Leonel-Maia commited on
Commit
84d172e
·
verified ·
1 Parent(s): afa45b0

End of training

Browse files
Files changed (5) hide show
  1. README.md +16 -4
  2. all_results.json +15 -0
  3. eval_results.json +9 -0
  4. train_results.json +9 -0
  5. trainer_state.json +946 -0
README.md CHANGED
@@ -4,11 +4,23 @@ license: apache-2.0
4
  base_model: openai/whisper-small
5
  tags:
6
  - generated_from_trainer
 
 
7
  metrics:
8
  - wer
9
  model-index:
10
  - name: whisper-small-splitted
11
- results: []
 
 
 
 
 
 
 
 
 
 
12
  ---
13
 
14
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -16,10 +28,10 @@ should probably proofread and complete it, then remove this comment. -->
16
 
17
  # whisper-small-splitted
18
 
19
- This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 0.1156
22
- - Wer: 0.0992
23
 
24
  ## Model description
25
 
 
4
  base_model: openai/whisper-small
5
  tags:
6
  - generated_from_trainer
7
+ datasets:
8
+ - Leonel-Maia/fongbe-splitted
9
  metrics:
10
  - wer
11
  model-index:
12
  - name: whisper-small-splitted
13
+ results:
14
+ - task:
15
+ name: Automatic Speech Recognition
16
+ type: automatic-speech-recognition
17
+ dataset:
18
+ name: Leonel-Maia/fongbe-splitted
19
+ type: Leonel-Maia/fongbe-splitted
20
+ metrics:
21
+ - name: Wer
22
+ type: wer
23
+ value: 0.11257288805358778
24
  ---
25
 
26
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
28
 
29
  # whisper-small-splitted
30
 
31
+ This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Leonel-Maia/fongbe-splitted dataset.
32
  It achieves the following results on the evaluation set:
33
+ - Loss: 0.1084
34
+ - Wer: 0.1126
35
 
36
  ## Model description
37
 
all_results.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 12.656151419558359,
3
+ "eval_loss": 0.10838954150676727,
4
+ "eval_runtime": 1153.8599,
5
+ "eval_samples": 3259,
6
+ "eval_samples_per_second": 2.824,
7
+ "eval_steps_per_second": 0.706,
8
+ "eval_wer": 0.11257288805358778,
9
+ "total_flos": 5.55469178535936e+19,
10
+ "train_loss": 0.17092564110457897,
11
+ "train_runtime": 92948.5763,
12
+ "train_samples": 15213,
13
+ "train_samples_per_second": 9.82,
14
+ "train_steps_per_second": 0.153
15
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 12.656151419558359,
3
+ "eval_loss": 0.10838954150676727,
4
+ "eval_runtime": 1153.8599,
5
+ "eval_samples": 3259,
6
+ "eval_samples_per_second": 2.824,
7
+ "eval_steps_per_second": 0.706,
8
+ "eval_wer": 0.11257288805358778
9
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 12.656151419558359,
3
+ "total_flos": 5.55469178535936e+19,
4
+ "train_loss": 0.17092564110457897,
5
+ "train_runtime": 92948.5763,
6
+ "train_samples": 15213,
7
+ "train_samples_per_second": 9.82,
8
+ "train_steps_per_second": 0.153
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,946 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 1500,
3
+ "best_metric": 0.10838954150676727,
4
+ "best_model_checkpoint": "./whisper-small-splitted/checkpoint-1500",
5
+ "epoch": 12.656151419558359,
6
+ "eval_steps": 500,
7
+ "global_step": 3000,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.10515247108307045,
14
+ "grad_norm": 37.98896408081055,
15
+ "learning_rate": 4.800000000000001e-07,
16
+ "loss": 4.0211,
17
+ "step": 25
18
+ },
19
+ {
20
+ "epoch": 0.2103049421661409,
21
+ "grad_norm": 28.073650360107422,
22
+ "learning_rate": 9.800000000000001e-07,
23
+ "loss": 3.1838,
24
+ "step": 50
25
+ },
26
+ {
27
+ "epoch": 0.31545741324921134,
28
+ "grad_norm": 12.10400676727295,
29
+ "learning_rate": 1.48e-06,
30
+ "loss": 2.0459,
31
+ "step": 75
32
+ },
33
+ {
34
+ "epoch": 0.4206098843322818,
35
+ "grad_norm": 8.05373477935791,
36
+ "learning_rate": 1.98e-06,
37
+ "loss": 1.3254,
38
+ "step": 100
39
+ },
40
+ {
41
+ "epoch": 0.5257623554153522,
42
+ "grad_norm": 6.276533126831055,
43
+ "learning_rate": 2.4800000000000004e-06,
44
+ "loss": 1.0124,
45
+ "step": 125
46
+ },
47
+ {
48
+ "epoch": 0.6309148264984227,
49
+ "grad_norm": 5.450071334838867,
50
+ "learning_rate": 2.9800000000000003e-06,
51
+ "loss": 0.7972,
52
+ "step": 150
53
+ },
54
+ {
55
+ "epoch": 0.7360672975814931,
56
+ "grad_norm": 6.463825225830078,
57
+ "learning_rate": 3.48e-06,
58
+ "loss": 0.6987,
59
+ "step": 175
60
+ },
61
+ {
62
+ "epoch": 0.8412197686645636,
63
+ "grad_norm": 6.113511085510254,
64
+ "learning_rate": 3.980000000000001e-06,
65
+ "loss": 0.6371,
66
+ "step": 200
67
+ },
68
+ {
69
+ "epoch": 0.9463722397476341,
70
+ "grad_norm": 5.203503131866455,
71
+ "learning_rate": 4.48e-06,
72
+ "loss": 0.5403,
73
+ "step": 225
74
+ },
75
+ {
76
+ "epoch": 1.0546792849631967,
77
+ "grad_norm": 3.6247618198394775,
78
+ "learning_rate": 4.980000000000001e-06,
79
+ "loss": 0.4597,
80
+ "step": 250
81
+ },
82
+ {
83
+ "epoch": 1.159831756046267,
84
+ "grad_norm": 4.7159552574157715,
85
+ "learning_rate": 5.480000000000001e-06,
86
+ "loss": 0.4074,
87
+ "step": 275
88
+ },
89
+ {
90
+ "epoch": 1.2649842271293377,
91
+ "grad_norm": 5.012338638305664,
92
+ "learning_rate": 5.98e-06,
93
+ "loss": 0.3846,
94
+ "step": 300
95
+ },
96
+ {
97
+ "epoch": 1.370136698212408,
98
+ "grad_norm": 4.738580703735352,
99
+ "learning_rate": 6.480000000000001e-06,
100
+ "loss": 0.3306,
101
+ "step": 325
102
+ },
103
+ {
104
+ "epoch": 1.4752891692954784,
105
+ "grad_norm": 4.1813578605651855,
106
+ "learning_rate": 6.98e-06,
107
+ "loss": 0.3145,
108
+ "step": 350
109
+ },
110
+ {
111
+ "epoch": 1.5804416403785488,
112
+ "grad_norm": 5.1083855628967285,
113
+ "learning_rate": 7.48e-06,
114
+ "loss": 0.2901,
115
+ "step": 375
116
+ },
117
+ {
118
+ "epoch": 1.6855941114616193,
119
+ "grad_norm": 3.654137372970581,
120
+ "learning_rate": 7.980000000000002e-06,
121
+ "loss": 0.2473,
122
+ "step": 400
123
+ },
124
+ {
125
+ "epoch": 1.7907465825446898,
126
+ "grad_norm": 5.576925754547119,
127
+ "learning_rate": 8.48e-06,
128
+ "loss": 0.2408,
129
+ "step": 425
130
+ },
131
+ {
132
+ "epoch": 1.8958990536277602,
133
+ "grad_norm": 5.050299167633057,
134
+ "learning_rate": 8.98e-06,
135
+ "loss": 0.222,
136
+ "step": 450
137
+ },
138
+ {
139
+ "epoch": 2.004206098843323,
140
+ "grad_norm": 4.0772705078125,
141
+ "learning_rate": 9.48e-06,
142
+ "loss": 0.2311,
143
+ "step": 475
144
+ },
145
+ {
146
+ "epoch": 2.1093585699263935,
147
+ "grad_norm": 3.189274311065674,
148
+ "learning_rate": 9.980000000000001e-06,
149
+ "loss": 0.1324,
150
+ "step": 500
151
+ },
152
+ {
153
+ "epoch": 2.1093585699263935,
154
+ "eval_loss": 0.1909431368112564,
155
+ "eval_runtime": 1339.9903,
156
+ "eval_samples_per_second": 2.432,
157
+ "eval_steps_per_second": 0.608,
158
+ "eval_wer": 0.3306828586574141,
159
+ "step": 500
160
+ },
161
+ {
162
+ "epoch": 2.2145110410094637,
163
+ "grad_norm": 3.6630983352661133,
164
+ "learning_rate": 9.982507288629738e-06,
165
+ "loss": 0.1425,
166
+ "step": 525
167
+ },
168
+ {
169
+ "epoch": 2.319663512092534,
170
+ "grad_norm": 3.793405532836914,
171
+ "learning_rate": 9.964285714285714e-06,
172
+ "loss": 0.1394,
173
+ "step": 550
174
+ },
175
+ {
176
+ "epoch": 2.4248159831756047,
177
+ "grad_norm": 3.442934036254883,
178
+ "learning_rate": 9.94606413994169e-06,
179
+ "loss": 0.1445,
180
+ "step": 575
181
+ },
182
+ {
183
+ "epoch": 2.5299684542586753,
184
+ "grad_norm": 2.4245970249176025,
185
+ "learning_rate": 9.927842565597668e-06,
186
+ "loss": 0.1247,
187
+ "step": 600
188
+ },
189
+ {
190
+ "epoch": 2.6351209253417456,
191
+ "grad_norm": 2.134147882461548,
192
+ "learning_rate": 9.909620991253646e-06,
193
+ "loss": 0.1246,
194
+ "step": 625
195
+ },
196
+ {
197
+ "epoch": 2.740273396424816,
198
+ "grad_norm": 3.135070562362671,
199
+ "learning_rate": 9.891399416909622e-06,
200
+ "loss": 0.1233,
201
+ "step": 650
202
+ },
203
+ {
204
+ "epoch": 2.8454258675078865,
205
+ "grad_norm": 2.723932981491089,
206
+ "learning_rate": 9.873177842565598e-06,
207
+ "loss": 0.119,
208
+ "step": 675
209
+ },
210
+ {
211
+ "epoch": 2.9505783385909568,
212
+ "grad_norm": 3.9131784439086914,
213
+ "learning_rate": 9.854956268221574e-06,
214
+ "loss": 0.1334,
215
+ "step": 700
216
+ },
217
+ {
218
+ "epoch": 3.0588853838065195,
219
+ "grad_norm": 4.044662952423096,
220
+ "learning_rate": 9.836734693877552e-06,
221
+ "loss": 0.0902,
222
+ "step": 725
223
+ },
224
+ {
225
+ "epoch": 3.1640378548895898,
226
+ "grad_norm": 2.1829943656921387,
227
+ "learning_rate": 9.818513119533528e-06,
228
+ "loss": 0.0702,
229
+ "step": 750
230
+ },
231
+ {
232
+ "epoch": 3.2691903259726605,
233
+ "grad_norm": 2.548558235168457,
234
+ "learning_rate": 9.800291545189506e-06,
235
+ "loss": 0.0712,
236
+ "step": 775
237
+ },
238
+ {
239
+ "epoch": 3.3743427970557307,
240
+ "grad_norm": 1.3480279445648193,
241
+ "learning_rate": 9.782069970845482e-06,
242
+ "loss": 0.0654,
243
+ "step": 800
244
+ },
245
+ {
246
+ "epoch": 3.4794952681388014,
247
+ "grad_norm": 2.091961622238159,
248
+ "learning_rate": 9.763848396501458e-06,
249
+ "loss": 0.0658,
250
+ "step": 825
251
+ },
252
+ {
253
+ "epoch": 3.5846477392218716,
254
+ "grad_norm": 2.2356009483337402,
255
+ "learning_rate": 9.745626822157436e-06,
256
+ "loss": 0.0682,
257
+ "step": 850
258
+ },
259
+ {
260
+ "epoch": 3.6898002103049423,
261
+ "grad_norm": 2.5908384323120117,
262
+ "learning_rate": 9.727405247813412e-06,
263
+ "loss": 0.056,
264
+ "step": 875
265
+ },
266
+ {
267
+ "epoch": 3.7949526813880126,
268
+ "grad_norm": 2.618177890777588,
269
+ "learning_rate": 9.70918367346939e-06,
270
+ "loss": 0.0623,
271
+ "step": 900
272
+ },
273
+ {
274
+ "epoch": 3.900105152471083,
275
+ "grad_norm": 2.1948764324188232,
276
+ "learning_rate": 9.690962099125366e-06,
277
+ "loss": 0.0752,
278
+ "step": 925
279
+ },
280
+ {
281
+ "epoch": 4.008412197686646,
282
+ "grad_norm": 1.5540448427200317,
283
+ "learning_rate": 9.672740524781342e-06,
284
+ "loss": 0.0638,
285
+ "step": 950
286
+ },
287
+ {
288
+ "epoch": 4.113564668769716,
289
+ "grad_norm": 1.905039668083191,
290
+ "learning_rate": 9.654518950437318e-06,
291
+ "loss": 0.0397,
292
+ "step": 975
293
+ },
294
+ {
295
+ "epoch": 4.218717139852787,
296
+ "grad_norm": 2.011355400085449,
297
+ "learning_rate": 9.636297376093296e-06,
298
+ "loss": 0.0383,
299
+ "step": 1000
300
+ },
301
+ {
302
+ "epoch": 4.218717139852787,
303
+ "eval_loss": 0.1153440847992897,
304
+ "eval_runtime": 1137.0265,
305
+ "eval_samples_per_second": 2.866,
306
+ "eval_steps_per_second": 0.717,
307
+ "eval_wer": 0.11305479254011855,
308
+ "step": 1000
309
+ },
310
+ {
311
+ "epoch": 4.323869610935857,
312
+ "grad_norm": 1.983497142791748,
313
+ "learning_rate": 9.618075801749272e-06,
314
+ "loss": 0.0376,
315
+ "step": 1025
316
+ },
317
+ {
318
+ "epoch": 4.429022082018927,
319
+ "grad_norm": 2.7597391605377197,
320
+ "learning_rate": 9.599854227405248e-06,
321
+ "loss": 0.0384,
322
+ "step": 1050
323
+ },
324
+ {
325
+ "epoch": 4.534174553101998,
326
+ "grad_norm": 2.3190319538116455,
327
+ "learning_rate": 9.581632653061226e-06,
328
+ "loss": 0.041,
329
+ "step": 1075
330
+ },
331
+ {
332
+ "epoch": 4.639327024185068,
333
+ "grad_norm": 1.6487782001495361,
334
+ "learning_rate": 9.563411078717202e-06,
335
+ "loss": 0.0373,
336
+ "step": 1100
337
+ },
338
+ {
339
+ "epoch": 4.744479495268139,
340
+ "grad_norm": 1.493482232093811,
341
+ "learning_rate": 9.54518950437318e-06,
342
+ "loss": 0.0453,
343
+ "step": 1125
344
+ },
345
+ {
346
+ "epoch": 4.849631966351209,
347
+ "grad_norm": 1.9630647897720337,
348
+ "learning_rate": 9.526967930029156e-06,
349
+ "loss": 0.0447,
350
+ "step": 1150
351
+ },
352
+ {
353
+ "epoch": 4.95478443743428,
354
+ "grad_norm": 2.1700949668884277,
355
+ "learning_rate": 9.508746355685132e-06,
356
+ "loss": 0.0416,
357
+ "step": 1175
358
+ },
359
+ {
360
+ "epoch": 5.063091482649842,
361
+ "grad_norm": 2.0949151515960693,
362
+ "learning_rate": 9.490524781341108e-06,
363
+ "loss": 0.0296,
364
+ "step": 1200
365
+ },
366
+ {
367
+ "epoch": 5.168243953732913,
368
+ "grad_norm": 1.232323169708252,
369
+ "learning_rate": 9.472303206997086e-06,
370
+ "loss": 0.0227,
371
+ "step": 1225
372
+ },
373
+ {
374
+ "epoch": 5.273396424815983,
375
+ "grad_norm": 6.17132568359375,
376
+ "learning_rate": 9.454081632653062e-06,
377
+ "loss": 0.0284,
378
+ "step": 1250
379
+ },
380
+ {
381
+ "epoch": 5.378548895899054,
382
+ "grad_norm": 0.9878859519958496,
383
+ "learning_rate": 9.43586005830904e-06,
384
+ "loss": 0.0211,
385
+ "step": 1275
386
+ },
387
+ {
388
+ "epoch": 5.483701366982124,
389
+ "grad_norm": 2.1974642276763916,
390
+ "learning_rate": 9.417638483965016e-06,
391
+ "loss": 0.0203,
392
+ "step": 1300
393
+ },
394
+ {
395
+ "epoch": 5.588853838065194,
396
+ "grad_norm": 1.2606751918792725,
397
+ "learning_rate": 9.399416909620992e-06,
398
+ "loss": 0.0271,
399
+ "step": 1325
400
+ },
401
+ {
402
+ "epoch": 5.694006309148265,
403
+ "grad_norm": 1.851982593536377,
404
+ "learning_rate": 9.38119533527697e-06,
405
+ "loss": 0.0271,
406
+ "step": 1350
407
+ },
408
+ {
409
+ "epoch": 5.799158780231336,
410
+ "grad_norm": 1.9509834051132202,
411
+ "learning_rate": 9.362973760932945e-06,
412
+ "loss": 0.0261,
413
+ "step": 1375
414
+ },
415
+ {
416
+ "epoch": 5.904311251314406,
417
+ "grad_norm": 1.3138453960418701,
418
+ "learning_rate": 9.344752186588922e-06,
419
+ "loss": 0.0269,
420
+ "step": 1400
421
+ },
422
+ {
423
+ "epoch": 6.012618296529968,
424
+ "grad_norm": 1.0709035396575928,
425
+ "learning_rate": 9.326530612244898e-06,
426
+ "loss": 0.0292,
427
+ "step": 1425
428
+ },
429
+ {
430
+ "epoch": 6.117770767613039,
431
+ "grad_norm": 0.9309059381484985,
432
+ "learning_rate": 9.308309037900875e-06,
433
+ "loss": 0.0128,
434
+ "step": 1450
435
+ },
436
+ {
437
+ "epoch": 6.22292323869611,
438
+ "grad_norm": 1.078444004058838,
439
+ "learning_rate": 9.290087463556851e-06,
440
+ "loss": 0.0189,
441
+ "step": 1475
442
+ },
443
+ {
444
+ "epoch": 6.3280757097791795,
445
+ "grad_norm": 2.0890743732452393,
446
+ "learning_rate": 9.27186588921283e-06,
447
+ "loss": 0.0184,
448
+ "step": 1500
449
+ },
450
+ {
451
+ "epoch": 6.3280757097791795,
452
+ "eval_loss": 0.10838954150676727,
453
+ "eval_runtime": 1142.8117,
454
+ "eval_samples_per_second": 2.852,
455
+ "eval_steps_per_second": 0.713,
456
+ "eval_wer": 0.11257288805358778,
457
+ "step": 1500
458
+ },
459
+ {
460
+ "epoch": 6.43322818086225,
461
+ "grad_norm": 0.5473038554191589,
462
+ "learning_rate": 9.253644314868805e-06,
463
+ "loss": 0.0153,
464
+ "step": 1525
465
+ },
466
+ {
467
+ "epoch": 6.538380651945321,
468
+ "grad_norm": 1.075249433517456,
469
+ "learning_rate": 9.235422740524781e-06,
470
+ "loss": 0.0172,
471
+ "step": 1550
472
+ },
473
+ {
474
+ "epoch": 6.643533123028391,
475
+ "grad_norm": 1.237701177597046,
476
+ "learning_rate": 9.21720116618076e-06,
477
+ "loss": 0.0167,
478
+ "step": 1575
479
+ },
480
+ {
481
+ "epoch": 6.748685594111461,
482
+ "grad_norm": 1.9220877885818481,
483
+ "learning_rate": 9.198979591836735e-06,
484
+ "loss": 0.0175,
485
+ "step": 1600
486
+ },
487
+ {
488
+ "epoch": 6.853838065194532,
489
+ "grad_norm": 1.435571312904358,
490
+ "learning_rate": 9.180758017492713e-06,
491
+ "loss": 0.0182,
492
+ "step": 1625
493
+ },
494
+ {
495
+ "epoch": 6.958990536277603,
496
+ "grad_norm": 1.9936102628707886,
497
+ "learning_rate": 9.162536443148689e-06,
498
+ "loss": 0.0158,
499
+ "step": 1650
500
+ },
501
+ {
502
+ "epoch": 7.067297581493165,
503
+ "grad_norm": 1.101319670677185,
504
+ "learning_rate": 9.144314868804665e-06,
505
+ "loss": 0.0147,
506
+ "step": 1675
507
+ },
508
+ {
509
+ "epoch": 7.172450052576235,
510
+ "grad_norm": 1.3670756816864014,
511
+ "learning_rate": 9.126093294460641e-06,
512
+ "loss": 0.0127,
513
+ "step": 1700
514
+ },
515
+ {
516
+ "epoch": 7.277602523659306,
517
+ "grad_norm": 0.6474692225456238,
518
+ "learning_rate": 9.107871720116619e-06,
519
+ "loss": 0.0107,
520
+ "step": 1725
521
+ },
522
+ {
523
+ "epoch": 7.382754994742377,
524
+ "grad_norm": 0.941483736038208,
525
+ "learning_rate": 9.089650145772595e-06,
526
+ "loss": 0.0129,
527
+ "step": 1750
528
+ },
529
+ {
530
+ "epoch": 7.4879074658254465,
531
+ "grad_norm": 0.7835807204246521,
532
+ "learning_rate": 9.071428571428573e-06,
533
+ "loss": 0.015,
534
+ "step": 1775
535
+ },
536
+ {
537
+ "epoch": 7.593059936908517,
538
+ "grad_norm": 0.7577527761459351,
539
+ "learning_rate": 9.053206997084549e-06,
540
+ "loss": 0.0119,
541
+ "step": 1800
542
+ },
543
+ {
544
+ "epoch": 7.698212407991588,
545
+ "grad_norm": 1.8214268684387207,
546
+ "learning_rate": 9.034985422740525e-06,
547
+ "loss": 0.0116,
548
+ "step": 1825
549
+ },
550
+ {
551
+ "epoch": 7.803364879074659,
552
+ "grad_norm": 0.8769547343254089,
553
+ "learning_rate": 9.016763848396503e-06,
554
+ "loss": 0.0131,
555
+ "step": 1850
556
+ },
557
+ {
558
+ "epoch": 7.908517350157728,
559
+ "grad_norm": 1.227581262588501,
560
+ "learning_rate": 8.998542274052479e-06,
561
+ "loss": 0.012,
562
+ "step": 1875
563
+ },
564
+ {
565
+ "epoch": 8.016824395373291,
566
+ "grad_norm": 1.28843355178833,
567
+ "learning_rate": 8.980320699708455e-06,
568
+ "loss": 0.0134,
569
+ "step": 1900
570
+ },
571
+ {
572
+ "epoch": 8.121976866456361,
573
+ "grad_norm": 1.5814619064331055,
574
+ "learning_rate": 8.962099125364431e-06,
575
+ "loss": 0.0086,
576
+ "step": 1925
577
+ },
578
+ {
579
+ "epoch": 8.227129337539433,
580
+ "grad_norm": 0.8933718800544739,
581
+ "learning_rate": 8.943877551020409e-06,
582
+ "loss": 0.0089,
583
+ "step": 1950
584
+ },
585
+ {
586
+ "epoch": 8.332281808622502,
587
+ "grad_norm": 0.9096396565437317,
588
+ "learning_rate": 8.925655976676385e-06,
589
+ "loss": 0.0097,
590
+ "step": 1975
591
+ },
592
+ {
593
+ "epoch": 8.437434279705574,
594
+ "grad_norm": 0.42109328508377075,
595
+ "learning_rate": 8.907434402332363e-06,
596
+ "loss": 0.0098,
597
+ "step": 2000
598
+ },
599
+ {
600
+ "epoch": 8.437434279705574,
601
+ "eval_loss": 0.11219166219234467,
602
+ "eval_runtime": 1135.5499,
603
+ "eval_samples_per_second": 2.87,
604
+ "eval_steps_per_second": 0.718,
605
+ "eval_wer": 0.10139270396607393,
606
+ "step": 2000
607
+ },
608
+ {
609
+ "epoch": 8.542586750788644,
610
+ "grad_norm": 1.2390159368515015,
611
+ "learning_rate": 8.889212827988339e-06,
612
+ "loss": 0.0106,
613
+ "step": 2025
614
+ },
615
+ {
616
+ "epoch": 8.647739221871714,
617
+ "grad_norm": 0.9200573563575745,
618
+ "learning_rate": 8.870991253644315e-06,
619
+ "loss": 0.0082,
620
+ "step": 2050
621
+ },
622
+ {
623
+ "epoch": 8.752891692954785,
624
+ "grad_norm": 0.814224362373352,
625
+ "learning_rate": 8.852769679300293e-06,
626
+ "loss": 0.0105,
627
+ "step": 2075
628
+ },
629
+ {
630
+ "epoch": 8.858044164037855,
631
+ "grad_norm": 0.7412490248680115,
632
+ "learning_rate": 8.834548104956269e-06,
633
+ "loss": 0.0101,
634
+ "step": 2100
635
+ },
636
+ {
637
+ "epoch": 8.963196635120925,
638
+ "grad_norm": 0.8766764402389526,
639
+ "learning_rate": 8.816326530612247e-06,
640
+ "loss": 0.0092,
641
+ "step": 2125
642
+ },
643
+ {
644
+ "epoch": 9.071503680336487,
645
+ "grad_norm": 0.09623388946056366,
646
+ "learning_rate": 8.798104956268223e-06,
647
+ "loss": 0.0088,
648
+ "step": 2150
649
+ },
650
+ {
651
+ "epoch": 9.176656151419559,
652
+ "grad_norm": 0.6915100812911987,
653
+ "learning_rate": 8.779883381924199e-06,
654
+ "loss": 0.0073,
655
+ "step": 2175
656
+ },
657
+ {
658
+ "epoch": 9.281808622502629,
659
+ "grad_norm": 0.6448748707771301,
660
+ "learning_rate": 8.761661807580175e-06,
661
+ "loss": 0.0089,
662
+ "step": 2200
663
+ },
664
+ {
665
+ "epoch": 9.386961093585699,
666
+ "grad_norm": 1.2592213153839111,
667
+ "learning_rate": 8.743440233236153e-06,
668
+ "loss": 0.0096,
669
+ "step": 2225
670
+ },
671
+ {
672
+ "epoch": 9.49211356466877,
673
+ "grad_norm": 0.8664531111717224,
674
+ "learning_rate": 8.72521865889213e-06,
675
+ "loss": 0.0073,
676
+ "step": 2250
677
+ },
678
+ {
679
+ "epoch": 9.59726603575184,
680
+ "grad_norm": 0.49285972118377686,
681
+ "learning_rate": 8.706997084548106e-06,
682
+ "loss": 0.0078,
683
+ "step": 2275
684
+ },
685
+ {
686
+ "epoch": 9.70241850683491,
687
+ "grad_norm": 0.9507968425750732,
688
+ "learning_rate": 8.688775510204082e-06,
689
+ "loss": 0.0092,
690
+ "step": 2300
691
+ },
692
+ {
693
+ "epoch": 9.807570977917981,
694
+ "grad_norm": 1.0036101341247559,
695
+ "learning_rate": 8.670553935860059e-06,
696
+ "loss": 0.0103,
697
+ "step": 2325
698
+ },
699
+ {
700
+ "epoch": 9.912723449001051,
701
+ "grad_norm": 0.8855874538421631,
702
+ "learning_rate": 8.653061224489798e-06,
703
+ "loss": 0.0111,
704
+ "step": 2350
705
+ },
706
+ {
707
+ "epoch": 10.021030494216614,
708
+ "grad_norm": 0.918958306312561,
709
+ "learning_rate": 8.634839650145774e-06,
710
+ "loss": 0.0097,
711
+ "step": 2375
712
+ },
713
+ {
714
+ "epoch": 10.126182965299684,
715
+ "grad_norm": 1.442686915397644,
716
+ "learning_rate": 8.61661807580175e-06,
717
+ "loss": 0.0083,
718
+ "step": 2400
719
+ },
720
+ {
721
+ "epoch": 10.231335436382755,
722
+ "grad_norm": 1.1829726696014404,
723
+ "learning_rate": 8.598396501457726e-06,
724
+ "loss": 0.0076,
725
+ "step": 2425
726
+ },
727
+ {
728
+ "epoch": 10.336487907465825,
729
+ "grad_norm": 1.1316958665847778,
730
+ "learning_rate": 8.580174927113704e-06,
731
+ "loss": 0.0086,
732
+ "step": 2450
733
+ },
734
+ {
735
+ "epoch": 10.441640378548897,
736
+ "grad_norm": 0.5993586778640747,
737
+ "learning_rate": 8.56195335276968e-06,
738
+ "loss": 0.0085,
739
+ "step": 2475
740
+ },
741
+ {
742
+ "epoch": 10.546792849631967,
743
+ "grad_norm": 1.1639779806137085,
744
+ "learning_rate": 8.543731778425656e-06,
745
+ "loss": 0.0076,
746
+ "step": 2500
747
+ },
748
+ {
749
+ "epoch": 10.546792849631967,
750
+ "eval_loss": 0.1101309061050415,
751
+ "eval_runtime": 1133.8828,
752
+ "eval_samples_per_second": 2.874,
753
+ "eval_steps_per_second": 0.719,
754
+ "eval_wer": 0.09657365910076623,
755
+ "step": 2500
756
+ },
757
+ {
758
+ "epoch": 10.651945320715036,
759
+ "grad_norm": 1.1275087594985962,
760
+ "learning_rate": 8.525510204081632e-06,
761
+ "loss": 0.008,
762
+ "step": 2525
763
+ },
764
+ {
765
+ "epoch": 10.757097791798108,
766
+ "grad_norm": 1.17903470993042,
767
+ "learning_rate": 8.50728862973761e-06,
768
+ "loss": 0.0082,
769
+ "step": 2550
770
+ },
771
+ {
772
+ "epoch": 10.862250262881178,
773
+ "grad_norm": 0.35520169138908386,
774
+ "learning_rate": 8.489067055393588e-06,
775
+ "loss": 0.0085,
776
+ "step": 2575
777
+ },
778
+ {
779
+ "epoch": 10.967402733964247,
780
+ "grad_norm": 1.5375522375106812,
781
+ "learning_rate": 8.470845481049564e-06,
782
+ "loss": 0.0062,
783
+ "step": 2600
784
+ },
785
+ {
786
+ "epoch": 11.07570977917981,
787
+ "grad_norm": 0.49886855483055115,
788
+ "learning_rate": 8.45262390670554e-06,
789
+ "loss": 0.006,
790
+ "step": 2625
791
+ },
792
+ {
793
+ "epoch": 11.180862250262882,
794
+ "grad_norm": 0.5618315935134888,
795
+ "learning_rate": 8.434402332361516e-06,
796
+ "loss": 0.0053,
797
+ "step": 2650
798
+ },
799
+ {
800
+ "epoch": 11.286014721345952,
801
+ "grad_norm": 0.4988507032394409,
802
+ "learning_rate": 8.416180758017494e-06,
803
+ "loss": 0.006,
804
+ "step": 2675
805
+ },
806
+ {
807
+ "epoch": 11.391167192429021,
808
+ "grad_norm": 0.4689825773239136,
809
+ "learning_rate": 8.39795918367347e-06,
810
+ "loss": 0.0059,
811
+ "step": 2700
812
+ },
813
+ {
814
+ "epoch": 11.496319663512093,
815
+ "grad_norm": 0.637856662273407,
816
+ "learning_rate": 8.379737609329447e-06,
817
+ "loss": 0.0066,
818
+ "step": 2725
819
+ },
820
+ {
821
+ "epoch": 11.601472134595163,
822
+ "grad_norm": 0.46745094656944275,
823
+ "learning_rate": 8.361516034985424e-06,
824
+ "loss": 0.007,
825
+ "step": 2750
826
+ },
827
+ {
828
+ "epoch": 11.706624605678233,
829
+ "grad_norm": 1.153106689453125,
830
+ "learning_rate": 8.3432944606414e-06,
831
+ "loss": 0.0077,
832
+ "step": 2775
833
+ },
834
+ {
835
+ "epoch": 11.811777076761304,
836
+ "grad_norm": 1.1093589067459106,
837
+ "learning_rate": 8.325072886297377e-06,
838
+ "loss": 0.0068,
839
+ "step": 2800
840
+ },
841
+ {
842
+ "epoch": 11.916929547844374,
843
+ "grad_norm": 0.7349961400032043,
844
+ "learning_rate": 8.306851311953353e-06,
845
+ "loss": 0.0075,
846
+ "step": 2825
847
+ },
848
+ {
849
+ "epoch": 12.025236593059937,
850
+ "grad_norm": 3.071120500564575,
851
+ "learning_rate": 8.288629737609331e-06,
852
+ "loss": 0.0067,
853
+ "step": 2850
854
+ },
855
+ {
856
+ "epoch": 12.130389064143007,
857
+ "grad_norm": 1.33828604221344,
858
+ "learning_rate": 8.270408163265307e-06,
859
+ "loss": 0.0049,
860
+ "step": 2875
861
+ },
862
+ {
863
+ "epoch": 12.235541535226078,
864
+ "grad_norm": 0.5961343050003052,
865
+ "learning_rate": 8.252186588921283e-06,
866
+ "loss": 0.0058,
867
+ "step": 2900
868
+ },
869
+ {
870
+ "epoch": 12.340694006309148,
871
+ "grad_norm": 0.5771873593330383,
872
+ "learning_rate": 8.23396501457726e-06,
873
+ "loss": 0.0049,
874
+ "step": 2925
875
+ },
876
+ {
877
+ "epoch": 12.44584647739222,
878
+ "grad_norm": 1.0507683753967285,
879
+ "learning_rate": 8.215743440233237e-06,
880
+ "loss": 0.0065,
881
+ "step": 2950
882
+ },
883
+ {
884
+ "epoch": 12.55099894847529,
885
+ "grad_norm": 0.579768180847168,
886
+ "learning_rate": 8.197521865889213e-06,
887
+ "loss": 0.0058,
888
+ "step": 2975
889
+ },
890
+ {
891
+ "epoch": 12.656151419558359,
892
+ "grad_norm": 0.8235365748405457,
893
+ "learning_rate": 8.17930029154519e-06,
894
+ "loss": 0.0075,
895
+ "step": 3000
896
+ },
897
+ {
898
+ "epoch": 12.656151419558359,
899
+ "eval_loss": 0.11558376252651215,
900
+ "eval_runtime": 1148.2367,
901
+ "eval_samples_per_second": 2.838,
902
+ "eval_steps_per_second": 0.71,
903
+ "eval_wer": 0.09917594332803238,
904
+ "step": 3000
905
+ },
906
+ {
907
+ "epoch": 12.656151419558359,
908
+ "step": 3000,
909
+ "total_flos": 5.55469178535936e+19,
910
+ "train_loss": 0.17092564110457897,
911
+ "train_runtime": 92948.5763,
912
+ "train_samples_per_second": 9.82,
913
+ "train_steps_per_second": 0.153
914
+ }
915
+ ],
916
+ "logging_steps": 25,
917
+ "max_steps": 14220,
918
+ "num_input_tokens_seen": 0,
919
+ "num_train_epochs": 60,
920
+ "save_steps": 500,
921
+ "stateful_callbacks": {
922
+ "EarlyStoppingCallback": {
923
+ "args": {
924
+ "early_stopping_patience": 3,
925
+ "early_stopping_threshold": 0.0
926
+ },
927
+ "attributes": {
928
+ "early_stopping_patience_counter": 3
929
+ }
930
+ },
931
+ "TrainerControl": {
932
+ "args": {
933
+ "should_epoch_stop": false,
934
+ "should_evaluate": false,
935
+ "should_log": false,
936
+ "should_save": true,
937
+ "should_training_stop": true
938
+ },
939
+ "attributes": {}
940
+ }
941
+ },
942
+ "total_flos": 5.55469178535936e+19,
943
+ "train_batch_size": 8,
944
+ "trial_name": null,
945
+ "trial_params": null
946
+ }