patrickvonplaten commited on
Commit
8d7a280
1 Parent(s): 439fe3a

End of training

Browse files
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "eval_loss": 0.3457382619380951,
4
+ "eval_runtime": 56.2561,
5
+ "eval_samples": 1680,
6
+ "eval_samples_per_second": 29.863,
7
+ "eval_steps_per_second": 29.863,
8
+ "eval_wer": 0.2151471297636276,
9
+ "train_loss": 0.6826766253185683,
10
+ "train_runtime": 4678.6424,
11
+ "train_samples": 4620,
12
+ "train_samples_per_second": 19.749,
13
+ "train_steps_per_second": 0.62
14
+ }
emissions.csv ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ timestamp,experiment_id,project_name,duration,emissions,energy_consumed,country_name,country_iso_code,region,on_cloud,cloud_provider,cloud_region
2
+ 2021-10-21T11:59:48,a10f5d0a-cd7c-4333-9a08-916890d3cbcf,codecarbon,4675.637314081192,0.011588395740298725,0.055058056549283564,United States,USA,new york,N,,
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "eval_loss": 0.3457382619380951,
4
+ "eval_runtime": 56.2561,
5
+ "eval_samples": 1680,
6
+ "eval_samples_per_second": 29.863,
7
+ "eval_steps_per_second": 29.863,
8
+ "eval_wer": 0.2151471297636276
9
+ }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:511aacd483c3b1c22cc94acec615d536cceeef51463d79de36f6a995e0c18dfb
3
  size 377672428
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94f3b45f5dbc1dbd5dcde66958569ffe62b3a3ee93c161c55969d8b7a8f93126
3
  size 377672428
runs/Oct21_10-40-40_brutasse/events.out.tfevents.1634812907.brutasse.28529.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:688579feed1102d320477ff4da239439d06a9fc00d7580f5d8c813af374031f1
3
- size 57323
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69467adf2c67965abcf85286094ccab3c6768b75ffba762e21d20d11b210e52b
3
+ size 59565
runs/Oct21_10-40-40_brutasse/events.out.tfevents.1634817648.brutasse.28529.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1c2664b4369fad1f72be0ec4654355c28cef28326aed448c4846fac34235dfb
3
+ size 358
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "train_loss": 0.6826766253185683,
4
+ "train_runtime": 4678.6424,
5
+ "train_samples": 4620,
6
+ "train_samples_per_second": 19.749,
7
+ "train_steps_per_second": 0.62
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,2026 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 20.0,
5
+ "global_step": 2900,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.07,
12
+ "learning_rate": 9e-07,
13
+ "loss": 9.7696,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.14,
18
+ "learning_rate": 1.9e-06,
19
+ "loss": 8.5301,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.21,
24
+ "learning_rate": 2.9e-06,
25
+ "loss": 7.8804,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.28,
30
+ "learning_rate": 3.8e-06,
31
+ "loss": 8.3082,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.34,
36
+ "learning_rate": 4.800000000000001e-06,
37
+ "loss": 7.4843,
38
+ "step": 50
39
+ },
40
+ {
41
+ "epoch": 0.41,
42
+ "learning_rate": 5.8e-06,
43
+ "loss": 6.4623,
44
+ "step": 60
45
+ },
46
+ {
47
+ "epoch": 0.48,
48
+ "learning_rate": 6.800000000000001e-06,
49
+ "loss": 5.0935,
50
+ "step": 70
51
+ },
52
+ {
53
+ "epoch": 0.55,
54
+ "learning_rate": 7.8e-06,
55
+ "loss": 4.1987,
56
+ "step": 80
57
+ },
58
+ {
59
+ "epoch": 0.62,
60
+ "learning_rate": 8.8e-06,
61
+ "loss": 3.3439,
62
+ "step": 90
63
+ },
64
+ {
65
+ "epoch": 0.69,
66
+ "learning_rate": 9.800000000000001e-06,
67
+ "loss": 3.1621,
68
+ "step": 100
69
+ },
70
+ {
71
+ "epoch": 0.69,
72
+ "eval_loss": 3.1101927757263184,
73
+ "eval_runtime": 55.8678,
74
+ "eval_samples_per_second": 30.071,
75
+ "eval_steps_per_second": 30.071,
76
+ "eval_wer": 1.0,
77
+ "step": 100
78
+ },
79
+ {
80
+ "epoch": 0.76,
81
+ "learning_rate": 1.08e-05,
82
+ "loss": 3.1218,
83
+ "step": 110
84
+ },
85
+ {
86
+ "epoch": 0.83,
87
+ "learning_rate": 1.18e-05,
88
+ "loss": 3.0538,
89
+ "step": 120
90
+ },
91
+ {
92
+ "epoch": 0.9,
93
+ "learning_rate": 1.2800000000000001e-05,
94
+ "loss": 3.0235,
95
+ "step": 130
96
+ },
97
+ {
98
+ "epoch": 0.97,
99
+ "learning_rate": 1.3800000000000002e-05,
100
+ "loss": 3.0046,
101
+ "step": 140
102
+ },
103
+ {
104
+ "epoch": 1.03,
105
+ "learning_rate": 1.48e-05,
106
+ "loss": 3.0439,
107
+ "step": 150
108
+ },
109
+ {
110
+ "epoch": 1.1,
111
+ "learning_rate": 1.58e-05,
112
+ "loss": 3.0008,
113
+ "step": 160
114
+ },
115
+ {
116
+ "epoch": 1.17,
117
+ "learning_rate": 1.6800000000000002e-05,
118
+ "loss": 2.9794,
119
+ "step": 170
120
+ },
121
+ {
122
+ "epoch": 1.24,
123
+ "learning_rate": 1.78e-05,
124
+ "loss": 2.9939,
125
+ "step": 180
126
+ },
127
+ {
128
+ "epoch": 1.31,
129
+ "learning_rate": 1.88e-05,
130
+ "loss": 2.9932,
131
+ "step": 190
132
+ },
133
+ {
134
+ "epoch": 1.38,
135
+ "learning_rate": 1.9800000000000004e-05,
136
+ "loss": 2.9592,
137
+ "step": 200
138
+ },
139
+ {
140
+ "epoch": 1.38,
141
+ "eval_loss": 2.960341215133667,
142
+ "eval_runtime": 55.5906,
143
+ "eval_samples_per_second": 30.221,
144
+ "eval_steps_per_second": 30.221,
145
+ "eval_wer": 1.0,
146
+ "step": 200
147
+ },
148
+ {
149
+ "epoch": 1.45,
150
+ "learning_rate": 2.08e-05,
151
+ "loss": 2.9513,
152
+ "step": 210
153
+ },
154
+ {
155
+ "epoch": 1.52,
156
+ "learning_rate": 2.18e-05,
157
+ "loss": 2.9676,
158
+ "step": 220
159
+ },
160
+ {
161
+ "epoch": 1.59,
162
+ "learning_rate": 2.2800000000000002e-05,
163
+ "loss": 2.9471,
164
+ "step": 230
165
+ },
166
+ {
167
+ "epoch": 1.66,
168
+ "learning_rate": 2.38e-05,
169
+ "loss": 2.9278,
170
+ "step": 240
171
+ },
172
+ {
173
+ "epoch": 1.72,
174
+ "learning_rate": 2.48e-05,
175
+ "loss": 2.92,
176
+ "step": 250
177
+ },
178
+ {
179
+ "epoch": 1.79,
180
+ "learning_rate": 2.58e-05,
181
+ "loss": 2.9487,
182
+ "step": 260
183
+ },
184
+ {
185
+ "epoch": 1.86,
186
+ "learning_rate": 2.6800000000000004e-05,
187
+ "loss": 2.8995,
188
+ "step": 270
189
+ },
190
+ {
191
+ "epoch": 1.93,
192
+ "learning_rate": 2.7800000000000005e-05,
193
+ "loss": 2.9073,
194
+ "step": 280
195
+ },
196
+ {
197
+ "epoch": 2.0,
198
+ "learning_rate": 2.88e-05,
199
+ "loss": 2.9357,
200
+ "step": 290
201
+ },
202
+ {
203
+ "epoch": 2.07,
204
+ "learning_rate": 2.98e-05,
205
+ "loss": 2.9116,
206
+ "step": 300
207
+ },
208
+ {
209
+ "epoch": 2.07,
210
+ "eval_loss": 2.8920886516571045,
211
+ "eval_runtime": 55.9266,
212
+ "eval_samples_per_second": 30.039,
213
+ "eval_steps_per_second": 30.039,
214
+ "eval_wer": 1.0,
215
+ "step": 300
216
+ },
217
+ {
218
+ "epoch": 2.14,
219
+ "learning_rate": 3.08e-05,
220
+ "loss": 2.8585,
221
+ "step": 310
222
+ },
223
+ {
224
+ "epoch": 2.21,
225
+ "learning_rate": 3.18e-05,
226
+ "loss": 2.8498,
227
+ "step": 320
228
+ },
229
+ {
230
+ "epoch": 2.28,
231
+ "learning_rate": 3.2800000000000004e-05,
232
+ "loss": 2.8704,
233
+ "step": 330
234
+ },
235
+ {
236
+ "epoch": 2.34,
237
+ "learning_rate": 3.38e-05,
238
+ "loss": 2.7755,
239
+ "step": 340
240
+ },
241
+ {
242
+ "epoch": 2.41,
243
+ "learning_rate": 3.48e-05,
244
+ "loss": 2.6958,
245
+ "step": 350
246
+ },
247
+ {
248
+ "epoch": 2.48,
249
+ "learning_rate": 3.58e-05,
250
+ "loss": 2.6205,
251
+ "step": 360
252
+ },
253
+ {
254
+ "epoch": 2.55,
255
+ "learning_rate": 3.68e-05,
256
+ "loss": 2.5635,
257
+ "step": 370
258
+ },
259
+ {
260
+ "epoch": 2.62,
261
+ "learning_rate": 3.7800000000000004e-05,
262
+ "loss": 2.3608,
263
+ "step": 380
264
+ },
265
+ {
266
+ "epoch": 2.69,
267
+ "learning_rate": 3.88e-05,
268
+ "loss": 2.1984,
269
+ "step": 390
270
+ },
271
+ {
272
+ "epoch": 2.76,
273
+ "learning_rate": 3.9800000000000005e-05,
274
+ "loss": 2.1332,
275
+ "step": 400
276
+ },
277
+ {
278
+ "epoch": 2.76,
279
+ "eval_loss": 1.971829891204834,
280
+ "eval_runtime": 55.7409,
281
+ "eval_samples_per_second": 30.139,
282
+ "eval_steps_per_second": 30.139,
283
+ "eval_wer": 0.995796292467783,
284
+ "step": 400
285
+ },
286
+ {
287
+ "epoch": 2.83,
288
+ "learning_rate": 4.08e-05,
289
+ "loss": 1.9219,
290
+ "step": 410
291
+ },
292
+ {
293
+ "epoch": 2.9,
294
+ "learning_rate": 4.18e-05,
295
+ "loss": 1.7055,
296
+ "step": 420
297
+ },
298
+ {
299
+ "epoch": 2.97,
300
+ "learning_rate": 4.2800000000000004e-05,
301
+ "loss": 1.5857,
302
+ "step": 430
303
+ },
304
+ {
305
+ "epoch": 3.03,
306
+ "learning_rate": 4.38e-05,
307
+ "loss": 1.6118,
308
+ "step": 440
309
+ },
310
+ {
311
+ "epoch": 3.1,
312
+ "learning_rate": 4.4800000000000005e-05,
313
+ "loss": 1.2915,
314
+ "step": 450
315
+ },
316
+ {
317
+ "epoch": 3.17,
318
+ "learning_rate": 4.58e-05,
319
+ "loss": 1.1766,
320
+ "step": 460
321
+ },
322
+ {
323
+ "epoch": 3.24,
324
+ "learning_rate": 4.6800000000000006e-05,
325
+ "loss": 1.1746,
326
+ "step": 470
327
+ },
328
+ {
329
+ "epoch": 3.31,
330
+ "learning_rate": 4.78e-05,
331
+ "loss": 1.1026,
332
+ "step": 480
333
+ },
334
+ {
335
+ "epoch": 3.38,
336
+ "learning_rate": 4.88e-05,
337
+ "loss": 0.8672,
338
+ "step": 490
339
+ },
340
+ {
341
+ "epoch": 3.45,
342
+ "learning_rate": 4.9800000000000004e-05,
343
+ "loss": 0.8477,
344
+ "step": 500
345
+ },
346
+ {
347
+ "epoch": 3.45,
348
+ "eval_loss": 0.7813268303871155,
349
+ "eval_runtime": 56.0095,
350
+ "eval_samples_per_second": 29.995,
351
+ "eval_steps_per_second": 29.995,
352
+ "eval_wer": 0.5237406105712907,
353
+ "step": 500
354
+ },
355
+ {
356
+ "epoch": 3.52,
357
+ "learning_rate": 5.08e-05,
358
+ "loss": 0.9733,
359
+ "step": 510
360
+ },
361
+ {
362
+ "epoch": 3.59,
363
+ "learning_rate": 5.1800000000000005e-05,
364
+ "loss": 0.6984,
365
+ "step": 520
366
+ },
367
+ {
368
+ "epoch": 3.66,
369
+ "learning_rate": 5.28e-05,
370
+ "loss": 0.65,
371
+ "step": 530
372
+ },
373
+ {
374
+ "epoch": 3.72,
375
+ "learning_rate": 5.380000000000001e-05,
376
+ "loss": 0.7546,
377
+ "step": 540
378
+ },
379
+ {
380
+ "epoch": 3.79,
381
+ "learning_rate": 5.4800000000000004e-05,
382
+ "loss": 0.7855,
383
+ "step": 550
384
+ },
385
+ {
386
+ "epoch": 3.86,
387
+ "learning_rate": 5.580000000000001e-05,
388
+ "loss": 0.5968,
389
+ "step": 560
390
+ },
391
+ {
392
+ "epoch": 3.93,
393
+ "learning_rate": 5.68e-05,
394
+ "loss": 0.6043,
395
+ "step": 570
396
+ },
397
+ {
398
+ "epoch": 4.0,
399
+ "learning_rate": 5.7799999999999995e-05,
400
+ "loss": 0.6865,
401
+ "step": 580
402
+ },
403
+ {
404
+ "epoch": 4.07,
405
+ "learning_rate": 5.88e-05,
406
+ "loss": 0.567,
407
+ "step": 590
408
+ },
409
+ {
410
+ "epoch": 4.14,
411
+ "learning_rate": 5.9800000000000003e-05,
412
+ "loss": 0.4251,
413
+ "step": 600
414
+ },
415
+ {
416
+ "epoch": 4.14,
417
+ "eval_loss": 0.5166197419166565,
418
+ "eval_runtime": 55.6677,
419
+ "eval_samples_per_second": 30.179,
420
+ "eval_steps_per_second": 30.179,
421
+ "eval_wer": 0.39824960374888013,
422
+ "step": 600
423
+ },
424
+ {
425
+ "epoch": 4.21,
426
+ "learning_rate": 6.08e-05,
427
+ "loss": 0.4704,
428
+ "step": 610
429
+ },
430
+ {
431
+ "epoch": 4.28,
432
+ "learning_rate": 6.18e-05,
433
+ "loss": 0.5981,
434
+ "step": 620
435
+ },
436
+ {
437
+ "epoch": 4.34,
438
+ "learning_rate": 6.280000000000001e-05,
439
+ "loss": 0.4192,
440
+ "step": 630
441
+ },
442
+ {
443
+ "epoch": 4.41,
444
+ "learning_rate": 6.38e-05,
445
+ "loss": 0.4157,
446
+ "step": 640
447
+ },
448
+ {
449
+ "epoch": 4.48,
450
+ "learning_rate": 6.48e-05,
451
+ "loss": 0.4709,
452
+ "step": 650
453
+ },
454
+ {
455
+ "epoch": 4.55,
456
+ "learning_rate": 6.58e-05,
457
+ "loss": 0.506,
458
+ "step": 660
459
+ },
460
+ {
461
+ "epoch": 4.62,
462
+ "learning_rate": 6.680000000000001e-05,
463
+ "loss": 0.3589,
464
+ "step": 670
465
+ },
466
+ {
467
+ "epoch": 4.69,
468
+ "learning_rate": 6.780000000000001e-05,
469
+ "loss": 0.3787,
470
+ "step": 680
471
+ },
472
+ {
473
+ "epoch": 4.76,
474
+ "learning_rate": 6.879999999999999e-05,
475
+ "loss": 0.5322,
476
+ "step": 690
477
+ },
478
+ {
479
+ "epoch": 4.83,
480
+ "learning_rate": 6.98e-05,
481
+ "loss": 0.3743,
482
+ "step": 700
483
+ },
484
+ {
485
+ "epoch": 4.83,
486
+ "eval_loss": 0.4399681091308594,
487
+ "eval_runtime": 56.2233,
488
+ "eval_samples_per_second": 29.881,
489
+ "eval_steps_per_second": 29.881,
490
+ "eval_wer": 0.35779753290607125,
491
+ "step": 700
492
+ },
493
+ {
494
+ "epoch": 4.9,
495
+ "learning_rate": 7.08e-05,
496
+ "loss": 0.3068,
497
+ "step": 710
498
+ },
499
+ {
500
+ "epoch": 4.97,
501
+ "learning_rate": 7.18e-05,
502
+ "loss": 0.3595,
503
+ "step": 720
504
+ },
505
+ {
506
+ "epoch": 5.03,
507
+ "learning_rate": 7.280000000000001e-05,
508
+ "loss": 0.4887,
509
+ "step": 730
510
+ },
511
+ {
512
+ "epoch": 5.1,
513
+ "learning_rate": 7.38e-05,
514
+ "loss": 0.2713,
515
+ "step": 740
516
+ },
517
+ {
518
+ "epoch": 5.17,
519
+ "learning_rate": 7.48e-05,
520
+ "loss": 0.2848,
521
+ "step": 750
522
+ },
523
+ {
524
+ "epoch": 5.24,
525
+ "learning_rate": 7.58e-05,
526
+ "loss": 0.3825,
527
+ "step": 760
528
+ },
529
+ {
530
+ "epoch": 5.31,
531
+ "learning_rate": 7.680000000000001e-05,
532
+ "loss": 0.3228,
533
+ "step": 770
534
+ },
535
+ {
536
+ "epoch": 5.38,
537
+ "learning_rate": 7.780000000000001e-05,
538
+ "loss": 0.2446,
539
+ "step": 780
540
+ },
541
+ {
542
+ "epoch": 5.45,
543
+ "learning_rate": 7.88e-05,
544
+ "loss": 0.2864,
545
+ "step": 790
546
+ },
547
+ {
548
+ "epoch": 5.52,
549
+ "learning_rate": 7.98e-05,
550
+ "loss": 0.4194,
551
+ "step": 800
552
+ },
553
+ {
554
+ "epoch": 5.52,
555
+ "eval_loss": 0.40769341588020325,
556
+ "eval_runtime": 56.215,
557
+ "eval_samples_per_second": 29.885,
558
+ "eval_steps_per_second": 29.885,
559
+ "eval_wer": 0.33698573495968576,
560
+ "step": 800
561
+ },
562
+ {
563
+ "epoch": 5.59,
564
+ "learning_rate": 8.080000000000001e-05,
565
+ "loss": 0.2523,
566
+ "step": 810
567
+ },
568
+ {
569
+ "epoch": 5.66,
570
+ "learning_rate": 8.18e-05,
571
+ "loss": 0.2337,
572
+ "step": 820
573
+ },
574
+ {
575
+ "epoch": 5.72,
576
+ "learning_rate": 8.28e-05,
577
+ "loss": 0.3113,
578
+ "step": 830
579
+ },
580
+ {
581
+ "epoch": 5.79,
582
+ "learning_rate": 8.38e-05,
583
+ "loss": 0.373,
584
+ "step": 840
585
+ },
586
+ {
587
+ "epoch": 5.86,
588
+ "learning_rate": 8.48e-05,
589
+ "loss": 0.2086,
590
+ "step": 850
591
+ },
592
+ {
593
+ "epoch": 5.93,
594
+ "learning_rate": 8.58e-05,
595
+ "loss": 0.2483,
596
+ "step": 860
597
+ },
598
+ {
599
+ "epoch": 6.0,
600
+ "learning_rate": 8.680000000000001e-05,
601
+ "loss": 0.3528,
602
+ "step": 870
603
+ },
604
+ {
605
+ "epoch": 6.07,
606
+ "learning_rate": 8.78e-05,
607
+ "loss": 0.225,
608
+ "step": 880
609
+ },
610
+ {
611
+ "epoch": 6.14,
612
+ "learning_rate": 8.88e-05,
613
+ "loss": 0.1819,
614
+ "step": 890
615
+ },
616
+ {
617
+ "epoch": 6.21,
618
+ "learning_rate": 8.98e-05,
619
+ "loss": 0.23,
620
+ "step": 900
621
+ },
622
+ {
623
+ "epoch": 6.21,
624
+ "eval_loss": 0.40175914764404297,
625
+ "eval_runtime": 56.0809,
626
+ "eval_samples_per_second": 29.957,
627
+ "eval_steps_per_second": 29.957,
628
+ "eval_wer": 0.3142443663427745,
629
+ "step": 900
630
+ },
631
+ {
632
+ "epoch": 6.28,
633
+ "learning_rate": 9.080000000000001e-05,
634
+ "loss": 0.3483,
635
+ "step": 910
636
+ },
637
+ {
638
+ "epoch": 6.34,
639
+ "learning_rate": 9.180000000000001e-05,
640
+ "loss": 0.1811,
641
+ "step": 920
642
+ },
643
+ {
644
+ "epoch": 6.41,
645
+ "learning_rate": 9.28e-05,
646
+ "loss": 0.1956,
647
+ "step": 930
648
+ },
649
+ {
650
+ "epoch": 6.48,
651
+ "learning_rate": 9.38e-05,
652
+ "loss": 0.2701,
653
+ "step": 940
654
+ },
655
+ {
656
+ "epoch": 6.55,
657
+ "learning_rate": 9.48e-05,
658
+ "loss": 0.2568,
659
+ "step": 950
660
+ },
661
+ {
662
+ "epoch": 6.62,
663
+ "learning_rate": 9.58e-05,
664
+ "loss": 0.1576,
665
+ "step": 960
666
+ },
667
+ {
668
+ "epoch": 6.69,
669
+ "learning_rate": 9.680000000000001e-05,
670
+ "loss": 0.2026,
671
+ "step": 970
672
+ },
673
+ {
674
+ "epoch": 6.76,
675
+ "learning_rate": 9.78e-05,
676
+ "loss": 0.354,
677
+ "step": 980
678
+ },
679
+ {
680
+ "epoch": 6.83,
681
+ "learning_rate": 9.88e-05,
682
+ "loss": 0.2064,
683
+ "step": 990
684
+ },
685
+ {
686
+ "epoch": 6.9,
687
+ "learning_rate": 9.98e-05,
688
+ "loss": 0.1554,
689
+ "step": 1000
690
+ },
691
+ {
692
+ "epoch": 6.9,
693
+ "eval_loss": 0.362291544675827,
694
+ "eval_runtime": 56.0865,
695
+ "eval_samples_per_second": 29.954,
696
+ "eval_steps_per_second": 29.954,
697
+ "eval_wer": 0.2994969333608986,
698
+ "step": 1000
699
+ },
700
+ {
701
+ "epoch": 6.97,
702
+ "learning_rate": 9.957894736842106e-05,
703
+ "loss": 0.2273,
704
+ "step": 1010
705
+ },
706
+ {
707
+ "epoch": 7.03,
708
+ "learning_rate": 9.905263157894737e-05,
709
+ "loss": 0.2759,
710
+ "step": 1020
711
+ },
712
+ {
713
+ "epoch": 7.1,
714
+ "learning_rate": 9.852631578947368e-05,
715
+ "loss": 0.1415,
716
+ "step": 1030
717
+ },
718
+ {
719
+ "epoch": 7.17,
720
+ "learning_rate": 9.8e-05,
721
+ "loss": 0.1495,
722
+ "step": 1040
723
+ },
724
+ {
725
+ "epoch": 7.24,
726
+ "learning_rate": 9.747368421052632e-05,
727
+ "loss": 0.2374,
728
+ "step": 1050
729
+ },
730
+ {
731
+ "epoch": 7.31,
732
+ "learning_rate": 9.694736842105264e-05,
733
+ "loss": 0.205,
734
+ "step": 1060
735
+ },
736
+ {
737
+ "epoch": 7.38,
738
+ "learning_rate": 9.642105263157896e-05,
739
+ "loss": 0.1435,
740
+ "step": 1070
741
+ },
742
+ {
743
+ "epoch": 7.45,
744
+ "learning_rate": 9.589473684210528e-05,
745
+ "loss": 0.1814,
746
+ "step": 1080
747
+ },
748
+ {
749
+ "epoch": 7.52,
750
+ "learning_rate": 9.536842105263159e-05,
751
+ "loss": 0.2679,
752
+ "step": 1090
753
+ },
754
+ {
755
+ "epoch": 7.59,
756
+ "learning_rate": 9.48421052631579e-05,
757
+ "loss": 0.1511,
758
+ "step": 1100
759
+ },
760
+ {
761
+ "epoch": 7.59,
762
+ "eval_loss": 0.343284547328949,
763
+ "eval_runtime": 56.3106,
764
+ "eval_samples_per_second": 29.835,
765
+ "eval_steps_per_second": 29.835,
766
+ "eval_wer": 0.26972641444421475,
767
+ "step": 1100
768
+ },
769
+ {
770
+ "epoch": 7.66,
771
+ "learning_rate": 9.431578947368421e-05,
772
+ "loss": 0.1261,
773
+ "step": 1110
774
+ },
775
+ {
776
+ "epoch": 7.72,
777
+ "learning_rate": 9.378947368421053e-05,
778
+ "loss": 0.1857,
779
+ "step": 1120
780
+ },
781
+ {
782
+ "epoch": 7.79,
783
+ "learning_rate": 9.326315789473684e-05,
784
+ "loss": 0.241,
785
+ "step": 1130
786
+ },
787
+ {
788
+ "epoch": 7.86,
789
+ "learning_rate": 9.273684210526317e-05,
790
+ "loss": 0.1263,
791
+ "step": 1140
792
+ },
793
+ {
794
+ "epoch": 7.93,
795
+ "learning_rate": 9.221052631578948e-05,
796
+ "loss": 0.1407,
797
+ "step": 1150
798
+ },
799
+ {
800
+ "epoch": 8.0,
801
+ "learning_rate": 9.16842105263158e-05,
802
+ "loss": 0.2252,
803
+ "step": 1160
804
+ },
805
+ {
806
+ "epoch": 8.07,
807
+ "learning_rate": 9.11578947368421e-05,
808
+ "loss": 0.1381,
809
+ "step": 1170
810
+ },
811
+ {
812
+ "epoch": 8.14,
813
+ "learning_rate": 9.063157894736842e-05,
814
+ "loss": 0.1075,
815
+ "step": 1180
816
+ },
817
+ {
818
+ "epoch": 8.21,
819
+ "learning_rate": 9.010526315789473e-05,
820
+ "loss": 0.1296,
821
+ "step": 1190
822
+ },
823
+ {
824
+ "epoch": 8.28,
825
+ "learning_rate": 8.957894736842106e-05,
826
+ "loss": 0.1983,
827
+ "step": 1200
828
+ },
829
+ {
830
+ "epoch": 8.28,
831
+ "eval_loss": 0.3538705110549927,
832
+ "eval_runtime": 56.1491,
833
+ "eval_samples_per_second": 29.92,
834
+ "eval_steps_per_second": 29.92,
835
+ "eval_wer": 0.2715181586382744,
836
+ "step": 1200
837
+ },
838
+ {
839
+ "epoch": 8.34,
840
+ "learning_rate": 8.905263157894737e-05,
841
+ "loss": 0.1206,
842
+ "step": 1210
843
+ },
844
+ {
845
+ "epoch": 8.41,
846
+ "learning_rate": 8.852631578947368e-05,
847
+ "loss": 0.1165,
848
+ "step": 1220
849
+ },
850
+ {
851
+ "epoch": 8.48,
852
+ "learning_rate": 8.800000000000001e-05,
853
+ "loss": 0.1739,
854
+ "step": 1230
855
+ },
856
+ {
857
+ "epoch": 8.55,
858
+ "learning_rate": 8.747368421052632e-05,
859
+ "loss": 0.1857,
860
+ "step": 1240
861
+ },
862
+ {
863
+ "epoch": 8.62,
864
+ "learning_rate": 8.694736842105264e-05,
865
+ "loss": 0.0872,
866
+ "step": 1250
867
+ },
868
+ {
869
+ "epoch": 8.69,
870
+ "learning_rate": 8.642105263157895e-05,
871
+ "loss": 0.1306,
872
+ "step": 1260
873
+ },
874
+ {
875
+ "epoch": 8.76,
876
+ "learning_rate": 8.589473684210528e-05,
877
+ "loss": 0.2083,
878
+ "step": 1270
879
+ },
880
+ {
881
+ "epoch": 8.83,
882
+ "learning_rate": 8.536842105263159e-05,
883
+ "loss": 0.1227,
884
+ "step": 1280
885
+ },
886
+ {
887
+ "epoch": 8.9,
888
+ "learning_rate": 8.48421052631579e-05,
889
+ "loss": 0.0972,
890
+ "step": 1290
891
+ },
892
+ {
893
+ "epoch": 8.97,
894
+ "learning_rate": 8.431578947368422e-05,
895
+ "loss": 0.1443,
896
+ "step": 1300
897
+ },
898
+ {
899
+ "epoch": 8.97,
900
+ "eval_loss": 0.3622087836265564,
901
+ "eval_runtime": 55.9308,
902
+ "eval_samples_per_second": 30.037,
903
+ "eval_steps_per_second": 30.037,
904
+ "eval_wer": 0.25511680793880503,
905
+ "step": 1300
906
+ },
907
+ {
908
+ "epoch": 9.03,
909
+ "learning_rate": 8.378947368421053e-05,
910
+ "loss": 0.1891,
911
+ "step": 1310
912
+ },
913
+ {
914
+ "epoch": 9.1,
915
+ "learning_rate": 8.326315789473684e-05,
916
+ "loss": 0.0701,
917
+ "step": 1320
918
+ },
919
+ {
920
+ "epoch": 9.17,
921
+ "learning_rate": 8.273684210526315e-05,
922
+ "loss": 0.0991,
923
+ "step": 1330
924
+ },
925
+ {
926
+ "epoch": 9.24,
927
+ "learning_rate": 8.221052631578948e-05,
928
+ "loss": 0.1548,
929
+ "step": 1340
930
+ },
931
+ {
932
+ "epoch": 9.31,
933
+ "learning_rate": 8.16842105263158e-05,
934
+ "loss": 0.132,
935
+ "step": 1350
936
+ },
937
+ {
938
+ "epoch": 9.38,
939
+ "learning_rate": 8.115789473684211e-05,
940
+ "loss": 0.0846,
941
+ "step": 1360
942
+ },
943
+ {
944
+ "epoch": 9.45,
945
+ "learning_rate": 8.063157894736842e-05,
946
+ "loss": 0.1054,
947
+ "step": 1370
948
+ },
949
+ {
950
+ "epoch": 9.52,
951
+ "learning_rate": 8.010526315789473e-05,
952
+ "loss": 0.1609,
953
+ "step": 1380
954
+ },
955
+ {
956
+ "epoch": 9.59,
957
+ "learning_rate": 7.957894736842106e-05,
958
+ "loss": 0.0877,
959
+ "step": 1390
960
+ },
961
+ {
962
+ "epoch": 9.66,
963
+ "learning_rate": 7.905263157894737e-05,
964
+ "loss": 0.0971,
965
+ "step": 1400
966
+ },
967
+ {
968
+ "epoch": 9.66,
969
+ "eval_loss": 0.35799121856689453,
970
+ "eval_runtime": 55.9651,
971
+ "eval_samples_per_second": 30.019,
972
+ "eval_steps_per_second": 30.019,
973
+ "eval_wer": 0.251877885741851,
974
+ "step": 1400
975
+ },
976
+ {
977
+ "epoch": 9.72,
978
+ "learning_rate": 7.852631578947369e-05,
979
+ "loss": 0.1346,
980
+ "step": 1410
981
+ },
982
+ {
983
+ "epoch": 9.79,
984
+ "learning_rate": 7.800000000000001e-05,
985
+ "loss": 0.1553,
986
+ "step": 1420
987
+ },
988
+ {
989
+ "epoch": 9.86,
990
+ "learning_rate": 7.747368421052633e-05,
991
+ "loss": 0.0879,
992
+ "step": 1430
993
+ },
994
+ {
995
+ "epoch": 9.93,
996
+ "learning_rate": 7.694736842105264e-05,
997
+ "loss": 0.0909,
998
+ "step": 1440
999
+ },
1000
+ {
1001
+ "epoch": 10.0,
1002
+ "learning_rate": 7.642105263157895e-05,
1003
+ "loss": 0.1714,
1004
+ "step": 1450
1005
+ },
1006
+ {
1007
+ "epoch": 10.07,
1008
+ "learning_rate": 7.589473684210526e-05,
1009
+ "loss": 0.0766,
1010
+ "step": 1460
1011
+ },
1012
+ {
1013
+ "epoch": 10.14,
1014
+ "learning_rate": 7.536842105263158e-05,
1015
+ "loss": 0.0693,
1016
+ "step": 1470
1017
+ },
1018
+ {
1019
+ "epoch": 10.21,
1020
+ "learning_rate": 7.48421052631579e-05,
1021
+ "loss": 0.0947,
1022
+ "step": 1480
1023
+ },
1024
+ {
1025
+ "epoch": 10.28,
1026
+ "learning_rate": 7.431578947368422e-05,
1027
+ "loss": 0.1642,
1028
+ "step": 1490
1029
+ },
1030
+ {
1031
+ "epoch": 10.34,
1032
+ "learning_rate": 7.378947368421053e-05,
1033
+ "loss": 0.0764,
1034
+ "step": 1500
1035
+ },
1036
+ {
1037
+ "epoch": 10.34,
1038
+ "eval_loss": 0.35290759801864624,
1039
+ "eval_runtime": 56.302,
1040
+ "eval_samples_per_second": 29.839,
1041
+ "eval_steps_per_second": 29.839,
1042
+ "eval_wer": 0.2437461236303494,
1043
+ "step": 1500
1044
+ },
1045
+ {
1046
+ "epoch": 10.41,
1047
+ "learning_rate": 7.326315789473684e-05,
1048
+ "loss": 0.0819,
1049
+ "step": 1510
1050
+ },
1051
+ {
1052
+ "epoch": 10.48,
1053
+ "learning_rate": 7.273684210526316e-05,
1054
+ "loss": 0.1125,
1055
+ "step": 1520
1056
+ },
1057
+ {
1058
+ "epoch": 10.55,
1059
+ "learning_rate": 7.221052631578947e-05,
1060
+ "loss": 0.1252,
1061
+ "step": 1530
1062
+ },
1063
+ {
1064
+ "epoch": 10.62,
1065
+ "learning_rate": 7.16842105263158e-05,
1066
+ "loss": 0.0672,
1067
+ "step": 1540
1068
+ },
1069
+ {
1070
+ "epoch": 10.69,
1071
+ "learning_rate": 7.115789473684211e-05,
1072
+ "loss": 0.0797,
1073
+ "step": 1550
1074
+ },
1075
+ {
1076
+ "epoch": 10.76,
1077
+ "learning_rate": 7.063157894736842e-05,
1078
+ "loss": 0.1408,
1079
+ "step": 1560
1080
+ },
1081
+ {
1082
+ "epoch": 10.83,
1083
+ "learning_rate": 7.010526315789473e-05,
1084
+ "loss": 0.0758,
1085
+ "step": 1570
1086
+ },
1087
+ {
1088
+ "epoch": 10.9,
1089
+ "learning_rate": 6.957894736842106e-05,
1090
+ "loss": 0.069,
1091
+ "step": 1580
1092
+ },
1093
+ {
1094
+ "epoch": 10.97,
1095
+ "learning_rate": 6.905263157894737e-05,
1096
+ "loss": 0.0953,
1097
+ "step": 1590
1098
+ },
1099
+ {
1100
+ "epoch": 11.03,
1101
+ "learning_rate": 6.852631578947369e-05,
1102
+ "loss": 0.1203,
1103
+ "step": 1600
1104
+ },
1105
+ {
1106
+ "epoch": 11.03,
1107
+ "eval_loss": 0.34554988145828247,
1108
+ "eval_runtime": 56.3041,
1109
+ "eval_samples_per_second": 29.838,
1110
+ "eval_steps_per_second": 29.838,
1111
+ "eval_wer": 0.24305699124801874,
1112
+ "step": 1600
1113
+ },
1114
+ {
1115
+ "epoch": 11.1,
1116
+ "learning_rate": 6.800000000000001e-05,
1117
+ "loss": 0.0573,
1118
+ "step": 1610
1119
+ },
1120
+ {
1121
+ "epoch": 11.17,
1122
+ "learning_rate": 6.747368421052633e-05,
1123
+ "loss": 0.0592,
1124
+ "step": 1620
1125
+ },
1126
+ {
1127
+ "epoch": 11.24,
1128
+ "learning_rate": 6.694736842105264e-05,
1129
+ "loss": 0.1146,
1130
+ "step": 1630
1131
+ },
1132
+ {
1133
+ "epoch": 11.31,
1134
+ "learning_rate": 6.642105263157895e-05,
1135
+ "loss": 0.0912,
1136
+ "step": 1640
1137
+ },
1138
+ {
1139
+ "epoch": 11.38,
1140
+ "learning_rate": 6.589473684210526e-05,
1141
+ "loss": 0.0698,
1142
+ "step": 1650
1143
+ },
1144
+ {
1145
+ "epoch": 11.45,
1146
+ "learning_rate": 6.536842105263158e-05,
1147
+ "loss": 0.0845,
1148
+ "step": 1660
1149
+ },
1150
+ {
1151
+ "epoch": 11.52,
1152
+ "learning_rate": 6.484210526315789e-05,
1153
+ "loss": 0.1377,
1154
+ "step": 1670
1155
+ },
1156
+ {
1157
+ "epoch": 11.59,
1158
+ "learning_rate": 6.431578947368422e-05,
1159
+ "loss": 0.0615,
1160
+ "step": 1680
1161
+ },
1162
+ {
1163
+ "epoch": 11.66,
1164
+ "learning_rate": 6.378947368421053e-05,
1165
+ "loss": 0.0577,
1166
+ "step": 1690
1167
+ },
1168
+ {
1169
+ "epoch": 11.72,
1170
+ "learning_rate": 6.326315789473684e-05,
1171
+ "loss": 0.0881,
1172
+ "step": 1700
1173
+ },
1174
+ {
1175
+ "epoch": 11.72,
1176
+ "eval_loss": 0.3647613823413849,
1177
+ "eval_runtime": 56.4856,
1178
+ "eval_samples_per_second": 29.742,
1179
+ "eval_steps_per_second": 29.742,
1180
+ "eval_wer": 0.24147198676865825,
1181
+ "step": 1700
1182
+ },
1183
+ {
1184
+ "epoch": 11.79,
1185
+ "learning_rate": 6.273684210526316e-05,
1186
+ "loss": 0.1276,
1187
+ "step": 1710
1188
+ },
1189
+ {
1190
+ "epoch": 11.86,
1191
+ "learning_rate": 6.221052631578947e-05,
1192
+ "loss": 0.0503,
1193
+ "step": 1720
1194
+ },
1195
+ {
1196
+ "epoch": 11.93,
1197
+ "learning_rate": 6.168421052631578e-05,
1198
+ "loss": 0.0747,
1199
+ "step": 1730
1200
+ },
1201
+ {
1202
+ "epoch": 12.0,
1203
+ "learning_rate": 6.115789473684211e-05,
1204
+ "loss": 0.1326,
1205
+ "step": 1740
1206
+ },
1207
+ {
1208
+ "epoch": 12.07,
1209
+ "learning_rate": 6.063157894736843e-05,
1210
+ "loss": 0.0642,
1211
+ "step": 1750
1212
+ },
1213
+ {
1214
+ "epoch": 12.14,
1215
+ "learning_rate": 6.010526315789474e-05,
1216
+ "loss": 0.0548,
1217
+ "step": 1760
1218
+ },
1219
+ {
1220
+ "epoch": 12.21,
1221
+ "learning_rate": 5.9578947368421055e-05,
1222
+ "loss": 0.0628,
1223
+ "step": 1770
1224
+ },
1225
+ {
1226
+ "epoch": 12.28,
1227
+ "learning_rate": 5.905263157894737e-05,
1228
+ "loss": 0.1033,
1229
+ "step": 1780
1230
+ },
1231
+ {
1232
+ "epoch": 12.34,
1233
+ "learning_rate": 5.852631578947369e-05,
1234
+ "loss": 0.06,
1235
+ "step": 1790
1236
+ },
1237
+ {
1238
+ "epoch": 12.41,
1239
+ "learning_rate": 5.8e-05,
1240
+ "loss": 0.0521,
1241
+ "step": 1800
1242
+ },
1243
+ {
1244
+ "epoch": 12.41,
1245
+ "eval_loss": 0.35644999146461487,
1246
+ "eval_runtime": 56.2186,
1247
+ "eval_samples_per_second": 29.883,
1248
+ "eval_steps_per_second": 29.883,
1249
+ "eval_wer": 0.23203087313072843,
1250
+ "step": 1800
1251
+ },
1252
+ {
1253
+ "epoch": 12.48,
1254
+ "learning_rate": 5.747368421052633e-05,
1255
+ "loss": 0.1038,
1256
+ "step": 1810
1257
+ },
1258
+ {
1259
+ "epoch": 12.55,
1260
+ "learning_rate": 5.694736842105264e-05,
1261
+ "loss": 0.0943,
1262
+ "step": 1820
1263
+ },
1264
+ {
1265
+ "epoch": 12.62,
1266
+ "learning_rate": 5.642105263157895e-05,
1267
+ "loss": 0.0463,
1268
+ "step": 1830
1269
+ },
1270
+ {
1271
+ "epoch": 12.69,
1272
+ "learning_rate": 5.5894736842105266e-05,
1273
+ "loss": 0.0514,
1274
+ "step": 1840
1275
+ },
1276
+ {
1277
+ "epoch": 12.76,
1278
+ "learning_rate": 5.536842105263158e-05,
1279
+ "loss": 0.1006,
1280
+ "step": 1850
1281
+ },
1282
+ {
1283
+ "epoch": 12.83,
1284
+ "learning_rate": 5.484210526315789e-05,
1285
+ "loss": 0.0524,
1286
+ "step": 1860
1287
+ },
1288
+ {
1289
+ "epoch": 12.9,
1290
+ "learning_rate": 5.431578947368421e-05,
1291
+ "loss": 0.0533,
1292
+ "step": 1870
1293
+ },
1294
+ {
1295
+ "epoch": 12.97,
1296
+ "learning_rate": 5.378947368421053e-05,
1297
+ "loss": 0.0683,
1298
+ "step": 1880
1299
+ },
1300
+ {
1301
+ "epoch": 13.03,
1302
+ "learning_rate": 5.326315789473685e-05,
1303
+ "loss": 0.0983,
1304
+ "step": 1890
1305
+ },
1306
+ {
1307
+ "epoch": 13.1,
1308
+ "learning_rate": 5.2736842105263164e-05,
1309
+ "loss": 0.0434,
1310
+ "step": 1900
1311
+ },
1312
+ {
1313
+ "epoch": 13.1,
1314
+ "eval_loss": 0.34853529930114746,
1315
+ "eval_runtime": 55.7329,
1316
+ "eval_samples_per_second": 30.144,
1317
+ "eval_steps_per_second": 30.144,
1318
+ "eval_wer": 0.2270002067397147,
1319
+ "step": 1900
1320
+ },
1321
+ {
1322
+ "epoch": 13.17,
1323
+ "learning_rate": 5.221052631578948e-05,
1324
+ "loss": 0.0451,
1325
+ "step": 1910
1326
+ },
1327
+ {
1328
+ "epoch": 13.24,
1329
+ "learning_rate": 5.168421052631579e-05,
1330
+ "loss": 0.0706,
1331
+ "step": 1920
1332
+ },
1333
+ {
1334
+ "epoch": 13.31,
1335
+ "learning_rate": 5.11578947368421e-05,
1336
+ "loss": 0.0723,
1337
+ "step": 1930
1338
+ },
1339
+ {
1340
+ "epoch": 13.38,
1341
+ "learning_rate": 5.0631578947368416e-05,
1342
+ "loss": 0.0433,
1343
+ "step": 1940
1344
+ },
1345
+ {
1346
+ "epoch": 13.45,
1347
+ "learning_rate": 5.010526315789474e-05,
1348
+ "loss": 0.0722,
1349
+ "step": 1950
1350
+ },
1351
+ {
1352
+ "epoch": 13.52,
1353
+ "learning_rate": 4.9578947368421055e-05,
1354
+ "loss": 0.0909,
1355
+ "step": 1960
1356
+ },
1357
+ {
1358
+ "epoch": 13.59,
1359
+ "learning_rate": 4.9052631578947375e-05,
1360
+ "loss": 0.056,
1361
+ "step": 1970
1362
+ },
1363
+ {
1364
+ "epoch": 13.66,
1365
+ "learning_rate": 4.852631578947369e-05,
1366
+ "loss": 0.0521,
1367
+ "step": 1980
1368
+ },
1369
+ {
1370
+ "epoch": 13.72,
1371
+ "learning_rate": 4.8e-05,
1372
+ "loss": 0.0793,
1373
+ "step": 1990
1374
+ },
1375
+ {
1376
+ "epoch": 13.79,
1377
+ "learning_rate": 4.747368421052632e-05,
1378
+ "loss": 0.0864,
1379
+ "step": 2000
1380
+ },
1381
+ {
1382
+ "epoch": 13.79,
1383
+ "eval_loss": 0.3516710102558136,
1384
+ "eval_runtime": 56.0246,
1385
+ "eval_samples_per_second": 29.987,
1386
+ "eval_steps_per_second": 29.987,
1387
+ "eval_wer": 0.22279649920749775,
1388
+ "step": 2000
1389
+ },
1390
+ {
1391
+ "epoch": 13.86,
1392
+ "learning_rate": 4.6947368421052634e-05,
1393
+ "loss": 0.0422,
1394
+ "step": 2010
1395
+ },
1396
+ {
1397
+ "epoch": 13.93,
1398
+ "learning_rate": 4.642105263157895e-05,
1399
+ "loss": 0.0489,
1400
+ "step": 2020
1401
+ },
1402
+ {
1403
+ "epoch": 14.0,
1404
+ "learning_rate": 4.589473684210526e-05,
1405
+ "loss": 0.0975,
1406
+ "step": 2030
1407
+ },
1408
+ {
1409
+ "epoch": 14.07,
1410
+ "learning_rate": 4.536842105263158e-05,
1411
+ "loss": 0.0459,
1412
+ "step": 2040
1413
+ },
1414
+ {
1415
+ "epoch": 14.14,
1416
+ "learning_rate": 4.48421052631579e-05,
1417
+ "loss": 0.05,
1418
+ "step": 2050
1419
+ },
1420
+ {
1421
+ "epoch": 14.21,
1422
+ "learning_rate": 4.431578947368421e-05,
1423
+ "loss": 0.057,
1424
+ "step": 2060
1425
+ },
1426
+ {
1427
+ "epoch": 14.28,
1428
+ "learning_rate": 4.378947368421053e-05,
1429
+ "loss": 0.0887,
1430
+ "step": 2070
1431
+ },
1432
+ {
1433
+ "epoch": 14.34,
1434
+ "learning_rate": 4.3263157894736845e-05,
1435
+ "loss": 0.033,
1436
+ "step": 2080
1437
+ },
1438
+ {
1439
+ "epoch": 14.41,
1440
+ "learning_rate": 4.273684210526316e-05,
1441
+ "loss": 0.0533,
1442
+ "step": 2090
1443
+ },
1444
+ {
1445
+ "epoch": 14.48,
1446
+ "learning_rate": 4.221052631578948e-05,
1447
+ "loss": 0.0651,
1448
+ "step": 2100
1449
+ },
1450
+ {
1451
+ "epoch": 14.48,
1452
+ "eval_loss": 0.35055288672447205,
1453
+ "eval_runtime": 56.3016,
1454
+ "eval_samples_per_second": 29.839,
1455
+ "eval_steps_per_second": 29.839,
1456
+ "eval_wer": 0.2285162979808421,
1457
+ "step": 2100
1458
+ },
1459
+ {
1460
+ "epoch": 14.55,
1461
+ "learning_rate": 4.168421052631579e-05,
1462
+ "loss": 0.0905,
1463
+ "step": 2110
1464
+ },
1465
+ {
1466
+ "epoch": 14.62,
1467
+ "learning_rate": 4.1157894736842104e-05,
1468
+ "loss": 0.0387,
1469
+ "step": 2120
1470
+ },
1471
+ {
1472
+ "epoch": 14.69,
1473
+ "learning_rate": 4.0631578947368424e-05,
1474
+ "loss": 0.0607,
1475
+ "step": 2130
1476
+ },
1477
+ {
1478
+ "epoch": 14.76,
1479
+ "learning_rate": 4.010526315789474e-05,
1480
+ "loss": 0.0808,
1481
+ "step": 2140
1482
+ },
1483
+ {
1484
+ "epoch": 14.83,
1485
+ "learning_rate": 3.9578947368421056e-05,
1486
+ "loss": 0.0396,
1487
+ "step": 2150
1488
+ },
1489
+ {
1490
+ "epoch": 14.9,
1491
+ "learning_rate": 3.905263157894737e-05,
1492
+ "loss": 0.036,
1493
+ "step": 2160
1494
+ },
1495
+ {
1496
+ "epoch": 14.97,
1497
+ "learning_rate": 3.852631578947369e-05,
1498
+ "loss": 0.0439,
1499
+ "step": 2170
1500
+ },
1501
+ {
1502
+ "epoch": 15.03,
1503
+ "learning_rate": 3.8e-05,
1504
+ "loss": 0.0885,
1505
+ "step": 2180
1506
+ },
1507
+ {
1508
+ "epoch": 15.1,
1509
+ "learning_rate": 3.7473684210526315e-05,
1510
+ "loss": 0.0313,
1511
+ "step": 2190
1512
+ },
1513
+ {
1514
+ "epoch": 15.17,
1515
+ "learning_rate": 3.6947368421052635e-05,
1516
+ "loss": 0.0423,
1517
+ "step": 2200
1518
+ },
1519
+ {
1520
+ "epoch": 15.17,
1521
+ "eval_loss": 0.34284868836402893,
1522
+ "eval_runtime": 56.494,
1523
+ "eval_samples_per_second": 29.738,
1524
+ "eval_steps_per_second": 29.738,
1525
+ "eval_wer": 0.2246571566397905,
1526
+ "step": 2200
1527
+ },
1528
+ {
1529
+ "epoch": 15.24,
1530
+ "learning_rate": 3.642105263157895e-05,
1531
+ "loss": 0.0724,
1532
+ "step": 2210
1533
+ },
1534
+ {
1535
+ "epoch": 15.31,
1536
+ "learning_rate": 3.589473684210527e-05,
1537
+ "loss": 0.0547,
1538
+ "step": 2220
1539
+ },
1540
+ {
1541
+ "epoch": 15.38,
1542
+ "learning_rate": 3.536842105263158e-05,
1543
+ "loss": 0.0358,
1544
+ "step": 2230
1545
+ },
1546
+ {
1547
+ "epoch": 15.45,
1548
+ "learning_rate": 3.48421052631579e-05,
1549
+ "loss": 0.0557,
1550
+ "step": 2240
1551
+ },
1552
+ {
1553
+ "epoch": 15.52,
1554
+ "learning_rate": 3.431578947368421e-05,
1555
+ "loss": 0.085,
1556
+ "step": 2250
1557
+ },
1558
+ {
1559
+ "epoch": 15.59,
1560
+ "learning_rate": 3.3789473684210526e-05,
1561
+ "loss": 0.0429,
1562
+ "step": 2260
1563
+ },
1564
+ {
1565
+ "epoch": 15.66,
1566
+ "learning_rate": 3.3263157894736846e-05,
1567
+ "loss": 0.0397,
1568
+ "step": 2270
1569
+ },
1570
+ {
1571
+ "epoch": 15.72,
1572
+ "learning_rate": 3.273684210526316e-05,
1573
+ "loss": 0.0613,
1574
+ "step": 2280
1575
+ },
1576
+ {
1577
+ "epoch": 15.79,
1578
+ "learning_rate": 3.221052631578947e-05,
1579
+ "loss": 0.0634,
1580
+ "step": 2290
1581
+ },
1582
+ {
1583
+ "epoch": 15.86,
1584
+ "learning_rate": 3.168421052631579e-05,
1585
+ "loss": 0.0302,
1586
+ "step": 2300
1587
+ },
1588
+ {
1589
+ "epoch": 15.86,
1590
+ "eval_loss": 0.3372061848640442,
1591
+ "eval_runtime": 56.0069,
1592
+ "eval_samples_per_second": 29.996,
1593
+ "eval_steps_per_second": 29.996,
1594
+ "eval_wer": 0.21983322996347598,
1595
+ "step": 2300
1596
+ },
1597
+ {
1598
+ "epoch": 15.93,
1599
+ "learning_rate": 3.115789473684211e-05,
1600
+ "loss": 0.0465,
1601
+ "step": 2310
1602
+ },
1603
+ {
1604
+ "epoch": 16.0,
1605
+ "learning_rate": 3.0631578947368425e-05,
1606
+ "loss": 0.0689,
1607
+ "step": 2320
1608
+ },
1609
+ {
1610
+ "epoch": 16.07,
1611
+ "learning_rate": 3.0105263157894737e-05,
1612
+ "loss": 0.042,
1613
+ "step": 2330
1614
+ },
1615
+ {
1616
+ "epoch": 16.14,
1617
+ "learning_rate": 2.9578947368421057e-05,
1618
+ "loss": 0.0352,
1619
+ "step": 2340
1620
+ },
1621
+ {
1622
+ "epoch": 16.21,
1623
+ "learning_rate": 2.905263157894737e-05,
1624
+ "loss": 0.0379,
1625
+ "step": 2350
1626
+ },
1627
+ {
1628
+ "epoch": 16.28,
1629
+ "learning_rate": 2.8526315789473683e-05,
1630
+ "loss": 0.0766,
1631
+ "step": 2360
1632
+ },
1633
+ {
1634
+ "epoch": 16.34,
1635
+ "learning_rate": 2.8000000000000003e-05,
1636
+ "loss": 0.0382,
1637
+ "step": 2370
1638
+ },
1639
+ {
1640
+ "epoch": 16.41,
1641
+ "learning_rate": 2.747368421052632e-05,
1642
+ "loss": 0.0417,
1643
+ "step": 2380
1644
+ },
1645
+ {
1646
+ "epoch": 16.48,
1647
+ "learning_rate": 2.6947368421052632e-05,
1648
+ "loss": 0.0459,
1649
+ "step": 2390
1650
+ },
1651
+ {
1652
+ "epoch": 16.55,
1653
+ "learning_rate": 2.6421052631578945e-05,
1654
+ "loss": 0.0548,
1655
+ "step": 2400
1656
+ },
1657
+ {
1658
+ "epoch": 16.55,
1659
+ "eval_loss": 0.34956642985343933,
1660
+ "eval_runtime": 56.0254,
1661
+ "eval_samples_per_second": 29.986,
1662
+ "eval_steps_per_second": 29.986,
1663
+ "eval_wer": 0.21962649024877678,
1664
+ "step": 2400
1665
+ },
1666
+ {
1667
+ "epoch": 16.62,
1668
+ "learning_rate": 2.5894736842105265e-05,
1669
+ "loss": 0.0317,
1670
+ "step": 2410
1671
+ },
1672
+ {
1673
+ "epoch": 16.69,
1674
+ "learning_rate": 2.536842105263158e-05,
1675
+ "loss": 0.0362,
1676
+ "step": 2420
1677
+ },
1678
+ {
1679
+ "epoch": 16.76,
1680
+ "learning_rate": 2.4842105263157898e-05,
1681
+ "loss": 0.064,
1682
+ "step": 2430
1683
+ },
1684
+ {
1685
+ "epoch": 16.83,
1686
+ "learning_rate": 2.431578947368421e-05,
1687
+ "loss": 0.0436,
1688
+ "step": 2440
1689
+ },
1690
+ {
1691
+ "epoch": 16.9,
1692
+ "learning_rate": 2.3789473684210527e-05,
1693
+ "loss": 0.0367,
1694
+ "step": 2450
1695
+ },
1696
+ {
1697
+ "epoch": 16.97,
1698
+ "learning_rate": 2.3263157894736844e-05,
1699
+ "loss": 0.0443,
1700
+ "step": 2460
1701
+ },
1702
+ {
1703
+ "epoch": 17.03,
1704
+ "learning_rate": 2.273684210526316e-05,
1705
+ "loss": 0.0785,
1706
+ "step": 2470
1707
+ },
1708
+ {
1709
+ "epoch": 17.1,
1710
+ "learning_rate": 2.2210526315789476e-05,
1711
+ "loss": 0.0316,
1712
+ "step": 2480
1713
+ },
1714
+ {
1715
+ "epoch": 17.17,
1716
+ "learning_rate": 2.168421052631579e-05,
1717
+ "loss": 0.0411,
1718
+ "step": 2490
1719
+ },
1720
+ {
1721
+ "epoch": 17.24,
1722
+ "learning_rate": 2.1157894736842106e-05,
1723
+ "loss": 0.0674,
1724
+ "step": 2500
1725
+ },
1726
+ {
1727
+ "epoch": 17.24,
1728
+ "eval_loss": 0.3407049775123596,
1729
+ "eval_runtime": 55.8959,
1730
+ "eval_samples_per_second": 30.056,
1731
+ "eval_steps_per_second": 30.056,
1732
+ "eval_wer": 0.21659430776652194,
1733
+ "step": 2500
1734
+ },
1735
+ {
1736
+ "epoch": 17.31,
1737
+ "learning_rate": 2.0631578947368422e-05,
1738
+ "loss": 0.0392,
1739
+ "step": 2510
1740
+ },
1741
+ {
1742
+ "epoch": 17.38,
1743
+ "learning_rate": 2.010526315789474e-05,
1744
+ "loss": 0.0248,
1745
+ "step": 2520
1746
+ },
1747
+ {
1748
+ "epoch": 17.45,
1749
+ "learning_rate": 1.957894736842105e-05,
1750
+ "loss": 0.0336,
1751
+ "step": 2530
1752
+ },
1753
+ {
1754
+ "epoch": 17.52,
1755
+ "learning_rate": 1.9052631578947368e-05,
1756
+ "loss": 0.056,
1757
+ "step": 2540
1758
+ },
1759
+ {
1760
+ "epoch": 17.59,
1761
+ "learning_rate": 1.8526315789473687e-05,
1762
+ "loss": 0.0397,
1763
+ "step": 2550
1764
+ },
1765
+ {
1766
+ "epoch": 17.66,
1767
+ "learning_rate": 1.8e-05,
1768
+ "loss": 0.0307,
1769
+ "step": 2560
1770
+ },
1771
+ {
1772
+ "epoch": 17.72,
1773
+ "learning_rate": 1.7473684210526317e-05,
1774
+ "loss": 0.0417,
1775
+ "step": 2570
1776
+ },
1777
+ {
1778
+ "epoch": 17.79,
1779
+ "learning_rate": 1.694736842105263e-05,
1780
+ "loss": 0.0646,
1781
+ "step": 2580
1782
+ },
1783
+ {
1784
+ "epoch": 17.86,
1785
+ "learning_rate": 1.642105263157895e-05,
1786
+ "loss": 0.0231,
1787
+ "step": 2590
1788
+ },
1789
+ {
1790
+ "epoch": 17.93,
1791
+ "learning_rate": 1.5894736842105266e-05,
1792
+ "loss": 0.0291,
1793
+ "step": 2600
1794
+ },
1795
+ {
1796
+ "epoch": 17.93,
1797
+ "eval_loss": 0.3512323200702667,
1798
+ "eval_runtime": 56.0829,
1799
+ "eval_samples_per_second": 29.956,
1800
+ "eval_steps_per_second": 29.956,
1801
+ "eval_wer": 0.2170767004341534,
1802
+ "step": 2600
1803
+ },
1804
+ {
1805
+ "epoch": 18.0,
1806
+ "learning_rate": 1.536842105263158e-05,
1807
+ "loss": 0.0675,
1808
+ "step": 2610
1809
+ },
1810
+ {
1811
+ "epoch": 18.07,
1812
+ "learning_rate": 1.4842105263157895e-05,
1813
+ "loss": 0.0357,
1814
+ "step": 2620
1815
+ },
1816
+ {
1817
+ "epoch": 18.14,
1818
+ "learning_rate": 1.431578947368421e-05,
1819
+ "loss": 0.0254,
1820
+ "step": 2630
1821
+ },
1822
+ {
1823
+ "epoch": 18.21,
1824
+ "learning_rate": 1.3789473684210526e-05,
1825
+ "loss": 0.0414,
1826
+ "step": 2640
1827
+ },
1828
+ {
1829
+ "epoch": 18.28,
1830
+ "learning_rate": 1.3263157894736844e-05,
1831
+ "loss": 0.0572,
1832
+ "step": 2650
1833
+ },
1834
+ {
1835
+ "epoch": 18.34,
1836
+ "learning_rate": 1.2736842105263157e-05,
1837
+ "loss": 0.0272,
1838
+ "step": 2660
1839
+ },
1840
+ {
1841
+ "epoch": 18.41,
1842
+ "learning_rate": 1.2210526315789474e-05,
1843
+ "loss": 0.0359,
1844
+ "step": 2670
1845
+ },
1846
+ {
1847
+ "epoch": 18.48,
1848
+ "learning_rate": 1.168421052631579e-05,
1849
+ "loss": 0.0543,
1850
+ "step": 2680
1851
+ },
1852
+ {
1853
+ "epoch": 18.55,
1854
+ "learning_rate": 1.1157894736842106e-05,
1855
+ "loss": 0.0579,
1856
+ "step": 2690
1857
+ },
1858
+ {
1859
+ "epoch": 18.62,
1860
+ "learning_rate": 1.0631578947368421e-05,
1861
+ "loss": 0.0298,
1862
+ "step": 2700
1863
+ },
1864
+ {
1865
+ "epoch": 18.62,
1866
+ "eval_loss": 0.3363025188446045,
1867
+ "eval_runtime": 56.3537,
1868
+ "eval_samples_per_second": 29.812,
1869
+ "eval_steps_per_second": 29.812,
1870
+ "eval_wer": 0.21576734890772517,
1871
+ "step": 2700
1872
+ },
1873
+ {
1874
+ "epoch": 18.69,
1875
+ "learning_rate": 1.0105263157894738e-05,
1876
+ "loss": 0.0423,
1877
+ "step": 2710
1878
+ },
1879
+ {
1880
+ "epoch": 18.76,
1881
+ "learning_rate": 9.578947368421052e-06,
1882
+ "loss": 0.0574,
1883
+ "step": 2720
1884
+ },
1885
+ {
1886
+ "epoch": 18.83,
1887
+ "learning_rate": 9.05263157894737e-06,
1888
+ "loss": 0.0267,
1889
+ "step": 2730
1890
+ },
1891
+ {
1892
+ "epoch": 18.9,
1893
+ "learning_rate": 8.526315789473685e-06,
1894
+ "loss": 0.035,
1895
+ "step": 2740
1896
+ },
1897
+ {
1898
+ "epoch": 18.97,
1899
+ "learning_rate": 8.000000000000001e-06,
1900
+ "loss": 0.0429,
1901
+ "step": 2750
1902
+ },
1903
+ {
1904
+ "epoch": 19.03,
1905
+ "learning_rate": 7.473684210526316e-06,
1906
+ "loss": 0.0634,
1907
+ "step": 2760
1908
+ },
1909
+ {
1910
+ "epoch": 19.1,
1911
+ "learning_rate": 6.9473684210526315e-06,
1912
+ "loss": 0.0315,
1913
+ "step": 2770
1914
+ },
1915
+ {
1916
+ "epoch": 19.17,
1917
+ "learning_rate": 6.421052631578947e-06,
1918
+ "loss": 0.0238,
1919
+ "step": 2780
1920
+ },
1921
+ {
1922
+ "epoch": 19.24,
1923
+ "learning_rate": 5.8947368421052634e-06,
1924
+ "loss": 0.0556,
1925
+ "step": 2790
1926
+ },
1927
+ {
1928
+ "epoch": 19.31,
1929
+ "learning_rate": 5.36842105263158e-06,
1930
+ "loss": 0.0419,
1931
+ "step": 2800
1932
+ },
1933
+ {
1934
+ "epoch": 19.31,
1935
+ "eval_loss": 0.34925296902656555,
1936
+ "eval_runtime": 56.3051,
1937
+ "eval_samples_per_second": 29.837,
1938
+ "eval_steps_per_second": 29.837,
1939
+ "eval_wer": 0.21452691061953,
1940
+ "step": 2800
1941
+ },
1942
+ {
1943
+ "epoch": 19.38,
1944
+ "learning_rate": 4.842105263157895e-06,
1945
+ "loss": 0.0274,
1946
+ "step": 2810
1947
+ },
1948
+ {
1949
+ "epoch": 19.45,
1950
+ "learning_rate": 4.315789473684211e-06,
1951
+ "loss": 0.0365,
1952
+ "step": 2820
1953
+ },
1954
+ {
1955
+ "epoch": 19.52,
1956
+ "learning_rate": 3.789473684210527e-06,
1957
+ "loss": 0.0517,
1958
+ "step": 2830
1959
+ },
1960
+ {
1961
+ "epoch": 19.59,
1962
+ "learning_rate": 3.2631578947368423e-06,
1963
+ "loss": 0.0274,
1964
+ "step": 2840
1965
+ },
1966
+ {
1967
+ "epoch": 19.66,
1968
+ "learning_rate": 2.7368421052631583e-06,
1969
+ "loss": 0.0162,
1970
+ "step": 2850
1971
+ },
1972
+ {
1973
+ "epoch": 19.72,
1974
+ "learning_rate": 2.2105263157894738e-06,
1975
+ "loss": 0.0378,
1976
+ "step": 2860
1977
+ },
1978
+ {
1979
+ "epoch": 19.79,
1980
+ "learning_rate": 1.6842105263157895e-06,
1981
+ "loss": 0.0544,
1982
+ "step": 2870
1983
+ },
1984
+ {
1985
+ "epoch": 19.86,
1986
+ "learning_rate": 1.1578947368421055e-06,
1987
+ "loss": 0.0249,
1988
+ "step": 2880
1989
+ },
1990
+ {
1991
+ "epoch": 19.93,
1992
+ "learning_rate": 6.315789473684211e-07,
1993
+ "loss": 0.0284,
1994
+ "step": 2890
1995
+ },
1996
+ {
1997
+ "epoch": 20.0,
1998
+ "learning_rate": 1.0526315789473685e-07,
1999
+ "loss": 0.046,
2000
+ "step": 2900
2001
+ },
2002
+ {
2003
+ "epoch": 20.0,
2004
+ "eval_loss": 0.3457382619380951,
2005
+ "eval_runtime": 56.4432,
2006
+ "eval_samples_per_second": 29.764,
2007
+ "eval_steps_per_second": 29.764,
2008
+ "eval_wer": 0.2151471297636276,
2009
+ "step": 2900
2010
+ },
2011
+ {
2012
+ "epoch": 20.0,
2013
+ "step": 2900,
2014
+ "total_flos": 2.6571394113263534e+18,
2015
+ "train_loss": 0.6826766253185683,
2016
+ "train_runtime": 4678.6424,
2017
+ "train_samples_per_second": 19.749,
2018
+ "train_steps_per_second": 0.62
2019
+ }
2020
+ ],
2021
+ "max_steps": 2900,
2022
+ "num_train_epochs": 20,
2023
+ "total_flos": 2.6571394113263534e+18,
2024
+ "trial_name": null,
2025
+ "trial_params": null
2026
+ }