vumichien commited on
Commit
ad0dc1d
1 Parent(s): 164dc80

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 +2119 -0
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "eval_accuracy": 0.9149749926448956,
4
+ "eval_loss": 0.30149412155151367,
5
+ "eval_runtime": 148.5016,
6
+ "eval_samples_per_second": 45.777,
7
+ "eval_steps_per_second": 0.721,
8
+ "train_loss": 0.8136612295581911,
9
+ "train_runtime": 27726.7705,
10
+ "train_samples_per_second": 36.855,
11
+ "train_steps_per_second": 1.152
12
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "eval_accuracy": 0.9149749926448956,
4
+ "eval_loss": 0.30149412155151367,
5
+ "eval_runtime": 148.5016,
6
+ "eval_samples_per_second": 45.777,
7
+ "eval_steps_per_second": 0.721
8
+ }
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "train_loss": 0.8136612295581911,
4
+ "train_runtime": 27726.7705,
5
+ "train_samples_per_second": 36.855,
6
+ "train_steps_per_second": 1.152
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,2119 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9149749926448956,
3
+ "best_model_checkpoint": "trillsson3-ft-keyword-spotting-14/checkpoint-23955",
4
+ "epoch": 20.0,
5
+ "global_step": 31940,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.06,
12
+ "learning_rate": 8.73512836568566e-06,
13
+ "loss": 8.0793,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 0.13,
18
+ "learning_rate": 1.812773951158422e-05,
19
+ "loss": 7.1754,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 0.19,
24
+ "learning_rate": 2.7520350657482777e-05,
25
+ "loss": 5.6723,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 0.25,
30
+ "learning_rate": 3.691296180338134e-05,
31
+ "loss": 4.1468,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 0.31,
36
+ "learning_rate": 4.6305572949279896e-05,
37
+ "loss": 3.3576,
38
+ "step": 500
39
+ },
40
+ {
41
+ "epoch": 0.38,
42
+ "learning_rate": 5.5698184095178454e-05,
43
+ "loss": 2.955,
44
+ "step": 600
45
+ },
46
+ {
47
+ "epoch": 0.44,
48
+ "learning_rate": 6.509079524107701e-05,
49
+ "loss": 2.7689,
50
+ "step": 700
51
+ },
52
+ {
53
+ "epoch": 0.5,
54
+ "learning_rate": 7.448340638697557e-05,
55
+ "loss": 2.5858,
56
+ "step": 800
57
+ },
58
+ {
59
+ "epoch": 0.56,
60
+ "learning_rate": 8.387601753287413e-05,
61
+ "loss": 2.3275,
62
+ "step": 900
63
+ },
64
+ {
65
+ "epoch": 0.63,
66
+ "learning_rate": 9.326862867877268e-05,
67
+ "loss": 2.0771,
68
+ "step": 1000
69
+ },
70
+ {
71
+ "epoch": 0.69,
72
+ "learning_rate": 0.00010266123982467126,
73
+ "loss": 1.8958,
74
+ "step": 1100
75
+ },
76
+ {
77
+ "epoch": 0.75,
78
+ "learning_rate": 0.0001120538509705698,
79
+ "loss": 1.6778,
80
+ "step": 1200
81
+ },
82
+ {
83
+ "epoch": 0.81,
84
+ "learning_rate": 0.00012144646211646837,
85
+ "loss": 1.5346,
86
+ "step": 1300
87
+ },
88
+ {
89
+ "epoch": 0.88,
90
+ "learning_rate": 0.00013083907326236693,
91
+ "loss": 1.4238,
92
+ "step": 1400
93
+ },
94
+ {
95
+ "epoch": 0.94,
96
+ "learning_rate": 0.0001402316844082655,
97
+ "loss": 1.2824,
98
+ "step": 1500
99
+ },
100
+ {
101
+ "epoch": 1.0,
102
+ "eval_accuracy": 0.6891732862606649,
103
+ "eval_loss": 0.7817752957344055,
104
+ "eval_runtime": 152.2428,
105
+ "eval_samples_per_second": 44.652,
106
+ "eval_steps_per_second": 0.703,
107
+ "step": 1597
108
+ },
109
+ {
110
+ "epoch": 1.0,
111
+ "learning_rate": 0.00014962429555416404,
112
+ "loss": 1.2121,
113
+ "step": 1600
114
+ },
115
+ {
116
+ "epoch": 1.06,
117
+ "learning_rate": 0.0001590169067000626,
118
+ "loss": 1.1391,
119
+ "step": 1700
120
+ },
121
+ {
122
+ "epoch": 1.13,
123
+ "learning_rate": 0.00016840951784596116,
124
+ "loss": 1.12,
125
+ "step": 1800
126
+ },
127
+ {
128
+ "epoch": 1.19,
129
+ "learning_rate": 0.00017780212899185972,
130
+ "loss": 1.053,
131
+ "step": 1900
132
+ },
133
+ {
134
+ "epoch": 1.25,
135
+ "learning_rate": 0.0001871947401377583,
136
+ "loss": 1.0245,
137
+ "step": 2000
138
+ },
139
+ {
140
+ "epoch": 1.31,
141
+ "learning_rate": 0.00019658735128365683,
142
+ "loss": 0.9887,
143
+ "step": 2100
144
+ },
145
+ {
146
+ "epoch": 1.38,
147
+ "learning_rate": 0.0002059799624295554,
148
+ "loss": 0.9671,
149
+ "step": 2200
150
+ },
151
+ {
152
+ "epoch": 1.44,
153
+ "learning_rate": 0.00021537257357545395,
154
+ "loss": 0.8912,
155
+ "step": 2300
156
+ },
157
+ {
158
+ "epoch": 1.5,
159
+ "learning_rate": 0.00022476518472135253,
160
+ "loss": 0.9015,
161
+ "step": 2400
162
+ },
163
+ {
164
+ "epoch": 1.57,
165
+ "learning_rate": 0.0002341577958672511,
166
+ "loss": 0.89,
167
+ "step": 2500
168
+ },
169
+ {
170
+ "epoch": 1.63,
171
+ "learning_rate": 0.00024355040701314962,
172
+ "loss": 0.8173,
173
+ "step": 2600
174
+ },
175
+ {
176
+ "epoch": 1.69,
177
+ "learning_rate": 0.0002529430181590482,
178
+ "loss": 0.8338,
179
+ "step": 2700
180
+ },
181
+ {
182
+ "epoch": 1.75,
183
+ "learning_rate": 0.00026233562930494674,
184
+ "loss": 0.8313,
185
+ "step": 2800
186
+ },
187
+ {
188
+ "epoch": 1.82,
189
+ "learning_rate": 0.0002717282404508453,
190
+ "loss": 0.8395,
191
+ "step": 2900
192
+ },
193
+ {
194
+ "epoch": 1.88,
195
+ "learning_rate": 0.00028112085159674385,
196
+ "loss": 0.8036,
197
+ "step": 3000
198
+ },
199
+ {
200
+ "epoch": 1.94,
201
+ "learning_rate": 0.00029051346274264244,
202
+ "loss": 0.8003,
203
+ "step": 3100
204
+ },
205
+ {
206
+ "epoch": 2.0,
207
+ "eval_accuracy": 0.8734922035892909,
208
+ "eval_loss": 0.44425612688064575,
209
+ "eval_runtime": 147.8186,
210
+ "eval_samples_per_second": 45.989,
211
+ "eval_steps_per_second": 0.724,
212
+ "step": 3194
213
+ },
214
+ {
215
+ "epoch": 2.0,
216
+ "learning_rate": 0.00029990607388854097,
217
+ "loss": 0.7951,
218
+ "step": 3200
219
+ },
220
+ {
221
+ "epoch": 2.07,
222
+ "learning_rate": 0.0002989668127739511,
223
+ "loss": 0.8114,
224
+ "step": 3300
225
+ },
226
+ {
227
+ "epoch": 2.13,
228
+ "learning_rate": 0.00029792318931329574,
229
+ "loss": 0.8343,
230
+ "step": 3400
231
+ },
232
+ {
233
+ "epoch": 2.19,
234
+ "learning_rate": 0.00029687956585264035,
235
+ "loss": 0.7617,
236
+ "step": 3500
237
+ },
238
+ {
239
+ "epoch": 2.25,
240
+ "learning_rate": 0.00029583594239198497,
241
+ "loss": 0.7779,
242
+ "step": 3600
243
+ },
244
+ {
245
+ "epoch": 2.32,
246
+ "learning_rate": 0.0002947923189313295,
247
+ "loss": 0.7512,
248
+ "step": 3700
249
+ },
250
+ {
251
+ "epoch": 2.38,
252
+ "learning_rate": 0.00029374869547067414,
253
+ "loss": 0.7962,
254
+ "step": 3800
255
+ },
256
+ {
257
+ "epoch": 2.44,
258
+ "learning_rate": 0.00029270507201001875,
259
+ "loss": 0.7134,
260
+ "step": 3900
261
+ },
262
+ {
263
+ "epoch": 2.5,
264
+ "learning_rate": 0.00029166144854936337,
265
+ "loss": 0.786,
266
+ "step": 4000
267
+ },
268
+ {
269
+ "epoch": 2.57,
270
+ "learning_rate": 0.000290617825088708,
271
+ "loss": 0.7808,
272
+ "step": 4100
273
+ },
274
+ {
275
+ "epoch": 2.63,
276
+ "learning_rate": 0.0002895742016280526,
277
+ "loss": 0.7379,
278
+ "step": 4200
279
+ },
280
+ {
281
+ "epoch": 2.69,
282
+ "learning_rate": 0.00028853057816739715,
283
+ "loss": 0.7545,
284
+ "step": 4300
285
+ },
286
+ {
287
+ "epoch": 2.76,
288
+ "learning_rate": 0.00028748695470674177,
289
+ "loss": 0.7557,
290
+ "step": 4400
291
+ },
292
+ {
293
+ "epoch": 2.82,
294
+ "learning_rate": 0.0002864433312460864,
295
+ "loss": 0.7633,
296
+ "step": 4500
297
+ },
298
+ {
299
+ "epoch": 2.88,
300
+ "learning_rate": 0.000285399707785431,
301
+ "loss": 0.7414,
302
+ "step": 4600
303
+ },
304
+ {
305
+ "epoch": 2.94,
306
+ "learning_rate": 0.0002843560843247756,
307
+ "loss": 0.7232,
308
+ "step": 4700
309
+ },
310
+ {
311
+ "epoch": 3.0,
312
+ "eval_accuracy": 0.8833480435422183,
313
+ "eval_loss": 0.372787207365036,
314
+ "eval_runtime": 148.4355,
315
+ "eval_samples_per_second": 45.798,
316
+ "eval_steps_per_second": 0.721,
317
+ "step": 4791
318
+ },
319
+ {
320
+ "epoch": 3.01,
321
+ "learning_rate": 0.00028331246086412017,
322
+ "loss": 0.7661,
323
+ "step": 4800
324
+ },
325
+ {
326
+ "epoch": 3.07,
327
+ "learning_rate": 0.00028226883740346484,
328
+ "loss": 0.7224,
329
+ "step": 4900
330
+ },
331
+ {
332
+ "epoch": 3.13,
333
+ "learning_rate": 0.0002812252139428094,
334
+ "loss": 0.7186,
335
+ "step": 5000
336
+ },
337
+ {
338
+ "epoch": 3.19,
339
+ "learning_rate": 0.0002801920267167606,
340
+ "loss": 0.7151,
341
+ "step": 5100
342
+ },
343
+ {
344
+ "epoch": 3.26,
345
+ "learning_rate": 0.00027914840325610516,
346
+ "loss": 0.7175,
347
+ "step": 5200
348
+ },
349
+ {
350
+ "epoch": 3.32,
351
+ "learning_rate": 0.0002781047797954498,
352
+ "loss": 0.6871,
353
+ "step": 5300
354
+ },
355
+ {
356
+ "epoch": 3.38,
357
+ "learning_rate": 0.0002770611563347944,
358
+ "loss": 0.7446,
359
+ "step": 5400
360
+ },
361
+ {
362
+ "epoch": 3.44,
363
+ "learning_rate": 0.000276017532874139,
364
+ "loss": 0.7128,
365
+ "step": 5500
366
+ },
367
+ {
368
+ "epoch": 3.51,
369
+ "learning_rate": 0.0002749739094134836,
370
+ "loss": 0.6896,
371
+ "step": 5600
372
+ },
373
+ {
374
+ "epoch": 3.57,
375
+ "learning_rate": 0.0002739302859528282,
376
+ "loss": 0.714,
377
+ "step": 5700
378
+ },
379
+ {
380
+ "epoch": 3.63,
381
+ "learning_rate": 0.0002728866624921728,
382
+ "loss": 0.7478,
383
+ "step": 5800
384
+ },
385
+ {
386
+ "epoch": 3.69,
387
+ "learning_rate": 0.0002718430390315174,
388
+ "loss": 0.6853,
389
+ "step": 5900
390
+ },
391
+ {
392
+ "epoch": 3.76,
393
+ "learning_rate": 0.000270799415570862,
394
+ "loss": 0.7165,
395
+ "step": 6000
396
+ },
397
+ {
398
+ "epoch": 3.82,
399
+ "learning_rate": 0.0002697557921102066,
400
+ "loss": 0.704,
401
+ "step": 6100
402
+ },
403
+ {
404
+ "epoch": 3.88,
405
+ "learning_rate": 0.00026871216864955125,
406
+ "loss": 0.6954,
407
+ "step": 6200
408
+ },
409
+ {
410
+ "epoch": 3.94,
411
+ "learning_rate": 0.0002676685451888958,
412
+ "loss": 0.73,
413
+ "step": 6300
414
+ },
415
+ {
416
+ "epoch": 4.0,
417
+ "eval_accuracy": 0.8973227419829362,
418
+ "eval_loss": 0.346542090177536,
419
+ "eval_runtime": 147.4068,
420
+ "eval_samples_per_second": 46.117,
421
+ "eval_steps_per_second": 0.726,
422
+ "step": 6388
423
+ },
424
+ {
425
+ "epoch": 4.01,
426
+ "learning_rate": 0.0002666249217282404,
427
+ "loss": 0.7149,
428
+ "step": 6400
429
+ },
430
+ {
431
+ "epoch": 4.07,
432
+ "learning_rate": 0.00026558129826758503,
433
+ "loss": 0.7107,
434
+ "step": 6500
435
+ },
436
+ {
437
+ "epoch": 4.13,
438
+ "learning_rate": 0.00026453767480692965,
439
+ "loss": 0.6744,
440
+ "step": 6600
441
+ },
442
+ {
443
+ "epoch": 4.2,
444
+ "learning_rate": 0.0002634940513462742,
445
+ "loss": 0.727,
446
+ "step": 6700
447
+ },
448
+ {
449
+ "epoch": 4.26,
450
+ "learning_rate": 0.0002624504278856189,
451
+ "loss": 0.7,
452
+ "step": 6800
453
+ },
454
+ {
455
+ "epoch": 4.32,
456
+ "learning_rate": 0.00026140680442496343,
457
+ "loss": 0.7239,
458
+ "step": 6900
459
+ },
460
+ {
461
+ "epoch": 4.38,
462
+ "learning_rate": 0.00026036318096430805,
463
+ "loss": 0.6847,
464
+ "step": 7000
465
+ },
466
+ {
467
+ "epoch": 4.45,
468
+ "learning_rate": 0.00025931955750365266,
469
+ "loss": 0.7151,
470
+ "step": 7100
471
+ },
472
+ {
473
+ "epoch": 4.51,
474
+ "learning_rate": 0.0002582759340429973,
475
+ "loss": 0.7228,
476
+ "step": 7200
477
+ },
478
+ {
479
+ "epoch": 4.57,
480
+ "learning_rate": 0.0002572427468169484,
481
+ "loss": 0.6925,
482
+ "step": 7300
483
+ },
484
+ {
485
+ "epoch": 4.63,
486
+ "learning_rate": 0.000256199123356293,
487
+ "loss": 0.7064,
488
+ "step": 7400
489
+ },
490
+ {
491
+ "epoch": 4.7,
492
+ "learning_rate": 0.00025515549989563765,
493
+ "loss": 0.6861,
494
+ "step": 7500
495
+ },
496
+ {
497
+ "epoch": 4.76,
498
+ "learning_rate": 0.0002541118764349822,
499
+ "loss": 0.7057,
500
+ "step": 7600
501
+ },
502
+ {
503
+ "epoch": 4.82,
504
+ "learning_rate": 0.0002530682529743269,
505
+ "loss": 0.6811,
506
+ "step": 7700
507
+ },
508
+ {
509
+ "epoch": 4.88,
510
+ "learning_rate": 0.00025202462951367144,
511
+ "loss": 0.6676,
512
+ "step": 7800
513
+ },
514
+ {
515
+ "epoch": 4.95,
516
+ "learning_rate": 0.00025098100605301605,
517
+ "loss": 0.7015,
518
+ "step": 7900
519
+ },
520
+ {
521
+ "epoch": 5.0,
522
+ "eval_accuracy": 0.910856134157105,
523
+ "eval_loss": 0.3211327791213989,
524
+ "eval_runtime": 150.4771,
525
+ "eval_samples_per_second": 45.176,
526
+ "eval_steps_per_second": 0.711,
527
+ "step": 7985
528
+ },
529
+ {
530
+ "epoch": 5.01,
531
+ "learning_rate": 0.00024993738259236067,
532
+ "loss": 0.708,
533
+ "step": 8000
534
+ },
535
+ {
536
+ "epoch": 5.07,
537
+ "learning_rate": 0.0002488937591317053,
538
+ "loss": 0.7372,
539
+ "step": 8100
540
+ },
541
+ {
542
+ "epoch": 5.13,
543
+ "learning_rate": 0.00024785013567104984,
544
+ "loss": 0.7089,
545
+ "step": 8200
546
+ },
547
+ {
548
+ "epoch": 5.2,
549
+ "learning_rate": 0.00024680651221039445,
550
+ "loss": 0.6962,
551
+ "step": 8300
552
+ },
553
+ {
554
+ "epoch": 5.26,
555
+ "learning_rate": 0.00024576288874973907,
556
+ "loss": 0.7209,
557
+ "step": 8400
558
+ },
559
+ {
560
+ "epoch": 5.32,
561
+ "learning_rate": 0.0002447192652890837,
562
+ "loss": 0.7091,
563
+ "step": 8500
564
+ },
565
+ {
566
+ "epoch": 5.39,
567
+ "learning_rate": 0.0002436756418284283,
568
+ "loss": 0.6898,
569
+ "step": 8600
570
+ },
571
+ {
572
+ "epoch": 5.45,
573
+ "learning_rate": 0.00024263201836777288,
574
+ "loss": 0.726,
575
+ "step": 8700
576
+ },
577
+ {
578
+ "epoch": 5.51,
579
+ "learning_rate": 0.00024158839490711747,
580
+ "loss": 0.682,
581
+ "step": 8800
582
+ },
583
+ {
584
+ "epoch": 5.57,
585
+ "learning_rate": 0.0002405447714464621,
586
+ "loss": 0.6855,
587
+ "step": 8900
588
+ },
589
+ {
590
+ "epoch": 5.64,
591
+ "learning_rate": 0.0002395011479858067,
592
+ "loss": 0.7108,
593
+ "step": 9000
594
+ },
595
+ {
596
+ "epoch": 5.7,
597
+ "learning_rate": 0.00023845752452515128,
598
+ "loss": 0.6945,
599
+ "step": 9100
600
+ },
601
+ {
602
+ "epoch": 5.76,
603
+ "learning_rate": 0.00023741390106449592,
604
+ "loss": 0.6621,
605
+ "step": 9200
606
+ },
607
+ {
608
+ "epoch": 5.82,
609
+ "learning_rate": 0.0002363702776038405,
610
+ "loss": 0.7176,
611
+ "step": 9300
612
+ },
613
+ {
614
+ "epoch": 5.89,
615
+ "learning_rate": 0.00023532665414318512,
616
+ "loss": 0.6824,
617
+ "step": 9400
618
+ },
619
+ {
620
+ "epoch": 5.95,
621
+ "learning_rate": 0.00023428303068252974,
622
+ "loss": 0.6981,
623
+ "step": 9500
624
+ },
625
+ {
626
+ "epoch": 6.0,
627
+ "eval_accuracy": 0.9080611944689615,
628
+ "eval_loss": 0.3200249671936035,
629
+ "eval_runtime": 148.3702,
630
+ "eval_samples_per_second": 45.818,
631
+ "eval_steps_per_second": 0.721,
632
+ "step": 9582
633
+ },
634
+ {
635
+ "epoch": 6.01,
636
+ "learning_rate": 0.00023323940722187433,
637
+ "loss": 0.6934,
638
+ "step": 9600
639
+ },
640
+ {
641
+ "epoch": 6.07,
642
+ "learning_rate": 0.00023219578376121894,
643
+ "loss": 0.7101,
644
+ "step": 9700
645
+ },
646
+ {
647
+ "epoch": 6.14,
648
+ "learning_rate": 0.00023115216030056353,
649
+ "loss": 0.6922,
650
+ "step": 9800
651
+ },
652
+ {
653
+ "epoch": 6.2,
654
+ "learning_rate": 0.00023010853683990814,
655
+ "loss": 0.7027,
656
+ "step": 9900
657
+ },
658
+ {
659
+ "epoch": 6.26,
660
+ "learning_rate": 0.00022906491337925275,
661
+ "loss": 0.6931,
662
+ "step": 10000
663
+ },
664
+ {
665
+ "epoch": 6.32,
666
+ "learning_rate": 0.00022802128991859734,
667
+ "loss": 0.6782,
668
+ "step": 10100
669
+ },
670
+ {
671
+ "epoch": 6.39,
672
+ "learning_rate": 0.00022697766645794195,
673
+ "loss": 0.6842,
674
+ "step": 10200
675
+ },
676
+ {
677
+ "epoch": 6.45,
678
+ "learning_rate": 0.00022593404299728657,
679
+ "loss": 0.7295,
680
+ "step": 10300
681
+ },
682
+ {
683
+ "epoch": 6.51,
684
+ "learning_rate": 0.00022489041953663115,
685
+ "loss": 0.7145,
686
+ "step": 10400
687
+ },
688
+ {
689
+ "epoch": 6.57,
690
+ "learning_rate": 0.00022384679607597577,
691
+ "loss": 0.6637,
692
+ "step": 10500
693
+ },
694
+ {
695
+ "epoch": 6.64,
696
+ "learning_rate": 0.00022280317261532038,
697
+ "loss": 0.6666,
698
+ "step": 10600
699
+ },
700
+ {
701
+ "epoch": 6.7,
702
+ "learning_rate": 0.00022175954915466497,
703
+ "loss": 0.7068,
704
+ "step": 10700
705
+ },
706
+ {
707
+ "epoch": 6.76,
708
+ "learning_rate": 0.00022071592569400955,
709
+ "loss": 0.6759,
710
+ "step": 10800
711
+ },
712
+ {
713
+ "epoch": 6.83,
714
+ "learning_rate": 0.0002196723022333542,
715
+ "loss": 0.6926,
716
+ "step": 10900
717
+ },
718
+ {
719
+ "epoch": 6.89,
720
+ "learning_rate": 0.00021862867877269878,
721
+ "loss": 0.6825,
722
+ "step": 11000
723
+ },
724
+ {
725
+ "epoch": 6.95,
726
+ "learning_rate": 0.00021758505531204342,
727
+ "loss": 0.6807,
728
+ "step": 11100
729
+ },
730
+ {
731
+ "epoch": 7.0,
732
+ "eval_accuracy": 0.9058546631362165,
733
+ "eval_loss": 0.3208906650543213,
734
+ "eval_runtime": 149.169,
735
+ "eval_samples_per_second": 45.572,
736
+ "eval_steps_per_second": 0.717,
737
+ "step": 11179
738
+ },
739
+ {
740
+ "epoch": 7.01,
741
+ "learning_rate": 0.000216541431851388,
742
+ "loss": 0.677,
743
+ "step": 11200
744
+ },
745
+ {
746
+ "epoch": 7.08,
747
+ "learning_rate": 0.00021550824462533916,
748
+ "loss": 0.7379,
749
+ "step": 11300
750
+ },
751
+ {
752
+ "epoch": 7.14,
753
+ "learning_rate": 0.00021447505739929034,
754
+ "loss": 0.7156,
755
+ "step": 11400
756
+ },
757
+ {
758
+ "epoch": 7.2,
759
+ "learning_rate": 0.00021343143393863493,
760
+ "loss": 0.6406,
761
+ "step": 11500
762
+ },
763
+ {
764
+ "epoch": 7.26,
765
+ "learning_rate": 0.0002123878104779795,
766
+ "loss": 0.6888,
767
+ "step": 11600
768
+ },
769
+ {
770
+ "epoch": 7.33,
771
+ "learning_rate": 0.00021134418701732413,
772
+ "loss": 0.6968,
773
+ "step": 11700
774
+ },
775
+ {
776
+ "epoch": 7.39,
777
+ "learning_rate": 0.00021030056355666874,
778
+ "loss": 0.6989,
779
+ "step": 11800
780
+ },
781
+ {
782
+ "epoch": 7.45,
783
+ "learning_rate": 0.00020925694009601333,
784
+ "loss": 0.6751,
785
+ "step": 11900
786
+ },
787
+ {
788
+ "epoch": 7.51,
789
+ "learning_rate": 0.00020821331663535794,
790
+ "loss": 0.6879,
791
+ "step": 12000
792
+ },
793
+ {
794
+ "epoch": 7.58,
795
+ "learning_rate": 0.00020716969317470255,
796
+ "loss": 0.678,
797
+ "step": 12100
798
+ },
799
+ {
800
+ "epoch": 7.64,
801
+ "learning_rate": 0.00020612606971404714,
802
+ "loss": 0.6501,
803
+ "step": 12200
804
+ },
805
+ {
806
+ "epoch": 7.7,
807
+ "learning_rate": 0.00020508244625339176,
808
+ "loss": 0.6679,
809
+ "step": 12300
810
+ },
811
+ {
812
+ "epoch": 7.76,
813
+ "learning_rate": 0.00020403882279273637,
814
+ "loss": 0.7116,
815
+ "step": 12400
816
+ },
817
+ {
818
+ "epoch": 7.83,
819
+ "learning_rate": 0.00020299519933208098,
820
+ "loss": 0.6899,
821
+ "step": 12500
822
+ },
823
+ {
824
+ "epoch": 7.89,
825
+ "learning_rate": 0.00020195157587142557,
826
+ "loss": 0.6892,
827
+ "step": 12600
828
+ },
829
+ {
830
+ "epoch": 7.95,
831
+ "learning_rate": 0.00020090795241077016,
832
+ "loss": 0.6873,
833
+ "step": 12700
834
+ },
835
+ {
836
+ "epoch": 8.0,
837
+ "eval_accuracy": 0.902177110914975,
838
+ "eval_loss": 0.3205910921096802,
839
+ "eval_runtime": 147.3348,
840
+ "eval_samples_per_second": 46.14,
841
+ "eval_steps_per_second": 0.726,
842
+ "step": 12776
843
+ },
844
+ {
845
+ "epoch": 8.02,
846
+ "learning_rate": 0.0001998643289501148,
847
+ "loss": 0.6847,
848
+ "step": 12800
849
+ },
850
+ {
851
+ "epoch": 8.08,
852
+ "learning_rate": 0.00019882070548945938,
853
+ "loss": 0.6833,
854
+ "step": 12900
855
+ },
856
+ {
857
+ "epoch": 8.14,
858
+ "learning_rate": 0.00019777708202880397,
859
+ "loss": 0.6737,
860
+ "step": 13000
861
+ },
862
+ {
863
+ "epoch": 8.2,
864
+ "learning_rate": 0.0001967334585681486,
865
+ "loss": 0.6568,
866
+ "step": 13100
867
+ },
868
+ {
869
+ "epoch": 8.27,
870
+ "learning_rate": 0.0001956898351074932,
871
+ "loss": 0.7059,
872
+ "step": 13200
873
+ },
874
+ {
875
+ "epoch": 8.33,
876
+ "learning_rate": 0.00019464621164683778,
877
+ "loss": 0.6831,
878
+ "step": 13300
879
+ },
880
+ {
881
+ "epoch": 8.39,
882
+ "learning_rate": 0.00019360258818618243,
883
+ "loss": 0.7174,
884
+ "step": 13400
885
+ },
886
+ {
887
+ "epoch": 8.45,
888
+ "learning_rate": 0.00019256940096013358,
889
+ "loss": 0.6966,
890
+ "step": 13500
891
+ },
892
+ {
893
+ "epoch": 8.52,
894
+ "learning_rate": 0.00019152577749947816,
895
+ "loss": 0.699,
896
+ "step": 13600
897
+ },
898
+ {
899
+ "epoch": 8.58,
900
+ "learning_rate": 0.00019048215403882278,
901
+ "loss": 0.6774,
902
+ "step": 13700
903
+ },
904
+ {
905
+ "epoch": 8.64,
906
+ "learning_rate": 0.0001894385305781674,
907
+ "loss": 0.6848,
908
+ "step": 13800
909
+ },
910
+ {
911
+ "epoch": 8.7,
912
+ "learning_rate": 0.00018839490711751198,
913
+ "loss": 0.6901,
914
+ "step": 13900
915
+ },
916
+ {
917
+ "epoch": 8.77,
918
+ "learning_rate": 0.00018735128365685656,
919
+ "loss": 0.6994,
920
+ "step": 14000
921
+ },
922
+ {
923
+ "epoch": 8.83,
924
+ "learning_rate": 0.0001863076601962012,
925
+ "loss": 0.6739,
926
+ "step": 14100
927
+ },
928
+ {
929
+ "epoch": 8.89,
930
+ "learning_rate": 0.0001852640367355458,
931
+ "loss": 0.6854,
932
+ "step": 14200
933
+ },
934
+ {
935
+ "epoch": 8.95,
936
+ "learning_rate": 0.00018422041327489038,
937
+ "loss": 0.6416,
938
+ "step": 14300
939
+ },
940
+ {
941
+ "epoch": 9.0,
942
+ "eval_accuracy": 0.9057075610473668,
943
+ "eval_loss": 0.31237688660621643,
944
+ "eval_runtime": 148.9632,
945
+ "eval_samples_per_second": 45.635,
946
+ "eval_steps_per_second": 0.718,
947
+ "step": 14373
948
+ },
949
+ {
950
+ "epoch": 9.02,
951
+ "learning_rate": 0.00018317678981423502,
952
+ "loss": 0.6711,
953
+ "step": 14400
954
+ },
955
+ {
956
+ "epoch": 9.08,
957
+ "learning_rate": 0.0001821331663535796,
958
+ "loss": 0.6965,
959
+ "step": 14500
960
+ },
961
+ {
962
+ "epoch": 9.14,
963
+ "learning_rate": 0.00018108954289292422,
964
+ "loss": 0.6896,
965
+ "step": 14600
966
+ },
967
+ {
968
+ "epoch": 9.2,
969
+ "learning_rate": 0.00018004591943226883,
970
+ "loss": 0.6944,
971
+ "step": 14700
972
+ },
973
+ {
974
+ "epoch": 9.27,
975
+ "learning_rate": 0.00017900229597161342,
976
+ "loss": 0.6981,
977
+ "step": 14800
978
+ },
979
+ {
980
+ "epoch": 9.33,
981
+ "learning_rate": 0.00017795867251095803,
982
+ "loss": 0.6819,
983
+ "step": 14900
984
+ },
985
+ {
986
+ "epoch": 9.39,
987
+ "learning_rate": 0.00017691504905030265,
988
+ "loss": 0.6869,
989
+ "step": 15000
990
+ },
991
+ {
992
+ "epoch": 9.46,
993
+ "learning_rate": 0.00017587142558964723,
994
+ "loss": 0.6974,
995
+ "step": 15100
996
+ },
997
+ {
998
+ "epoch": 9.52,
999
+ "learning_rate": 0.00017482780212899185,
1000
+ "loss": 0.6621,
1001
+ "step": 15200
1002
+ },
1003
+ {
1004
+ "epoch": 9.58,
1005
+ "learning_rate": 0.00017378417866833646,
1006
+ "loss": 0.6732,
1007
+ "step": 15300
1008
+ },
1009
+ {
1010
+ "epoch": 9.64,
1011
+ "learning_rate": 0.00017274055520768105,
1012
+ "loss": 0.678,
1013
+ "step": 15400
1014
+ },
1015
+ {
1016
+ "epoch": 9.71,
1017
+ "learning_rate": 0.00017169693174702566,
1018
+ "loss": 0.6494,
1019
+ "step": 15500
1020
+ },
1021
+ {
1022
+ "epoch": 9.77,
1023
+ "learning_rate": 0.00017065330828637025,
1024
+ "loss": 0.6909,
1025
+ "step": 15600
1026
+ },
1027
+ {
1028
+ "epoch": 9.83,
1029
+ "learning_rate": 0.00016960968482571486,
1030
+ "loss": 0.687,
1031
+ "step": 15700
1032
+ },
1033
+ {
1034
+ "epoch": 9.89,
1035
+ "learning_rate": 0.00016856606136505948,
1036
+ "loss": 0.6705,
1037
+ "step": 15800
1038
+ },
1039
+ {
1040
+ "epoch": 9.96,
1041
+ "learning_rate": 0.00016752243790440406,
1042
+ "loss": 0.6698,
1043
+ "step": 15900
1044
+ },
1045
+ {
1046
+ "epoch": 10.0,
1047
+ "eval_accuracy": 0.8949691085613416,
1048
+ "eval_loss": 0.3288457989692688,
1049
+ "eval_runtime": 150.3856,
1050
+ "eval_samples_per_second": 45.204,
1051
+ "eval_steps_per_second": 0.712,
1052
+ "step": 15970
1053
+ },
1054
+ {
1055
+ "epoch": 10.02,
1056
+ "learning_rate": 0.00016648925067835524,
1057
+ "loss": 0.6701,
1058
+ "step": 16000
1059
+ },
1060
+ {
1061
+ "epoch": 10.08,
1062
+ "learning_rate": 0.00016544562721769983,
1063
+ "loss": 0.6771,
1064
+ "step": 16100
1065
+ },
1066
+ {
1067
+ "epoch": 10.14,
1068
+ "learning_rate": 0.00016440200375704444,
1069
+ "loss": 0.6877,
1070
+ "step": 16200
1071
+ },
1072
+ {
1073
+ "epoch": 10.21,
1074
+ "learning_rate": 0.00016335838029638906,
1075
+ "loss": 0.6495,
1076
+ "step": 16300
1077
+ },
1078
+ {
1079
+ "epoch": 10.27,
1080
+ "learning_rate": 0.00016231475683573364,
1081
+ "loss": 0.6925,
1082
+ "step": 16400
1083
+ },
1084
+ {
1085
+ "epoch": 10.33,
1086
+ "learning_rate": 0.00016127113337507826,
1087
+ "loss": 0.647,
1088
+ "step": 16500
1089
+ },
1090
+ {
1091
+ "epoch": 10.39,
1092
+ "learning_rate": 0.00016022750991442287,
1093
+ "loss": 0.658,
1094
+ "step": 16600
1095
+ },
1096
+ {
1097
+ "epoch": 10.46,
1098
+ "learning_rate": 0.00015918388645376746,
1099
+ "loss": 0.7033,
1100
+ "step": 16700
1101
+ },
1102
+ {
1103
+ "epoch": 10.52,
1104
+ "learning_rate": 0.00015814026299311207,
1105
+ "loss": 0.6675,
1106
+ "step": 16800
1107
+ },
1108
+ {
1109
+ "epoch": 10.58,
1110
+ "learning_rate": 0.00015709663953245666,
1111
+ "loss": 0.6905,
1112
+ "step": 16900
1113
+ },
1114
+ {
1115
+ "epoch": 10.64,
1116
+ "learning_rate": 0.0001560530160718013,
1117
+ "loss": 0.6766,
1118
+ "step": 17000
1119
+ },
1120
+ {
1121
+ "epoch": 10.71,
1122
+ "learning_rate": 0.00015500939261114588,
1123
+ "loss": 0.684,
1124
+ "step": 17100
1125
+ },
1126
+ {
1127
+ "epoch": 10.77,
1128
+ "learning_rate": 0.00015396576915049047,
1129
+ "loss": 0.6382,
1130
+ "step": 17200
1131
+ },
1132
+ {
1133
+ "epoch": 10.83,
1134
+ "learning_rate": 0.0001529221456898351,
1135
+ "loss": 0.6737,
1136
+ "step": 17300
1137
+ },
1138
+ {
1139
+ "epoch": 10.9,
1140
+ "learning_rate": 0.0001518785222291797,
1141
+ "loss": 0.691,
1142
+ "step": 17400
1143
+ },
1144
+ {
1145
+ "epoch": 10.96,
1146
+ "learning_rate": 0.00015083489876852429,
1147
+ "loss": 0.716,
1148
+ "step": 17500
1149
+ },
1150
+ {
1151
+ "epoch": 11.0,
1152
+ "eval_accuracy": 0.8998234774933804,
1153
+ "eval_loss": 0.31469690799713135,
1154
+ "eval_runtime": 150.5759,
1155
+ "eval_samples_per_second": 45.147,
1156
+ "eval_steps_per_second": 0.711,
1157
+ "step": 17567
1158
+ },
1159
+ {
1160
+ "epoch": 11.02,
1161
+ "learning_rate": 0.0001497912753078689,
1162
+ "loss": 0.7326,
1163
+ "step": 17600
1164
+ },
1165
+ {
1166
+ "epoch": 11.08,
1167
+ "learning_rate": 0.0001487476518472135,
1168
+ "loss": 0.6747,
1169
+ "step": 17700
1170
+ },
1171
+ {
1172
+ "epoch": 11.15,
1173
+ "learning_rate": 0.00014770402838655813,
1174
+ "loss": 0.7075,
1175
+ "step": 17800
1176
+ },
1177
+ {
1178
+ "epoch": 11.21,
1179
+ "learning_rate": 0.0001466604049259027,
1180
+ "loss": 0.69,
1181
+ "step": 17900
1182
+ },
1183
+ {
1184
+ "epoch": 11.27,
1185
+ "learning_rate": 0.00014561678146524733,
1186
+ "loss": 0.6793,
1187
+ "step": 18000
1188
+ },
1189
+ {
1190
+ "epoch": 11.33,
1191
+ "learning_rate": 0.00014457315800459194,
1192
+ "loss": 0.6782,
1193
+ "step": 18100
1194
+ },
1195
+ {
1196
+ "epoch": 11.4,
1197
+ "learning_rate": 0.00014352953454393655,
1198
+ "loss": 0.6532,
1199
+ "step": 18200
1200
+ },
1201
+ {
1202
+ "epoch": 11.46,
1203
+ "learning_rate": 0.00014249634731788768,
1204
+ "loss": 0.7053,
1205
+ "step": 18300
1206
+ },
1207
+ {
1208
+ "epoch": 11.52,
1209
+ "learning_rate": 0.0001414527238572323,
1210
+ "loss": 0.6476,
1211
+ "step": 18400
1212
+ },
1213
+ {
1214
+ "epoch": 11.58,
1215
+ "learning_rate": 0.0001404091003965769,
1216
+ "loss": 0.6308,
1217
+ "step": 18500
1218
+ },
1219
+ {
1220
+ "epoch": 11.65,
1221
+ "learning_rate": 0.0001393654769359215,
1222
+ "loss": 0.6886,
1223
+ "step": 18600
1224
+ },
1225
+ {
1226
+ "epoch": 11.71,
1227
+ "learning_rate": 0.0001383218534752661,
1228
+ "loss": 0.6631,
1229
+ "step": 18700
1230
+ },
1231
+ {
1232
+ "epoch": 11.77,
1233
+ "learning_rate": 0.00013727823001461072,
1234
+ "loss": 0.7056,
1235
+ "step": 18800
1236
+ },
1237
+ {
1238
+ "epoch": 11.83,
1239
+ "learning_rate": 0.00013623460655395533,
1240
+ "loss": 0.6602,
1241
+ "step": 18900
1242
+ },
1243
+ {
1244
+ "epoch": 11.9,
1245
+ "learning_rate": 0.00013519098309329992,
1246
+ "loss": 0.6728,
1247
+ "step": 19000
1248
+ },
1249
+ {
1250
+ "epoch": 11.96,
1251
+ "learning_rate": 0.00013414735963264453,
1252
+ "loss": 0.6514,
1253
+ "step": 19100
1254
+ },
1255
+ {
1256
+ "epoch": 12.0,
1257
+ "eval_accuracy": 0.9111503383348043,
1258
+ "eval_loss": 0.3034283220767975,
1259
+ "eval_runtime": 145.7542,
1260
+ "eval_samples_per_second": 46.64,
1261
+ "eval_steps_per_second": 0.734,
1262
+ "step": 19164
1263
+ },
1264
+ {
1265
+ "epoch": 12.02,
1266
+ "learning_rate": 0.00013310373617198915,
1267
+ "loss": 0.6567,
1268
+ "step": 19200
1269
+ },
1270
+ {
1271
+ "epoch": 12.09,
1272
+ "learning_rate": 0.00013206011271133373,
1273
+ "loss": 0.6882,
1274
+ "step": 19300
1275
+ },
1276
+ {
1277
+ "epoch": 12.15,
1278
+ "learning_rate": 0.00013101648925067835,
1279
+ "loss": 0.6511,
1280
+ "step": 19400
1281
+ },
1282
+ {
1283
+ "epoch": 12.21,
1284
+ "learning_rate": 0.00012997286579002296,
1285
+ "loss": 0.6705,
1286
+ "step": 19500
1287
+ },
1288
+ {
1289
+ "epoch": 12.27,
1290
+ "learning_rate": 0.00012892924232936755,
1291
+ "loss": 0.6693,
1292
+ "step": 19600
1293
+ },
1294
+ {
1295
+ "epoch": 12.34,
1296
+ "learning_rate": 0.00012788561886871216,
1297
+ "loss": 0.68,
1298
+ "step": 19700
1299
+ },
1300
+ {
1301
+ "epoch": 12.4,
1302
+ "learning_rate": 0.00012684199540805675,
1303
+ "loss": 0.6767,
1304
+ "step": 19800
1305
+ },
1306
+ {
1307
+ "epoch": 12.46,
1308
+ "learning_rate": 0.00012579837194740136,
1309
+ "loss": 0.6768,
1310
+ "step": 19900
1311
+ },
1312
+ {
1313
+ "epoch": 12.52,
1314
+ "learning_rate": 0.00012475474848674598,
1315
+ "loss": 0.6662,
1316
+ "step": 20000
1317
+ },
1318
+ {
1319
+ "epoch": 12.59,
1320
+ "learning_rate": 0.00012371112502609056,
1321
+ "loss": 0.6511,
1322
+ "step": 20100
1323
+ },
1324
+ {
1325
+ "epoch": 12.65,
1326
+ "learning_rate": 0.00012266750156543518,
1327
+ "loss": 0.7057,
1328
+ "step": 20200
1329
+ },
1330
+ {
1331
+ "epoch": 12.71,
1332
+ "learning_rate": 0.00012163431433938634,
1333
+ "loss": 0.6699,
1334
+ "step": 20300
1335
+ },
1336
+ {
1337
+ "epoch": 12.77,
1338
+ "learning_rate": 0.00012059069087873094,
1339
+ "loss": 0.6541,
1340
+ "step": 20400
1341
+ },
1342
+ {
1343
+ "epoch": 12.84,
1344
+ "learning_rate": 0.00011954706741807554,
1345
+ "loss": 0.6741,
1346
+ "step": 20500
1347
+ },
1348
+ {
1349
+ "epoch": 12.9,
1350
+ "learning_rate": 0.00011850344395742016,
1351
+ "loss": 0.658,
1352
+ "step": 20600
1353
+ },
1354
+ {
1355
+ "epoch": 12.96,
1356
+ "learning_rate": 0.00011745982049676476,
1357
+ "loss": 0.6513,
1358
+ "step": 20700
1359
+ },
1360
+ {
1361
+ "epoch": 13.0,
1362
+ "eval_accuracy": 0.9092380111797588,
1363
+ "eval_loss": 0.30905914306640625,
1364
+ "eval_runtime": 146.3169,
1365
+ "eval_samples_per_second": 46.461,
1366
+ "eval_steps_per_second": 0.731,
1367
+ "step": 20761
1368
+ },
1369
+ {
1370
+ "epoch": 13.02,
1371
+ "learning_rate": 0.00011641619703610936,
1372
+ "loss": 0.6568,
1373
+ "step": 20800
1374
+ },
1375
+ {
1376
+ "epoch": 13.09,
1377
+ "learning_rate": 0.00011537257357545397,
1378
+ "loss": 0.6853,
1379
+ "step": 20900
1380
+ },
1381
+ {
1382
+ "epoch": 13.15,
1383
+ "learning_rate": 0.00011432895011479858,
1384
+ "loss": 0.6699,
1385
+ "step": 21000
1386
+ },
1387
+ {
1388
+ "epoch": 13.21,
1389
+ "learning_rate": 0.00011328532665414317,
1390
+ "loss": 0.6494,
1391
+ "step": 21100
1392
+ },
1393
+ {
1394
+ "epoch": 13.27,
1395
+ "learning_rate": 0.00011224170319348778,
1396
+ "loss": 0.7118,
1397
+ "step": 21200
1398
+ },
1399
+ {
1400
+ "epoch": 13.34,
1401
+ "learning_rate": 0.00011119807973283239,
1402
+ "loss": 0.6649,
1403
+ "step": 21300
1404
+ },
1405
+ {
1406
+ "epoch": 13.4,
1407
+ "learning_rate": 0.00011015445627217699,
1408
+ "loss": 0.6646,
1409
+ "step": 21400
1410
+ },
1411
+ {
1412
+ "epoch": 13.46,
1413
+ "learning_rate": 0.0001091108328115216,
1414
+ "loss": 0.6436,
1415
+ "step": 21500
1416
+ },
1417
+ {
1418
+ "epoch": 13.53,
1419
+ "learning_rate": 0.0001080672093508662,
1420
+ "loss": 0.6258,
1421
+ "step": 21600
1422
+ },
1423
+ {
1424
+ "epoch": 13.59,
1425
+ "learning_rate": 0.0001070235858902108,
1426
+ "loss": 0.6754,
1427
+ "step": 21700
1428
+ },
1429
+ {
1430
+ "epoch": 13.65,
1431
+ "learning_rate": 0.0001059799624295554,
1432
+ "loss": 0.6737,
1433
+ "step": 21800
1434
+ },
1435
+ {
1436
+ "epoch": 13.71,
1437
+ "learning_rate": 0.00010493633896890001,
1438
+ "loss": 0.6511,
1439
+ "step": 21900
1440
+ },
1441
+ {
1442
+ "epoch": 13.78,
1443
+ "learning_rate": 0.00010389271550824463,
1444
+ "loss": 0.6472,
1445
+ "step": 22000
1446
+ },
1447
+ {
1448
+ "epoch": 13.84,
1449
+ "learning_rate": 0.00010284909204758921,
1450
+ "loss": 0.6571,
1451
+ "step": 22100
1452
+ },
1453
+ {
1454
+ "epoch": 13.9,
1455
+ "learning_rate": 0.00010180546858693383,
1456
+ "loss": 0.693,
1457
+ "step": 22200
1458
+ },
1459
+ {
1460
+ "epoch": 13.96,
1461
+ "learning_rate": 0.00010076184512627843,
1462
+ "loss": 0.652,
1463
+ "step": 22300
1464
+ },
1465
+ {
1466
+ "epoch": 14.0,
1467
+ "eval_accuracy": 0.909973521624007,
1468
+ "eval_loss": 0.30560359358787537,
1469
+ "eval_runtime": 146.3619,
1470
+ "eval_samples_per_second": 46.447,
1471
+ "eval_steps_per_second": 0.731,
1472
+ "step": 22358
1473
+ },
1474
+ {
1475
+ "epoch": 14.03,
1476
+ "learning_rate": 9.971822166562303e-05,
1477
+ "loss": 0.6286,
1478
+ "step": 22400
1479
+ },
1480
+ {
1481
+ "epoch": 14.09,
1482
+ "learning_rate": 9.867459820496764e-05,
1483
+ "loss": 0.6503,
1484
+ "step": 22500
1485
+ },
1486
+ {
1487
+ "epoch": 14.15,
1488
+ "learning_rate": 9.763097474431224e-05,
1489
+ "loss": 0.6514,
1490
+ "step": 22600
1491
+ },
1492
+ {
1493
+ "epoch": 14.21,
1494
+ "learning_rate": 9.659778751826341e-05,
1495
+ "loss": 0.6728,
1496
+ "step": 22700
1497
+ },
1498
+ {
1499
+ "epoch": 14.28,
1500
+ "learning_rate": 9.555416405760801e-05,
1501
+ "loss": 0.6621,
1502
+ "step": 22800
1503
+ },
1504
+ {
1505
+ "epoch": 14.34,
1506
+ "learning_rate": 9.451054059695261e-05,
1507
+ "loss": 0.6771,
1508
+ "step": 22900
1509
+ },
1510
+ {
1511
+ "epoch": 14.4,
1512
+ "learning_rate": 9.346691713629722e-05,
1513
+ "loss": 0.6689,
1514
+ "step": 23000
1515
+ },
1516
+ {
1517
+ "epoch": 14.46,
1518
+ "learning_rate": 9.242329367564181e-05,
1519
+ "loss": 0.6712,
1520
+ "step": 23100
1521
+ },
1522
+ {
1523
+ "epoch": 14.53,
1524
+ "learning_rate": 9.137967021498642e-05,
1525
+ "loss": 0.6761,
1526
+ "step": 23200
1527
+ },
1528
+ {
1529
+ "epoch": 14.59,
1530
+ "learning_rate": 9.033604675433104e-05,
1531
+ "loss": 0.6327,
1532
+ "step": 23300
1533
+ },
1534
+ {
1535
+ "epoch": 14.65,
1536
+ "learning_rate": 8.929242329367564e-05,
1537
+ "loss": 0.6671,
1538
+ "step": 23400
1539
+ },
1540
+ {
1541
+ "epoch": 14.72,
1542
+ "learning_rate": 8.824879983302024e-05,
1543
+ "loss": 0.6598,
1544
+ "step": 23500
1545
+ },
1546
+ {
1547
+ "epoch": 14.78,
1548
+ "learning_rate": 8.720517637236485e-05,
1549
+ "loss": 0.6317,
1550
+ "step": 23600
1551
+ },
1552
+ {
1553
+ "epoch": 14.84,
1554
+ "learning_rate": 8.616155291170945e-05,
1555
+ "loss": 0.6615,
1556
+ "step": 23700
1557
+ },
1558
+ {
1559
+ "epoch": 14.9,
1560
+ "learning_rate": 8.511792945105405e-05,
1561
+ "loss": 0.6087,
1562
+ "step": 23800
1563
+ },
1564
+ {
1565
+ "epoch": 14.97,
1566
+ "learning_rate": 8.407430599039865e-05,
1567
+ "loss": 0.7105,
1568
+ "step": 23900
1569
+ },
1570
+ {
1571
+ "epoch": 15.0,
1572
+ "eval_accuracy": 0.9149749926448956,
1573
+ "eval_loss": 0.30149412155151367,
1574
+ "eval_runtime": 145.6909,
1575
+ "eval_samples_per_second": 46.66,
1576
+ "eval_steps_per_second": 0.734,
1577
+ "step": 23955
1578
+ },
1579
+ {
1580
+ "epoch": 15.03,
1581
+ "learning_rate": 8.303068252974326e-05,
1582
+ "loss": 0.6911,
1583
+ "step": 24000
1584
+ },
1585
+ {
1586
+ "epoch": 15.09,
1587
+ "learning_rate": 8.198705906908788e-05,
1588
+ "loss": 0.6717,
1589
+ "step": 24100
1590
+ },
1591
+ {
1592
+ "epoch": 15.15,
1593
+ "learning_rate": 8.094343560843246e-05,
1594
+ "loss": 0.6564,
1595
+ "step": 24200
1596
+ },
1597
+ {
1598
+ "epoch": 15.22,
1599
+ "learning_rate": 7.989981214777708e-05,
1600
+ "loss": 0.6446,
1601
+ "step": 24300
1602
+ },
1603
+ {
1604
+ "epoch": 15.28,
1605
+ "learning_rate": 7.885618868712169e-05,
1606
+ "loss": 0.6431,
1607
+ "step": 24400
1608
+ },
1609
+ {
1610
+ "epoch": 15.34,
1611
+ "learning_rate": 7.781256522646628e-05,
1612
+ "loss": 0.6762,
1613
+ "step": 24500
1614
+ },
1615
+ {
1616
+ "epoch": 15.4,
1617
+ "learning_rate": 7.676894176581089e-05,
1618
+ "loss": 0.6656,
1619
+ "step": 24600
1620
+ },
1621
+ {
1622
+ "epoch": 15.47,
1623
+ "learning_rate": 7.572531830515549e-05,
1624
+ "loss": 0.6337,
1625
+ "step": 24700
1626
+ },
1627
+ {
1628
+ "epoch": 15.53,
1629
+ "learning_rate": 7.468169484450009e-05,
1630
+ "loss": 0.6541,
1631
+ "step": 24800
1632
+ },
1633
+ {
1634
+ "epoch": 15.59,
1635
+ "learning_rate": 7.363807138384469e-05,
1636
+ "loss": 0.6772,
1637
+ "step": 24900
1638
+ },
1639
+ {
1640
+ "epoch": 15.65,
1641
+ "learning_rate": 7.260488415779586e-05,
1642
+ "loss": 0.629,
1643
+ "step": 25000
1644
+ },
1645
+ {
1646
+ "epoch": 15.72,
1647
+ "learning_rate": 7.156126069714046e-05,
1648
+ "loss": 0.6998,
1649
+ "step": 25100
1650
+ },
1651
+ {
1652
+ "epoch": 15.78,
1653
+ "learning_rate": 7.051763723648507e-05,
1654
+ "loss": 0.6686,
1655
+ "step": 25200
1656
+ },
1657
+ {
1658
+ "epoch": 15.84,
1659
+ "learning_rate": 6.947401377582967e-05,
1660
+ "loss": 0.6822,
1661
+ "step": 25300
1662
+ },
1663
+ {
1664
+ "epoch": 15.9,
1665
+ "learning_rate": 6.843039031517429e-05,
1666
+ "loss": 0.6143,
1667
+ "step": 25400
1668
+ },
1669
+ {
1670
+ "epoch": 15.97,
1671
+ "learning_rate": 6.738676685451889e-05,
1672
+ "loss": 0.6337,
1673
+ "step": 25500
1674
+ },
1675
+ {
1676
+ "epoch": 16.0,
1677
+ "eval_accuracy": 0.9090909090909091,
1678
+ "eval_loss": 0.30700910091400146,
1679
+ "eval_runtime": 146.384,
1680
+ "eval_samples_per_second": 46.44,
1681
+ "eval_steps_per_second": 0.731,
1682
+ "step": 25552
1683
+ },
1684
+ {
1685
+ "epoch": 16.03,
1686
+ "learning_rate": 6.634314339386349e-05,
1687
+ "loss": 0.6502,
1688
+ "step": 25600
1689
+ },
1690
+ {
1691
+ "epoch": 16.09,
1692
+ "learning_rate": 6.52995199332081e-05,
1693
+ "loss": 0.6448,
1694
+ "step": 25700
1695
+ },
1696
+ {
1697
+ "epoch": 16.16,
1698
+ "learning_rate": 6.42558964725527e-05,
1699
+ "loss": 0.6272,
1700
+ "step": 25800
1701
+ },
1702
+ {
1703
+ "epoch": 16.22,
1704
+ "learning_rate": 6.32122730118973e-05,
1705
+ "loss": 0.6316,
1706
+ "step": 25900
1707
+ },
1708
+ {
1709
+ "epoch": 16.28,
1710
+ "learning_rate": 6.21686495512419e-05,
1711
+ "loss": 0.6442,
1712
+ "step": 26000
1713
+ },
1714
+ {
1715
+ "epoch": 16.34,
1716
+ "learning_rate": 6.11250260905865e-05,
1717
+ "loss": 0.6489,
1718
+ "step": 26100
1719
+ },
1720
+ {
1721
+ "epoch": 16.41,
1722
+ "learning_rate": 6.0081402629931114e-05,
1723
+ "loss": 0.6603,
1724
+ "step": 26200
1725
+ },
1726
+ {
1727
+ "epoch": 16.47,
1728
+ "learning_rate": 5.9037779169275714e-05,
1729
+ "loss": 0.6449,
1730
+ "step": 26300
1731
+ },
1732
+ {
1733
+ "epoch": 16.53,
1734
+ "learning_rate": 5.799415570862033e-05,
1735
+ "loss": 0.661,
1736
+ "step": 26400
1737
+ },
1738
+ {
1739
+ "epoch": 16.59,
1740
+ "learning_rate": 5.695053224796493e-05,
1741
+ "loss": 0.642,
1742
+ "step": 26500
1743
+ },
1744
+ {
1745
+ "epoch": 16.66,
1746
+ "learning_rate": 5.5906908787309535e-05,
1747
+ "loss": 0.6498,
1748
+ "step": 26600
1749
+ },
1750
+ {
1751
+ "epoch": 16.72,
1752
+ "learning_rate": 5.4863285326654135e-05,
1753
+ "loss": 0.6902,
1754
+ "step": 26700
1755
+ },
1756
+ {
1757
+ "epoch": 16.78,
1758
+ "learning_rate": 5.381966186599874e-05,
1759
+ "loss": 0.6336,
1760
+ "step": 26800
1761
+ },
1762
+ {
1763
+ "epoch": 16.84,
1764
+ "learning_rate": 5.277603840534335e-05,
1765
+ "loss": 0.6393,
1766
+ "step": 26900
1767
+ },
1768
+ {
1769
+ "epoch": 16.91,
1770
+ "learning_rate": 5.173241494468795e-05,
1771
+ "loss": 0.6496,
1772
+ "step": 27000
1773
+ },
1774
+ {
1775
+ "epoch": 16.97,
1776
+ "learning_rate": 5.0688791484032557e-05,
1777
+ "loss": 0.63,
1778
+ "step": 27100
1779
+ },
1780
+ {
1781
+ "epoch": 17.0,
1782
+ "eval_accuracy": 0.913503971756399,
1783
+ "eval_loss": 0.30175167322158813,
1784
+ "eval_runtime": 149.8031,
1785
+ "eval_samples_per_second": 45.38,
1786
+ "eval_steps_per_second": 0.714,
1787
+ "step": 27149
1788
+ },
1789
+ {
1790
+ "epoch": 17.03,
1791
+ "learning_rate": 4.9645168023377163e-05,
1792
+ "loss": 0.6195,
1793
+ "step": 27200
1794
+ },
1795
+ {
1796
+ "epoch": 17.09,
1797
+ "learning_rate": 4.861198079732832e-05,
1798
+ "loss": 0.6299,
1799
+ "step": 27300
1800
+ },
1801
+ {
1802
+ "epoch": 17.16,
1803
+ "learning_rate": 4.756835733667292e-05,
1804
+ "loss": 0.6939,
1805
+ "step": 27400
1806
+ },
1807
+ {
1808
+ "epoch": 17.22,
1809
+ "learning_rate": 4.652473387601752e-05,
1810
+ "loss": 0.6416,
1811
+ "step": 27500
1812
+ },
1813
+ {
1814
+ "epoch": 17.28,
1815
+ "learning_rate": 4.5481110415362136e-05,
1816
+ "loss": 0.6626,
1817
+ "step": 27600
1818
+ },
1819
+ {
1820
+ "epoch": 17.35,
1821
+ "learning_rate": 4.4437486954706736e-05,
1822
+ "loss": 0.6444,
1823
+ "step": 27700
1824
+ },
1825
+ {
1826
+ "epoch": 17.41,
1827
+ "learning_rate": 4.339386349405134e-05,
1828
+ "loss": 0.6238,
1829
+ "step": 27800
1830
+ },
1831
+ {
1832
+ "epoch": 17.47,
1833
+ "learning_rate": 4.2350240033395944e-05,
1834
+ "loss": 0.6505,
1835
+ "step": 27900
1836
+ },
1837
+ {
1838
+ "epoch": 17.53,
1839
+ "learning_rate": 4.130661657274056e-05,
1840
+ "loss": 0.6346,
1841
+ "step": 28000
1842
+ },
1843
+ {
1844
+ "epoch": 17.6,
1845
+ "learning_rate": 4.026299311208516e-05,
1846
+ "loss": 0.6319,
1847
+ "step": 28100
1848
+ },
1849
+ {
1850
+ "epoch": 17.66,
1851
+ "learning_rate": 3.921936965142976e-05,
1852
+ "loss": 0.6811,
1853
+ "step": 28200
1854
+ },
1855
+ {
1856
+ "epoch": 17.72,
1857
+ "learning_rate": 3.8175746190774365e-05,
1858
+ "loss": 0.6645,
1859
+ "step": 28300
1860
+ },
1861
+ {
1862
+ "epoch": 17.78,
1863
+ "learning_rate": 3.713212273011897e-05,
1864
+ "loss": 0.6512,
1865
+ "step": 28400
1866
+ },
1867
+ {
1868
+ "epoch": 17.85,
1869
+ "learning_rate": 3.608849926946357e-05,
1870
+ "loss": 0.6578,
1871
+ "step": 28500
1872
+ },
1873
+ {
1874
+ "epoch": 17.91,
1875
+ "learning_rate": 3.504487580880818e-05,
1876
+ "loss": 0.6358,
1877
+ "step": 28600
1878
+ },
1879
+ {
1880
+ "epoch": 17.97,
1881
+ "learning_rate": 3.4001252348152786e-05,
1882
+ "loss": 0.6672,
1883
+ "step": 28700
1884
+ },
1885
+ {
1886
+ "epoch": 18.0,
1887
+ "eval_accuracy": 0.9087967049132097,
1888
+ "eval_loss": 0.30836355686187744,
1889
+ "eval_runtime": 148.2197,
1890
+ "eval_samples_per_second": 45.864,
1891
+ "eval_steps_per_second": 0.722,
1892
+ "step": 28746
1893
+ },
1894
+ {
1895
+ "epoch": 18.03,
1896
+ "learning_rate": 3.295762888749739e-05,
1897
+ "loss": 0.65,
1898
+ "step": 28800
1899
+ },
1900
+ {
1901
+ "epoch": 18.1,
1902
+ "learning_rate": 3.191400542684199e-05,
1903
+ "loss": 0.6293,
1904
+ "step": 28900
1905
+ },
1906
+ {
1907
+ "epoch": 18.16,
1908
+ "learning_rate": 3.087038196618659e-05,
1909
+ "loss": 0.6059,
1910
+ "step": 29000
1911
+ },
1912
+ {
1913
+ "epoch": 18.22,
1914
+ "learning_rate": 2.98267585055312e-05,
1915
+ "loss": 0.5975,
1916
+ "step": 29100
1917
+ },
1918
+ {
1919
+ "epoch": 18.28,
1920
+ "learning_rate": 2.8783135044875807e-05,
1921
+ "loss": 0.6759,
1922
+ "step": 29200
1923
+ },
1924
+ {
1925
+ "epoch": 18.35,
1926
+ "learning_rate": 2.773951158422041e-05,
1927
+ "loss": 0.6457,
1928
+ "step": 29300
1929
+ },
1930
+ {
1931
+ "epoch": 18.41,
1932
+ "learning_rate": 2.6695888123565017e-05,
1933
+ "loss": 0.6715,
1934
+ "step": 29400
1935
+ },
1936
+ {
1937
+ "epoch": 18.47,
1938
+ "learning_rate": 2.565226466290962e-05,
1939
+ "loss": 0.6642,
1940
+ "step": 29500
1941
+ },
1942
+ {
1943
+ "epoch": 18.53,
1944
+ "learning_rate": 2.460864120225422e-05,
1945
+ "loss": 0.6432,
1946
+ "step": 29600
1947
+ },
1948
+ {
1949
+ "epoch": 18.6,
1950
+ "learning_rate": 2.3565017741598828e-05,
1951
+ "loss": 0.6441,
1952
+ "step": 29700
1953
+ },
1954
+ {
1955
+ "epoch": 18.66,
1956
+ "learning_rate": 2.2521394280943432e-05,
1957
+ "loss": 0.6843,
1958
+ "step": 29800
1959
+ },
1960
+ {
1961
+ "epoch": 18.72,
1962
+ "learning_rate": 2.147777082028804e-05,
1963
+ "loss": 0.6459,
1964
+ "step": 29900
1965
+ },
1966
+ {
1967
+ "epoch": 18.79,
1968
+ "learning_rate": 2.0434147359632642e-05,
1969
+ "loss": 0.6233,
1970
+ "step": 30000
1971
+ },
1972
+ {
1973
+ "epoch": 18.85,
1974
+ "learning_rate": 1.939052389897725e-05,
1975
+ "loss": 0.6634,
1976
+ "step": 30100
1977
+ },
1978
+ {
1979
+ "epoch": 18.91,
1980
+ "learning_rate": 1.8346900438321853e-05,
1981
+ "loss": 0.6701,
1982
+ "step": 30200
1983
+ },
1984
+ {
1985
+ "epoch": 18.97,
1986
+ "learning_rate": 1.7303276977666456e-05,
1987
+ "loss": 0.6479,
1988
+ "step": 30300
1989
+ },
1990
+ {
1991
+ "epoch": 19.0,
1992
+ "eval_accuracy": 0.9101206237128567,
1993
+ "eval_loss": 0.3060016632080078,
1994
+ "eval_runtime": 149.555,
1995
+ "eval_samples_per_second": 45.455,
1996
+ "eval_steps_per_second": 0.715,
1997
+ "step": 30343
1998
+ },
1999
+ {
2000
+ "epoch": 19.04,
2001
+ "learning_rate": 1.625965351701106e-05,
2002
+ "loss": 0.6155,
2003
+ "step": 30400
2004
+ },
2005
+ {
2006
+ "epoch": 19.1,
2007
+ "learning_rate": 1.5216030056355665e-05,
2008
+ "loss": 0.6079,
2009
+ "step": 30500
2010
+ },
2011
+ {
2012
+ "epoch": 19.16,
2013
+ "learning_rate": 1.417240659570027e-05,
2014
+ "loss": 0.6709,
2015
+ "step": 30600
2016
+ },
2017
+ {
2018
+ "epoch": 19.22,
2019
+ "learning_rate": 1.3128783135044874e-05,
2020
+ "loss": 0.6604,
2021
+ "step": 30700
2022
+ },
2023
+ {
2024
+ "epoch": 19.29,
2025
+ "learning_rate": 1.208515967438948e-05,
2026
+ "loss": 0.6222,
2027
+ "step": 30800
2028
+ },
2029
+ {
2030
+ "epoch": 19.35,
2031
+ "learning_rate": 1.1041536213734085e-05,
2032
+ "loss": 0.6428,
2033
+ "step": 30900
2034
+ },
2035
+ {
2036
+ "epoch": 19.41,
2037
+ "learning_rate": 9.997912753078688e-06,
2038
+ "loss": 0.6664,
2039
+ "step": 31000
2040
+ },
2041
+ {
2042
+ "epoch": 19.47,
2043
+ "learning_rate": 8.954289292423293e-06,
2044
+ "loss": 0.6489,
2045
+ "step": 31100
2046
+ },
2047
+ {
2048
+ "epoch": 19.54,
2049
+ "learning_rate": 7.910665831767897e-06,
2050
+ "loss": 0.6414,
2051
+ "step": 31200
2052
+ },
2053
+ {
2054
+ "epoch": 19.6,
2055
+ "learning_rate": 6.877478605719056e-06,
2056
+ "loss": 0.6599,
2057
+ "step": 31300
2058
+ },
2059
+ {
2060
+ "epoch": 19.66,
2061
+ "learning_rate": 5.83385514506366e-06,
2062
+ "loss": 0.6433,
2063
+ "step": 31400
2064
+ },
2065
+ {
2066
+ "epoch": 19.72,
2067
+ "learning_rate": 4.790231684408265e-06,
2068
+ "loss": 0.6325,
2069
+ "step": 31500
2070
+ },
2071
+ {
2072
+ "epoch": 19.79,
2073
+ "learning_rate": 3.7466082237528697e-06,
2074
+ "loss": 0.6363,
2075
+ "step": 31600
2076
+ },
2077
+ {
2078
+ "epoch": 19.85,
2079
+ "learning_rate": 2.7029847630974745e-06,
2080
+ "loss": 0.6408,
2081
+ "step": 31700
2082
+ },
2083
+ {
2084
+ "epoch": 19.91,
2085
+ "learning_rate": 1.6593613024420787e-06,
2086
+ "loss": 0.645,
2087
+ "step": 31800
2088
+ },
2089
+ {
2090
+ "epoch": 19.97,
2091
+ "learning_rate": 6.157378417866834e-07,
2092
+ "loss": 0.6658,
2093
+ "step": 31900
2094
+ },
2095
+ {
2096
+ "epoch": 20.0,
2097
+ "eval_accuracy": 0.9089438070020595,
2098
+ "eval_loss": 0.3071773946285248,
2099
+ "eval_runtime": 150.0563,
2100
+ "eval_samples_per_second": 45.303,
2101
+ "eval_steps_per_second": 0.713,
2102
+ "step": 31940
2103
+ },
2104
+ {
2105
+ "epoch": 20.0,
2106
+ "step": 31940,
2107
+ "total_flos": 0.0,
2108
+ "train_loss": 0.8136612295581911,
2109
+ "train_runtime": 27726.7705,
2110
+ "train_samples_per_second": 36.855,
2111
+ "train_steps_per_second": 1.152
2112
+ }
2113
+ ],
2114
+ "max_steps": 31940,
2115
+ "num_train_epochs": 20,
2116
+ "total_flos": 0.0,
2117
+ "trial_name": null,
2118
+ "trial_params": null
2119
+ }