,
1777
+ "loss": 0.3034,
1778
+ "step": 29500
1779
+ },
1780
+ {
1781
+ "epoch": 4.64,
1782
+ "learning_rate": 1.427450980392157e-06,
1783
+ "loss": 0.2773,
1784
+ "step": 29600
1785
+ },
1786
+ {
1787
+ "epoch": 4.66,
1788
+ "learning_rate": 1.3647058823529413e-06,
1789
+ "loss": 0.2741,
1790
+ "step": 29700
1791
+ },
1792
+ {
1793
+ "epoch": 4.67,
1794
+ "learning_rate": 1.3019607843137256e-06,
1795
+ "loss": 0.3051,
1796
+ "step": 29800
1797
+ },
1798
+ {
1799
+ "epoch": 4.69,
1800
+ "learning_rate": 1.23921568627451e-06,
1801
+ "loss": 0.3271,
1802
+ "step": 29900
1803
+ },
1804
+ {
1805
+ "epoch": 4.71,
1806
+ "learning_rate": 1.1764705882352942e-06,
1807
+ "loss": 0.2903,
1808
+ "step": 30000
1809
+ },
1810
+ {
1811
+ "epoch": 4.72,
1812
+ "learning_rate": 1.1137254901960786e-06,
1813
+ "loss": 0.3264,
1814
+ "step": 30100
1815
+ },
1816
+ {
1817
+ "epoch": 4.74,
1818
+ "learning_rate": 1.0509803921568629e-06,
1819
+ "loss": 0.2549,
1820
+ "step": 30200
1821
+ },
1822
+ {
1823
+ "epoch": 4.75,
1824
+ "learning_rate": 9.88235294117647e-07,
1825
+ "loss": 0.2874,
1826
+ "step": 30300
1827
+ },
1828
+ {
1829
+ "epoch": 4.77,
1830
+ "learning_rate": 9.254901960784314e-07,
1831
+ "loss": 0.2807,
1832
+ "step": 30400
1833
+ },
1834
+ {
1835
+ "epoch": 4.78,
1836
+ "learning_rate": 8.627450980392157e-07,
1837
+ "loss": 0.326,
1838
+ "step": 30500
1839
+ },
1840
+ {
1841
+ "epoch": 4.8,
1842
+ "learning_rate": 8.000000000000001e-07,
1843
+ "loss": 0.2802,
1844
+ "step": 30600
1845
+ },
1846
+ {
1847
+ "epoch": 4.82,
1848
+ "learning_rate": 7.372549019607844e-07,
1849
+ "loss": 0.2875,
1850
+ "step": 30700
1851
+ },
1852
+ {
1853
+ "epoch": 4.83,
1854
+ "learning_rate": 6.745098039215686e-07,
1855
+ "loss": 0.2507,
1856
+ "step": 30800
1857
+ },
1858
+ {
1859
+ "epoch": 4.85,
1860
+ "learning_rate": 6.11764705882353e-07,
1861
+ "loss": 0.3013,
1862
+ "step": 30900
1863
+ },
1864
+ {
1865
+ "epoch": 4.86,
1866
+ "learning_rate": 5.490196078431373e-07,
1867
+ "loss": 0.298,
1868
+ "step": 31000
1869
+ },
1870
+ {
1871
+ "epoch": 4.88,
1872
+ "learning_rate": 4.862745098039216e-07,
1873
+ "loss": 0.3424,
1874
+ "step": 31100
1875
+ },
1876
+ {
1877
+ "epoch": 4.89,
1878
+ "learning_rate": 4.235294117647059e-07,
1879
+ "loss": 0.3628,
1880
+ "step": 31200
1881
+ },
1882
+ {
1883
+ "epoch": 4.91,
1884
+ "learning_rate": 3.6078431372549024e-07,
1885
+ "loss": 0.2314,
1886
+ "step": 31300
1887
+ },
1888
+ {
1889
+ "epoch": 4.93,
1890
+ "learning_rate": 2.9803921568627456e-07,
1891
+ "loss": 0.3012,
1892
+ "step": 31400
1893
+ },
1894
+ {
1895
+ "epoch": 4.94,
1896
+ "learning_rate": 2.3529411764705883e-07,
1897
+ "loss": 0.2644,
1898
+ "step": 31500
1899
+ },
1900
+ {
1901
+ "epoch": 4.96,
1902
+ "learning_rate": 1.7254901960784313e-07,
1903
+ "loss": 0.3057,
1904
+ "step": 31600
1905
+ },
1906
+ {
1907
+ "epoch": 4.97,
1908
+ "learning_rate": 1.0980392156862746e-07,
1909
+ "loss": 0.2536,
1910
+ "step": 31700
1911
+ },
1912
+ {
1913
+ "epoch": 4.99,
1914
+ "learning_rate": 4.705882352941176e-08,
1915
+ "loss": 0.2628,
1916
+ "step": 31800
1917
+ },
1918
+ {
1919
+ "epoch": 5.0,
1920
+ "step": 31875,
1921
+ "total_flos": 1.976187420555264e+19,
1922
+ "train_loss": 0.3782404615215227,
1923
+ "train_runtime": 8280.4212,
1924
+ "train_samples_per_second": 30.796,
1925
+ "train_steps_per_second": 3.849
1926
+ }
1927
+ ],
1928
+ "max_steps": 31875,
1929
+ "num_train_epochs": 5,
1930
+ "total_flos": 1.976187420555264e+19,
1931
+ "trial_name": null,
1932
+ "trial_params": null
1933
+ }