GIanlucaRub commited on
Commit
c27e9d2
1 Parent(s): 962ad37

Copied from Google Drive

Browse files
checkpoint-1000/config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-tiny",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "architectures": [
6
+ "WhisperForConditionalGeneration"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "begin_suppress_tokens": [
10
+ 220,
11
+ 50257
12
+ ],
13
+ "bos_token_id": 50257,
14
+ "d_model": 384,
15
+ "decoder_attention_heads": 6,
16
+ "decoder_ffn_dim": 1536,
17
+ "decoder_layerdrop": 0.0,
18
+ "decoder_layers": 4,
19
+ "decoder_start_token_id": 50258,
20
+ "dropout": 0.0,
21
+ "encoder_attention_heads": 6,
22
+ "encoder_ffn_dim": 1536,
23
+ "encoder_layerdrop": 0.0,
24
+ "encoder_layers": 4,
25
+ "eos_token_id": 50257,
26
+ "forced_decoder_ids": null,
27
+ "init_std": 0.02,
28
+ "is_encoder_decoder": true,
29
+ "max_length": 448,
30
+ "max_source_positions": 1500,
31
+ "max_target_positions": 448,
32
+ "model_type": "whisper",
33
+ "num_hidden_layers": 4,
34
+ "num_mel_bins": 80,
35
+ "pad_token_id": 50257,
36
+ "scale_embedding": false,
37
+ "suppress_tokens": [],
38
+ "torch_dtype": "float32",
39
+ "transformers_version": "4.26.0.dev0",
40
+ "use_cache": false,
41
+ "vocab_size": 51865
42
+ }
checkpoint-1000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0de738a16f4eecd6e388ee09cdb50ceca88886aaca4a0641dda753d4b4d0d2da
3
+ size 302181925
checkpoint-1000/preprocessor_config.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af494fa3097ae6332d93cccd2610455ea0a580b0dc345c61ee20f505469c3ea3
3
+ size 151097331
checkpoint-1000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ea848d0d77a0cdd0978f4ed102a0fafa8789dea99f96090a1e465681bbcc5874
3
+ size 14503
checkpoint-1000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:433cf2b714804ef6d53017e239e9089080272abc62f2f115b90d9cab4e56532d
3
+ size 559
checkpoint-1000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d46707535128e9b43d9cf7ea1bf8f7bbc821c42b4fb15b9b3a2e8dbf60058c0f
3
+ size 623
checkpoint-1000/trainer_state.json ADDED
@@ -0,0 +1,265 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 56.48225180255131,
3
+ "best_model_checkpoint": "./whisper-tiny-it-4/checkpoint-1000",
4
+ "epoch": 0.9541984732824428,
5
+ "global_step": 1000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.02,
12
+ "learning_rate": 2.0000000000000003e-06,
13
+ "loss": 3.5722,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 4.5e-06,
19
+ "loss": 2.2272,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.07,
24
+ "learning_rate": 7.000000000000001e-06,
25
+ "loss": 1.1115,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.1,
30
+ "learning_rate": 9.5e-06,
31
+ "loss": 0.8798,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.12,
36
+ "learning_rate": 1.2e-05,
37
+ "loss": 0.7707,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.14,
42
+ "learning_rate": 1.45e-05,
43
+ "loss": 0.754,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.17,
48
+ "learning_rate": 1.7000000000000003e-05,
49
+ "loss": 0.7692,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.19,
54
+ "learning_rate": 1.9500000000000003e-05,
55
+ "loss": 0.7261,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.21,
60
+ "learning_rate": 2.2000000000000003e-05,
61
+ "loss": 0.661,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.24,
66
+ "learning_rate": 2.45e-05,
67
+ "loss": 0.6989,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.26,
72
+ "learning_rate": 2.7000000000000002e-05,
73
+ "loss": 0.6835,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.29,
78
+ "learning_rate": 2.95e-05,
79
+ "loss": 0.6839,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.31,
84
+ "learning_rate": 3.2000000000000005e-05,
85
+ "loss": 0.6934,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.33,
90
+ "learning_rate": 3.45e-05,
91
+ "loss": 0.6656,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.36,
96
+ "learning_rate": 3.7e-05,
97
+ "loss": 0.7053,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.38,
102
+ "learning_rate": 3.9500000000000005e-05,
103
+ "loss": 0.6652,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.41,
108
+ "learning_rate": 4.2e-05,
109
+ "loss": 0.6747,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.43,
114
+ "learning_rate": 4.4500000000000004e-05,
115
+ "loss": 0.6654,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.45,
120
+ "learning_rate": 4.7e-05,
121
+ "loss": 0.6408,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.48,
126
+ "learning_rate": 4.9500000000000004e-05,
127
+ "loss": 0.6722,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.5,
132
+ "learning_rate": 4.971428571428572e-05,
133
+ "loss": 0.6924,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.52,
138
+ "learning_rate": 4.935714285714286e-05,
139
+ "loss": 0.6558,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.55,
144
+ "learning_rate": 4.9e-05,
145
+ "loss": 0.6947,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.57,
150
+ "learning_rate": 4.8642857142857145e-05,
151
+ "loss": 0.6687,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.6,
156
+ "learning_rate": 4.828571428571429e-05,
157
+ "loss": 0.6956,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.62,
162
+ "learning_rate": 4.7928571428571425e-05,
163
+ "loss": 0.6678,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.64,
168
+ "learning_rate": 4.757142857142857e-05,
169
+ "loss": 0.6195,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.67,
174
+ "learning_rate": 4.721428571428572e-05,
175
+ "loss": 0.652,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.69,
180
+ "learning_rate": 4.685714285714286e-05,
181
+ "loss": 0.6484,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.72,
186
+ "learning_rate": 4.6500000000000005e-05,
187
+ "loss": 0.6104,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.74,
192
+ "learning_rate": 4.6142857142857145e-05,
193
+ "loss": 0.6369,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.76,
198
+ "learning_rate": 4.5785714285714285e-05,
199
+ "loss": 0.586,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.79,
204
+ "learning_rate": 4.542857142857143e-05,
205
+ "loss": 0.6088,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.81,
210
+ "learning_rate": 4.507142857142858e-05,
211
+ "loss": 0.6312,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.83,
216
+ "learning_rate": 4.471428571428571e-05,
217
+ "loss": 0.6064,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 0.86,
222
+ "learning_rate": 4.435714285714286e-05,
223
+ "loss": 0.6264,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 0.88,
228
+ "learning_rate": 4.4000000000000006e-05,
229
+ "loss": 0.6792,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 0.91,
234
+ "learning_rate": 4.3642857142857146e-05,
235
+ "loss": 0.5809,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 0.93,
240
+ "learning_rate": 4.328571428571429e-05,
241
+ "loss": 0.6006,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 0.95,
246
+ "learning_rate": 4.292857142857143e-05,
247
+ "loss": 0.5919,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 0.95,
252
+ "eval_loss": 0.804903507232666,
253
+ "eval_runtime": 401.2067,
254
+ "eval_samples_per_second": 3.739,
255
+ "eval_steps_per_second": 0.469,
256
+ "eval_wer": 56.48225180255131,
257
+ "step": 1000
258
+ }
259
+ ],
260
+ "max_steps": 4000,
261
+ "num_train_epochs": 4,
262
+ "total_flos": 3.9390216192e+17,
263
+ "trial_name": null,
264
+ "trial_params": null
265
+ }
checkpoint-1000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ef0b51f0fd5d42552e2301c410bc79d1d62b8b879d179c2d243b38241ad5b64
3
+ size 3503
checkpoint-2000/config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-tiny",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "architectures": [
6
+ "WhisperForConditionalGeneration"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "begin_suppress_tokens": [
10
+ 220,
11
+ 50257
12
+ ],
13
+ "bos_token_id": 50257,
14
+ "d_model": 384,
15
+ "decoder_attention_heads": 6,
16
+ "decoder_ffn_dim": 1536,
17
+ "decoder_layerdrop": 0.0,
18
+ "decoder_layers": 4,
19
+ "decoder_start_token_id": 50258,
20
+ "dropout": 0.0,
21
+ "encoder_attention_heads": 6,
22
+ "encoder_ffn_dim": 1536,
23
+ "encoder_layerdrop": 0.0,
24
+ "encoder_layers": 4,
25
+ "eos_token_id": 50257,
26
+ "forced_decoder_ids": null,
27
+ "init_std": 0.02,
28
+ "is_encoder_decoder": true,
29
+ "max_length": 448,
30
+ "max_source_positions": 1500,
31
+ "max_target_positions": 448,
32
+ "model_type": "whisper",
33
+ "num_hidden_layers": 4,
34
+ "num_mel_bins": 80,
35
+ "pad_token_id": 50257,
36
+ "scale_embedding": false,
37
+ "suppress_tokens": [],
38
+ "torch_dtype": "float32",
39
+ "transformers_version": "4.26.0.dev0",
40
+ "use_cache": false,
41
+ "vocab_size": 51865
42
+ }
checkpoint-2000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:27d451f2b97b84f15a8b3beb574b2647197a425cc08eda7416facc6dba87e792
3
+ size 302181925
checkpoint-2000/preprocessor_config.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-2000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a31e77cf264a0f72d2f699df4cc8702d942571766cff50bbda24fa954abe53cb
3
+ size 151097331
checkpoint-2000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc0c5ebff70a60720012530c91eb8b0af3c25907cbe605e7fb7bbf7b2a359391
3
+ size 14503
checkpoint-2000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6e2335474b493e039967d85052dfaea6b7a1eb56ba93d9511798135a18bc820d
3
+ size 559
checkpoint-2000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89702cc09c7b559b4f030b96ce62cb55cc688040b5e5982ff7a137b2e4c38370
3
+ size 623
checkpoint-2000/trainer_state.json ADDED
@@ -0,0 +1,514 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 44.81419855795895,
3
+ "best_model_checkpoint": "./whisper-tiny-it-4/checkpoint-2000",
4
+ "epoch": 1.9083969465648853,
5
+ "global_step": 2000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.02,
12
+ "learning_rate": 2.0000000000000003e-06,
13
+ "loss": 3.5722,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 4.5e-06,
19
+ "loss": 2.2272,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.07,
24
+ "learning_rate": 7.000000000000001e-06,
25
+ "loss": 1.1115,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.1,
30
+ "learning_rate": 9.5e-06,
31
+ "loss": 0.8798,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.12,
36
+ "learning_rate": 1.2e-05,
37
+ "loss": 0.7707,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.14,
42
+ "learning_rate": 1.45e-05,
43
+ "loss": 0.754,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.17,
48
+ "learning_rate": 1.7000000000000003e-05,
49
+ "loss": 0.7692,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.19,
54
+ "learning_rate": 1.9500000000000003e-05,
55
+ "loss": 0.7261,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.21,
60
+ "learning_rate": 2.2000000000000003e-05,
61
+ "loss": 0.661,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.24,
66
+ "learning_rate": 2.45e-05,
67
+ "loss": 0.6989,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.26,
72
+ "learning_rate": 2.7000000000000002e-05,
73
+ "loss": 0.6835,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.29,
78
+ "learning_rate": 2.95e-05,
79
+ "loss": 0.6839,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.31,
84
+ "learning_rate": 3.2000000000000005e-05,
85
+ "loss": 0.6934,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.33,
90
+ "learning_rate": 3.45e-05,
91
+ "loss": 0.6656,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.36,
96
+ "learning_rate": 3.7e-05,
97
+ "loss": 0.7053,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.38,
102
+ "learning_rate": 3.9500000000000005e-05,
103
+ "loss": 0.6652,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.41,
108
+ "learning_rate": 4.2e-05,
109
+ "loss": 0.6747,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.43,
114
+ "learning_rate": 4.4500000000000004e-05,
115
+ "loss": 0.6654,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.45,
120
+ "learning_rate": 4.7e-05,
121
+ "loss": 0.6408,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.48,
126
+ "learning_rate": 4.9500000000000004e-05,
127
+ "loss": 0.6722,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.5,
132
+ "learning_rate": 4.971428571428572e-05,
133
+ "loss": 0.6924,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.52,
138
+ "learning_rate": 4.935714285714286e-05,
139
+ "loss": 0.6558,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.55,
144
+ "learning_rate": 4.9e-05,
145
+ "loss": 0.6947,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.57,
150
+ "learning_rate": 4.8642857142857145e-05,
151
+ "loss": 0.6687,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.6,
156
+ "learning_rate": 4.828571428571429e-05,
157
+ "loss": 0.6956,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.62,
162
+ "learning_rate": 4.7928571428571425e-05,
163
+ "loss": 0.6678,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.64,
168
+ "learning_rate": 4.757142857142857e-05,
169
+ "loss": 0.6195,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.67,
174
+ "learning_rate": 4.721428571428572e-05,
175
+ "loss": 0.652,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.69,
180
+ "learning_rate": 4.685714285714286e-05,
181
+ "loss": 0.6484,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.72,
186
+ "learning_rate": 4.6500000000000005e-05,
187
+ "loss": 0.6104,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.74,
192
+ "learning_rate": 4.6142857142857145e-05,
193
+ "loss": 0.6369,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.76,
198
+ "learning_rate": 4.5785714285714285e-05,
199
+ "loss": 0.586,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.79,
204
+ "learning_rate": 4.542857142857143e-05,
205
+ "loss": 0.6088,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.81,
210
+ "learning_rate": 4.507142857142858e-05,
211
+ "loss": 0.6312,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.83,
216
+ "learning_rate": 4.471428571428571e-05,
217
+ "loss": 0.6064,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 0.86,
222
+ "learning_rate": 4.435714285714286e-05,
223
+ "loss": 0.6264,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 0.88,
228
+ "learning_rate": 4.4000000000000006e-05,
229
+ "loss": 0.6792,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 0.91,
234
+ "learning_rate": 4.3642857142857146e-05,
235
+ "loss": 0.5809,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 0.93,
240
+ "learning_rate": 4.328571428571429e-05,
241
+ "loss": 0.6006,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 0.95,
246
+ "learning_rate": 4.292857142857143e-05,
247
+ "loss": 0.5919,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 0.95,
252
+ "eval_loss": 0.804903507232666,
253
+ "eval_runtime": 401.2067,
254
+ "eval_samples_per_second": 3.739,
255
+ "eval_steps_per_second": 0.469,
256
+ "eval_wer": 56.48225180255131,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 0.98,
261
+ "learning_rate": 4.257142857142857e-05,
262
+ "loss": 0.6182,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 1.0,
267
+ "learning_rate": 4.221428571428572e-05,
268
+ "loss": 0.5634,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 1.03,
273
+ "learning_rate": 4.185714285714286e-05,
274
+ "loss": 0.3226,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 1.05,
279
+ "learning_rate": 4.15e-05,
280
+ "loss": 0.3305,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 1.07,
285
+ "learning_rate": 4.1142857142857146e-05,
286
+ "loss": 0.3342,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 1.1,
291
+ "learning_rate": 4.0785714285714286e-05,
292
+ "loss": 0.3565,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 1.12,
297
+ "learning_rate": 4.042857142857143e-05,
298
+ "loss": 0.3389,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 1.15,
303
+ "learning_rate": 4.007142857142857e-05,
304
+ "loss": 0.3322,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 1.17,
309
+ "learning_rate": 3.971428571428571e-05,
310
+ "loss": 0.3462,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 1.19,
315
+ "learning_rate": 3.935714285714286e-05,
316
+ "loss": 0.3442,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 1.22,
321
+ "learning_rate": 3.9000000000000006e-05,
322
+ "loss": 0.3433,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 1.24,
327
+ "learning_rate": 3.8642857142857146e-05,
328
+ "loss": 0.3398,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 1.26,
333
+ "learning_rate": 3.8285714285714286e-05,
334
+ "loss": 0.32,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 1.29,
339
+ "learning_rate": 3.792857142857143e-05,
340
+ "loss": 0.3482,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 1.31,
345
+ "learning_rate": 3.757142857142857e-05,
346
+ "loss": 0.3189,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 1.34,
351
+ "learning_rate": 3.721428571428572e-05,
352
+ "loss": 0.3682,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 1.36,
357
+ "learning_rate": 3.685714285714286e-05,
358
+ "loss": 0.3379,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 1.38,
363
+ "learning_rate": 3.65e-05,
364
+ "loss": 0.3116,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 1.41,
369
+ "learning_rate": 3.6142857142857146e-05,
370
+ "loss": 0.354,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 1.43,
375
+ "learning_rate": 3.5785714285714286e-05,
376
+ "loss": 0.3559,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 1.46,
381
+ "learning_rate": 3.5428571428571426e-05,
382
+ "loss": 0.3395,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 1.48,
387
+ "learning_rate": 3.507142857142857e-05,
388
+ "loss": 0.3693,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 1.5,
393
+ "learning_rate": 3.471428571428571e-05,
394
+ "loss": 0.3307,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 1.53,
399
+ "learning_rate": 3.435714285714286e-05,
400
+ "loss": 0.3259,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 1.55,
405
+ "learning_rate": 3.4000000000000007e-05,
406
+ "loss": 0.3576,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 1.57,
411
+ "learning_rate": 3.364285714285714e-05,
412
+ "loss": 0.3083,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 1.6,
417
+ "learning_rate": 3.3285714285714286e-05,
418
+ "loss": 0.3413,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 1.62,
423
+ "learning_rate": 3.292857142857143e-05,
424
+ "loss": 0.3178,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 1.65,
429
+ "learning_rate": 3.257142857142857e-05,
430
+ "loss": 0.3331,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 1.67,
435
+ "learning_rate": 3.221428571428571e-05,
436
+ "loss": 0.3208,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 1.69,
441
+ "learning_rate": 3.185714285714286e-05,
442
+ "loss": 0.3275,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 1.72,
447
+ "learning_rate": 3.15e-05,
448
+ "loss": 0.3368,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 1.74,
453
+ "learning_rate": 3.114285714285715e-05,
454
+ "loss": 0.3335,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 1.77,
459
+ "learning_rate": 3.078571428571429e-05,
460
+ "loss": 0.2934,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 1.79,
465
+ "learning_rate": 3.042857142857143e-05,
466
+ "loss": 0.3081,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 1.81,
471
+ "learning_rate": 3.0071428571428573e-05,
472
+ "loss": 0.3298,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 1.84,
477
+ "learning_rate": 2.9714285714285717e-05,
478
+ "loss": 0.3052,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 1.86,
483
+ "learning_rate": 2.935714285714286e-05,
484
+ "loss": 0.3281,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 1.88,
489
+ "learning_rate": 2.9e-05,
490
+ "loss": 0.3194,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 1.91,
495
+ "learning_rate": 2.8642857142857144e-05,
496
+ "loss": 0.3181,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 1.91,
501
+ "eval_loss": 0.7393131256103516,
502
+ "eval_runtime": 356.4776,
503
+ "eval_samples_per_second": 4.208,
504
+ "eval_steps_per_second": 0.527,
505
+ "eval_wer": 44.81419855795895,
506
+ "step": 2000
507
+ }
508
+ ],
509
+ "max_steps": 4000,
510
+ "num_train_epochs": 4,
511
+ "total_flos": 7.8763199164416e+17,
512
+ "trial_name": null,
513
+ "trial_params": null
514
+ }
checkpoint-2000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ef0b51f0fd5d42552e2301c410bc79d1d62b8b879d179c2d243b38241ad5b64
3
+ size 3503
checkpoint-3000/config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-tiny",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "architectures": [
6
+ "WhisperForConditionalGeneration"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "begin_suppress_tokens": [
10
+ 220,
11
+ 50257
12
+ ],
13
+ "bos_token_id": 50257,
14
+ "d_model": 384,
15
+ "decoder_attention_heads": 6,
16
+ "decoder_ffn_dim": 1536,
17
+ "decoder_layerdrop": 0.0,
18
+ "decoder_layers": 4,
19
+ "decoder_start_token_id": 50258,
20
+ "dropout": 0.0,
21
+ "encoder_attention_heads": 6,
22
+ "encoder_ffn_dim": 1536,
23
+ "encoder_layerdrop": 0.0,
24
+ "encoder_layers": 4,
25
+ "eos_token_id": 50257,
26
+ "forced_decoder_ids": null,
27
+ "init_std": 0.02,
28
+ "is_encoder_decoder": true,
29
+ "max_length": 448,
30
+ "max_source_positions": 1500,
31
+ "max_target_positions": 448,
32
+ "model_type": "whisper",
33
+ "num_hidden_layers": 4,
34
+ "num_mel_bins": 80,
35
+ "pad_token_id": 50257,
36
+ "scale_embedding": false,
37
+ "suppress_tokens": [],
38
+ "torch_dtype": "float32",
39
+ "transformers_version": "4.26.0.dev0",
40
+ "use_cache": false,
41
+ "vocab_size": 51865
42
+ }
checkpoint-3000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b98448bde8ccf15224a8a3d9a1d0a318dae553383c78eed42504c2f8b5293680
3
+ size 302181925
checkpoint-3000/preprocessor_config.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-3000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:958daeb5289c8a0ea5df3873ae943faffb1eb78f233fe9e4f87e15056dfded16
3
+ size 151097331
checkpoint-3000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:859a71d2f0fba9c152e15e56a553ae6dae086eb0de26afe18c0e490f8f93544a
3
+ size 14503
checkpoint-3000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:31db255d829f3c9409ac0db3b6c130bb38ca53b911fdc5c4cd1c34c1fd24b10e
3
+ size 559
checkpoint-3000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:30fbf1ba5d1b834f21dea6db569a8904cfaab1a3670e8dc13daf70f60513afc5
3
+ size 623
checkpoint-3000/trainer_state.json ADDED
@@ -0,0 +1,763 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 42.748197448696615,
3
+ "best_model_checkpoint": "./whisper-tiny-it-4/checkpoint-3000",
4
+ "epoch": 2.8625954198473282,
5
+ "global_step": 3000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.02,
12
+ "learning_rate": 2.0000000000000003e-06,
13
+ "loss": 3.5722,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 4.5e-06,
19
+ "loss": 2.2272,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.07,
24
+ "learning_rate": 7.000000000000001e-06,
25
+ "loss": 1.1115,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.1,
30
+ "learning_rate": 9.5e-06,
31
+ "loss": 0.8798,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.12,
36
+ "learning_rate": 1.2e-05,
37
+ "loss": 0.7707,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.14,
42
+ "learning_rate": 1.45e-05,
43
+ "loss": 0.754,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.17,
48
+ "learning_rate": 1.7000000000000003e-05,
49
+ "loss": 0.7692,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.19,
54
+ "learning_rate": 1.9500000000000003e-05,
55
+ "loss": 0.7261,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.21,
60
+ "learning_rate": 2.2000000000000003e-05,
61
+ "loss": 0.661,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.24,
66
+ "learning_rate": 2.45e-05,
67
+ "loss": 0.6989,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.26,
72
+ "learning_rate": 2.7000000000000002e-05,
73
+ "loss": 0.6835,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.29,
78
+ "learning_rate": 2.95e-05,
79
+ "loss": 0.6839,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.31,
84
+ "learning_rate": 3.2000000000000005e-05,
85
+ "loss": 0.6934,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.33,
90
+ "learning_rate": 3.45e-05,
91
+ "loss": 0.6656,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.36,
96
+ "learning_rate": 3.7e-05,
97
+ "loss": 0.7053,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.38,
102
+ "learning_rate": 3.9500000000000005e-05,
103
+ "loss": 0.6652,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.41,
108
+ "learning_rate": 4.2e-05,
109
+ "loss": 0.6747,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.43,
114
+ "learning_rate": 4.4500000000000004e-05,
115
+ "loss": 0.6654,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.45,
120
+ "learning_rate": 4.7e-05,
121
+ "loss": 0.6408,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.48,
126
+ "learning_rate": 4.9500000000000004e-05,
127
+ "loss": 0.6722,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.5,
132
+ "learning_rate": 4.971428571428572e-05,
133
+ "loss": 0.6924,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.52,
138
+ "learning_rate": 4.935714285714286e-05,
139
+ "loss": 0.6558,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.55,
144
+ "learning_rate": 4.9e-05,
145
+ "loss": 0.6947,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.57,
150
+ "learning_rate": 4.8642857142857145e-05,
151
+ "loss": 0.6687,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.6,
156
+ "learning_rate": 4.828571428571429e-05,
157
+ "loss": 0.6956,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.62,
162
+ "learning_rate": 4.7928571428571425e-05,
163
+ "loss": 0.6678,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.64,
168
+ "learning_rate": 4.757142857142857e-05,
169
+ "loss": 0.6195,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.67,
174
+ "learning_rate": 4.721428571428572e-05,
175
+ "loss": 0.652,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.69,
180
+ "learning_rate": 4.685714285714286e-05,
181
+ "loss": 0.6484,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.72,
186
+ "learning_rate": 4.6500000000000005e-05,
187
+ "loss": 0.6104,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.74,
192
+ "learning_rate": 4.6142857142857145e-05,
193
+ "loss": 0.6369,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.76,
198
+ "learning_rate": 4.5785714285714285e-05,
199
+ "loss": 0.586,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.79,
204
+ "learning_rate": 4.542857142857143e-05,
205
+ "loss": 0.6088,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.81,
210
+ "learning_rate": 4.507142857142858e-05,
211
+ "loss": 0.6312,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.83,
216
+ "learning_rate": 4.471428571428571e-05,
217
+ "loss": 0.6064,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 0.86,
222
+ "learning_rate": 4.435714285714286e-05,
223
+ "loss": 0.6264,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 0.88,
228
+ "learning_rate": 4.4000000000000006e-05,
229
+ "loss": 0.6792,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 0.91,
234
+ "learning_rate": 4.3642857142857146e-05,
235
+ "loss": 0.5809,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 0.93,
240
+ "learning_rate": 4.328571428571429e-05,
241
+ "loss": 0.6006,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 0.95,
246
+ "learning_rate": 4.292857142857143e-05,
247
+ "loss": 0.5919,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 0.95,
252
+ "eval_loss": 0.804903507232666,
253
+ "eval_runtime": 401.2067,
254
+ "eval_samples_per_second": 3.739,
255
+ "eval_steps_per_second": 0.469,
256
+ "eval_wer": 56.48225180255131,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 0.98,
261
+ "learning_rate": 4.257142857142857e-05,
262
+ "loss": 0.6182,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 1.0,
267
+ "learning_rate": 4.221428571428572e-05,
268
+ "loss": 0.5634,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 1.03,
273
+ "learning_rate": 4.185714285714286e-05,
274
+ "loss": 0.3226,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 1.05,
279
+ "learning_rate": 4.15e-05,
280
+ "loss": 0.3305,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 1.07,
285
+ "learning_rate": 4.1142857142857146e-05,
286
+ "loss": 0.3342,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 1.1,
291
+ "learning_rate": 4.0785714285714286e-05,
292
+ "loss": 0.3565,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 1.12,
297
+ "learning_rate": 4.042857142857143e-05,
298
+ "loss": 0.3389,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 1.15,
303
+ "learning_rate": 4.007142857142857e-05,
304
+ "loss": 0.3322,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 1.17,
309
+ "learning_rate": 3.971428571428571e-05,
310
+ "loss": 0.3462,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 1.19,
315
+ "learning_rate": 3.935714285714286e-05,
316
+ "loss": 0.3442,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 1.22,
321
+ "learning_rate": 3.9000000000000006e-05,
322
+ "loss": 0.3433,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 1.24,
327
+ "learning_rate": 3.8642857142857146e-05,
328
+ "loss": 0.3398,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 1.26,
333
+ "learning_rate": 3.8285714285714286e-05,
334
+ "loss": 0.32,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 1.29,
339
+ "learning_rate": 3.792857142857143e-05,
340
+ "loss": 0.3482,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 1.31,
345
+ "learning_rate": 3.757142857142857e-05,
346
+ "loss": 0.3189,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 1.34,
351
+ "learning_rate": 3.721428571428572e-05,
352
+ "loss": 0.3682,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 1.36,
357
+ "learning_rate": 3.685714285714286e-05,
358
+ "loss": 0.3379,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 1.38,
363
+ "learning_rate": 3.65e-05,
364
+ "loss": 0.3116,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 1.41,
369
+ "learning_rate": 3.6142857142857146e-05,
370
+ "loss": 0.354,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 1.43,
375
+ "learning_rate": 3.5785714285714286e-05,
376
+ "loss": 0.3559,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 1.46,
381
+ "learning_rate": 3.5428571428571426e-05,
382
+ "loss": 0.3395,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 1.48,
387
+ "learning_rate": 3.507142857142857e-05,
388
+ "loss": 0.3693,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 1.5,
393
+ "learning_rate": 3.471428571428571e-05,
394
+ "loss": 0.3307,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 1.53,
399
+ "learning_rate": 3.435714285714286e-05,
400
+ "loss": 0.3259,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 1.55,
405
+ "learning_rate": 3.4000000000000007e-05,
406
+ "loss": 0.3576,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 1.57,
411
+ "learning_rate": 3.364285714285714e-05,
412
+ "loss": 0.3083,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 1.6,
417
+ "learning_rate": 3.3285714285714286e-05,
418
+ "loss": 0.3413,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 1.62,
423
+ "learning_rate": 3.292857142857143e-05,
424
+ "loss": 0.3178,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 1.65,
429
+ "learning_rate": 3.257142857142857e-05,
430
+ "loss": 0.3331,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 1.67,
435
+ "learning_rate": 3.221428571428571e-05,
436
+ "loss": 0.3208,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 1.69,
441
+ "learning_rate": 3.185714285714286e-05,
442
+ "loss": 0.3275,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 1.72,
447
+ "learning_rate": 3.15e-05,
448
+ "loss": 0.3368,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 1.74,
453
+ "learning_rate": 3.114285714285715e-05,
454
+ "loss": 0.3335,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 1.77,
459
+ "learning_rate": 3.078571428571429e-05,
460
+ "loss": 0.2934,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 1.79,
465
+ "learning_rate": 3.042857142857143e-05,
466
+ "loss": 0.3081,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 1.81,
471
+ "learning_rate": 3.0071428571428573e-05,
472
+ "loss": 0.3298,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 1.84,
477
+ "learning_rate": 2.9714285714285717e-05,
478
+ "loss": 0.3052,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 1.86,
483
+ "learning_rate": 2.935714285714286e-05,
484
+ "loss": 0.3281,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 1.88,
489
+ "learning_rate": 2.9e-05,
490
+ "loss": 0.3194,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 1.91,
495
+ "learning_rate": 2.8642857142857144e-05,
496
+ "loss": 0.3181,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 1.91,
501
+ "eval_loss": 0.7393131256103516,
502
+ "eval_runtime": 356.4776,
503
+ "eval_samples_per_second": 4.208,
504
+ "eval_steps_per_second": 0.527,
505
+ "eval_wer": 44.81419855795895,
506
+ "step": 2000
507
+ },
508
+ {
509
+ "epoch": 1.93,
510
+ "learning_rate": 2.8285714285714287e-05,
511
+ "loss": 0.3146,
512
+ "step": 2025
513
+ },
514
+ {
515
+ "epoch": 1.96,
516
+ "learning_rate": 2.792857142857143e-05,
517
+ "loss": 0.3081,
518
+ "step": 2050
519
+ },
520
+ {
521
+ "epoch": 1.98,
522
+ "learning_rate": 2.757142857142857e-05,
523
+ "loss": 0.3204,
524
+ "step": 2075
525
+ },
526
+ {
527
+ "epoch": 2.0,
528
+ "learning_rate": 2.7214285714285714e-05,
529
+ "loss": 0.2845,
530
+ "step": 2100
531
+ },
532
+ {
533
+ "epoch": 2.03,
534
+ "learning_rate": 2.6857142857142857e-05,
535
+ "loss": 0.1397,
536
+ "step": 2125
537
+ },
538
+ {
539
+ "epoch": 2.05,
540
+ "learning_rate": 2.6500000000000004e-05,
541
+ "loss": 0.1343,
542
+ "step": 2150
543
+ },
544
+ {
545
+ "epoch": 2.08,
546
+ "learning_rate": 2.6142857142857147e-05,
547
+ "loss": 0.146,
548
+ "step": 2175
549
+ },
550
+ {
551
+ "epoch": 2.1,
552
+ "learning_rate": 2.5785714285714284e-05,
553
+ "loss": 0.1424,
554
+ "step": 2200
555
+ },
556
+ {
557
+ "epoch": 2.12,
558
+ "learning_rate": 2.542857142857143e-05,
559
+ "loss": 0.1407,
560
+ "step": 2225
561
+ },
562
+ {
563
+ "epoch": 2.15,
564
+ "learning_rate": 2.5071428571428574e-05,
565
+ "loss": 0.14,
566
+ "step": 2250
567
+ },
568
+ {
569
+ "epoch": 2.17,
570
+ "learning_rate": 2.4714285714285714e-05,
571
+ "loss": 0.1545,
572
+ "step": 2275
573
+ },
574
+ {
575
+ "epoch": 2.19,
576
+ "learning_rate": 2.4357142857142857e-05,
577
+ "loss": 0.1481,
578
+ "step": 2300
579
+ },
580
+ {
581
+ "epoch": 2.22,
582
+ "learning_rate": 2.4e-05,
583
+ "loss": 0.1475,
584
+ "step": 2325
585
+ },
586
+ {
587
+ "epoch": 2.24,
588
+ "learning_rate": 2.3642857142857144e-05,
589
+ "loss": 0.1641,
590
+ "step": 2350
591
+ },
592
+ {
593
+ "epoch": 2.27,
594
+ "learning_rate": 2.3285714285714287e-05,
595
+ "loss": 0.1418,
596
+ "step": 2375
597
+ },
598
+ {
599
+ "epoch": 2.29,
600
+ "learning_rate": 2.292857142857143e-05,
601
+ "loss": 0.1476,
602
+ "step": 2400
603
+ },
604
+ {
605
+ "epoch": 2.31,
606
+ "learning_rate": 2.257142857142857e-05,
607
+ "loss": 0.1299,
608
+ "step": 2425
609
+ },
610
+ {
611
+ "epoch": 2.34,
612
+ "learning_rate": 2.2214285714285714e-05,
613
+ "loss": 0.1354,
614
+ "step": 2450
615
+ },
616
+ {
617
+ "epoch": 2.36,
618
+ "learning_rate": 2.185714285714286e-05,
619
+ "loss": 0.1379,
620
+ "step": 2475
621
+ },
622
+ {
623
+ "epoch": 2.39,
624
+ "learning_rate": 2.15e-05,
625
+ "loss": 0.1493,
626
+ "step": 2500
627
+ },
628
+ {
629
+ "epoch": 2.41,
630
+ "learning_rate": 2.1142857142857144e-05,
631
+ "loss": 0.1474,
632
+ "step": 2525
633
+ },
634
+ {
635
+ "epoch": 2.43,
636
+ "learning_rate": 2.0785714285714288e-05,
637
+ "loss": 0.147,
638
+ "step": 2550
639
+ },
640
+ {
641
+ "epoch": 2.46,
642
+ "learning_rate": 2.042857142857143e-05,
643
+ "loss": 0.1495,
644
+ "step": 2575
645
+ },
646
+ {
647
+ "epoch": 2.48,
648
+ "learning_rate": 2.007142857142857e-05,
649
+ "loss": 0.1502,
650
+ "step": 2600
651
+ },
652
+ {
653
+ "epoch": 2.5,
654
+ "learning_rate": 1.9714285714285714e-05,
655
+ "loss": 0.147,
656
+ "step": 2625
657
+ },
658
+ {
659
+ "epoch": 2.53,
660
+ "learning_rate": 1.9357142857142858e-05,
661
+ "loss": 0.148,
662
+ "step": 2650
663
+ },
664
+ {
665
+ "epoch": 2.55,
666
+ "learning_rate": 1.9e-05,
667
+ "loss": 0.1493,
668
+ "step": 2675
669
+ },
670
+ {
671
+ "epoch": 2.58,
672
+ "learning_rate": 1.864285714285714e-05,
673
+ "loss": 0.1469,
674
+ "step": 2700
675
+ },
676
+ {
677
+ "epoch": 2.6,
678
+ "learning_rate": 1.8285714285714288e-05,
679
+ "loss": 0.1372,
680
+ "step": 2725
681
+ },
682
+ {
683
+ "epoch": 2.62,
684
+ "learning_rate": 1.792857142857143e-05,
685
+ "loss": 0.1468,
686
+ "step": 2750
687
+ },
688
+ {
689
+ "epoch": 2.65,
690
+ "learning_rate": 1.757142857142857e-05,
691
+ "loss": 0.1418,
692
+ "step": 2775
693
+ },
694
+ {
695
+ "epoch": 2.67,
696
+ "learning_rate": 1.7214285714285715e-05,
697
+ "loss": 0.1389,
698
+ "step": 2800
699
+ },
700
+ {
701
+ "epoch": 2.7,
702
+ "learning_rate": 1.6857142857142858e-05,
703
+ "loss": 0.1347,
704
+ "step": 2825
705
+ },
706
+ {
707
+ "epoch": 2.72,
708
+ "learning_rate": 1.65e-05,
709
+ "loss": 0.1508,
710
+ "step": 2850
711
+ },
712
+ {
713
+ "epoch": 2.74,
714
+ "learning_rate": 1.614285714285714e-05,
715
+ "loss": 0.1347,
716
+ "step": 2875
717
+ },
718
+ {
719
+ "epoch": 2.77,
720
+ "learning_rate": 1.5785714285714288e-05,
721
+ "loss": 0.143,
722
+ "step": 2900
723
+ },
724
+ {
725
+ "epoch": 2.79,
726
+ "learning_rate": 1.5428571428571428e-05,
727
+ "loss": 0.1433,
728
+ "step": 2925
729
+ },
730
+ {
731
+ "epoch": 2.81,
732
+ "learning_rate": 1.5071428571428573e-05,
733
+ "loss": 0.1532,
734
+ "step": 2950
735
+ },
736
+ {
737
+ "epoch": 2.84,
738
+ "learning_rate": 1.4714285714285713e-05,
739
+ "loss": 0.1472,
740
+ "step": 2975
741
+ },
742
+ {
743
+ "epoch": 2.86,
744
+ "learning_rate": 1.4357142857142858e-05,
745
+ "loss": 0.1417,
746
+ "step": 3000
747
+ },
748
+ {
749
+ "epoch": 2.86,
750
+ "eval_loss": 0.7067434787750244,
751
+ "eval_runtime": 360.1108,
752
+ "eval_samples_per_second": 4.165,
753
+ "eval_steps_per_second": 0.522,
754
+ "eval_wer": 42.748197448696615,
755
+ "step": 3000
756
+ }
757
+ ],
758
+ "max_steps": 4000,
759
+ "num_train_epochs": 4,
760
+ "total_flos": 1.18136182136832e+18,
761
+ "trial_name": null,
762
+ "trial_params": null
763
+ }
checkpoint-3000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ef0b51f0fd5d42552e2301c410bc79d1d62b8b879d179c2d243b38241ad5b64
3
+ size 3503
checkpoint-4000/config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-tiny",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "architectures": [
6
+ "WhisperForConditionalGeneration"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "begin_suppress_tokens": [
10
+ 220,
11
+ 50257
12
+ ],
13
+ "bos_token_id": 50257,
14
+ "d_model": 384,
15
+ "decoder_attention_heads": 6,
16
+ "decoder_ffn_dim": 1536,
17
+ "decoder_layerdrop": 0.0,
18
+ "decoder_layers": 4,
19
+ "decoder_start_token_id": 50258,
20
+ "dropout": 0.0,
21
+ "encoder_attention_heads": 6,
22
+ "encoder_ffn_dim": 1536,
23
+ "encoder_layerdrop": 0.0,
24
+ "encoder_layers": 4,
25
+ "eos_token_id": 50257,
26
+ "forced_decoder_ids": null,
27
+ "init_std": 0.02,
28
+ "is_encoder_decoder": true,
29
+ "max_length": 448,
30
+ "max_source_positions": 1500,
31
+ "max_target_positions": 448,
32
+ "model_type": "whisper",
33
+ "num_hidden_layers": 4,
34
+ "num_mel_bins": 80,
35
+ "pad_token_id": 50257,
36
+ "scale_embedding": false,
37
+ "suppress_tokens": [],
38
+ "torch_dtype": "float32",
39
+ "transformers_version": "4.26.0.dev0",
40
+ "use_cache": false,
41
+ "vocab_size": 51865
42
+ }
checkpoint-4000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19e45ed3e290f7c3f12b8a5afba18eeec1d05416abee3e017aa0f20db76efb80
3
+ size 302181925
checkpoint-4000/preprocessor_config.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-4000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28d85eafe3f1ec7d57053a9b3af6f1a6d1e8e13bde7798b57eb98e9a7eeea434
3
+ size 151097331
checkpoint-4000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a308b67252ede9e1e6924b08c67261beff587ff8a8ae70c5987b9f3105827f9e
3
+ size 14503
checkpoint-4000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f5090a9fe1a2692162b6c7a502a38ccad55e430ca4f5f4f67294b8e6a83fa35
3
+ size 559
checkpoint-4000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a6a06b3cd60e0234b8cd1759173a69905545d08398139db446b316b46f96f7be
3
+ size 623
checkpoint-4000/trainer_state.json ADDED
@@ -0,0 +1,1012 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 41.3546866333888,
3
+ "best_model_checkpoint": "./whisper-tiny-it-4/checkpoint-4000",
4
+ "epoch": 3.816793893129771,
5
+ "global_step": 4000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.02,
12
+ "learning_rate": 2.0000000000000003e-06,
13
+ "loss": 3.5722,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 4.5e-06,
19
+ "loss": 2.2272,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.07,
24
+ "learning_rate": 7.000000000000001e-06,
25
+ "loss": 1.1115,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.1,
30
+ "learning_rate": 9.5e-06,
31
+ "loss": 0.8798,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.12,
36
+ "learning_rate": 1.2e-05,
37
+ "loss": 0.7707,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.14,
42
+ "learning_rate": 1.45e-05,
43
+ "loss": 0.754,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.17,
48
+ "learning_rate": 1.7000000000000003e-05,
49
+ "loss": 0.7692,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.19,
54
+ "learning_rate": 1.9500000000000003e-05,
55
+ "loss": 0.7261,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.21,
60
+ "learning_rate": 2.2000000000000003e-05,
61
+ "loss": 0.661,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.24,
66
+ "learning_rate": 2.45e-05,
67
+ "loss": 0.6989,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.26,
72
+ "learning_rate": 2.7000000000000002e-05,
73
+ "loss": 0.6835,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.29,
78
+ "learning_rate": 2.95e-05,
79
+ "loss": 0.6839,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.31,
84
+ "learning_rate": 3.2000000000000005e-05,
85
+ "loss": 0.6934,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.33,
90
+ "learning_rate": 3.45e-05,
91
+ "loss": 0.6656,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.36,
96
+ "learning_rate": 3.7e-05,
97
+ "loss": 0.7053,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.38,
102
+ "learning_rate": 3.9500000000000005e-05,
103
+ "loss": 0.6652,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.41,
108
+ "learning_rate": 4.2e-05,
109
+ "loss": 0.6747,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.43,
114
+ "learning_rate": 4.4500000000000004e-05,
115
+ "loss": 0.6654,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.45,
120
+ "learning_rate": 4.7e-05,
121
+ "loss": 0.6408,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.48,
126
+ "learning_rate": 4.9500000000000004e-05,
127
+ "loss": 0.6722,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.5,
132
+ "learning_rate": 4.971428571428572e-05,
133
+ "loss": 0.6924,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.52,
138
+ "learning_rate": 4.935714285714286e-05,
139
+ "loss": 0.6558,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.55,
144
+ "learning_rate": 4.9e-05,
145
+ "loss": 0.6947,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.57,
150
+ "learning_rate": 4.8642857142857145e-05,
151
+ "loss": 0.6687,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.6,
156
+ "learning_rate": 4.828571428571429e-05,
157
+ "loss": 0.6956,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.62,
162
+ "learning_rate": 4.7928571428571425e-05,
163
+ "loss": 0.6678,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.64,
168
+ "learning_rate": 4.757142857142857e-05,
169
+ "loss": 0.6195,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.67,
174
+ "learning_rate": 4.721428571428572e-05,
175
+ "loss": 0.652,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.69,
180
+ "learning_rate": 4.685714285714286e-05,
181
+ "loss": 0.6484,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.72,
186
+ "learning_rate": 4.6500000000000005e-05,
187
+ "loss": 0.6104,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.74,
192
+ "learning_rate": 4.6142857142857145e-05,
193
+ "loss": 0.6369,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.76,
198
+ "learning_rate": 4.5785714285714285e-05,
199
+ "loss": 0.586,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.79,
204
+ "learning_rate": 4.542857142857143e-05,
205
+ "loss": 0.6088,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.81,
210
+ "learning_rate": 4.507142857142858e-05,
211
+ "loss": 0.6312,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.83,
216
+ "learning_rate": 4.471428571428571e-05,
217
+ "loss": 0.6064,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 0.86,
222
+ "learning_rate": 4.435714285714286e-05,
223
+ "loss": 0.6264,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 0.88,
228
+ "learning_rate": 4.4000000000000006e-05,
229
+ "loss": 0.6792,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 0.91,
234
+ "learning_rate": 4.3642857142857146e-05,
235
+ "loss": 0.5809,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 0.93,
240
+ "learning_rate": 4.328571428571429e-05,
241
+ "loss": 0.6006,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 0.95,
246
+ "learning_rate": 4.292857142857143e-05,
247
+ "loss": 0.5919,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 0.95,
252
+ "eval_loss": 0.804903507232666,
253
+ "eval_runtime": 401.2067,
254
+ "eval_samples_per_second": 3.739,
255
+ "eval_steps_per_second": 0.469,
256
+ "eval_wer": 56.48225180255131,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 0.98,
261
+ "learning_rate": 4.257142857142857e-05,
262
+ "loss": 0.6182,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 1.0,
267
+ "learning_rate": 4.221428571428572e-05,
268
+ "loss": 0.5634,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 1.03,
273
+ "learning_rate": 4.185714285714286e-05,
274
+ "loss": 0.3226,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 1.05,
279
+ "learning_rate": 4.15e-05,
280
+ "loss": 0.3305,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 1.07,
285
+ "learning_rate": 4.1142857142857146e-05,
286
+ "loss": 0.3342,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 1.1,
291
+ "learning_rate": 4.0785714285714286e-05,
292
+ "loss": 0.3565,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 1.12,
297
+ "learning_rate": 4.042857142857143e-05,
298
+ "loss": 0.3389,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 1.15,
303
+ "learning_rate": 4.007142857142857e-05,
304
+ "loss": 0.3322,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 1.17,
309
+ "learning_rate": 3.971428571428571e-05,
310
+ "loss": 0.3462,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 1.19,
315
+ "learning_rate": 3.935714285714286e-05,
316
+ "loss": 0.3442,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 1.22,
321
+ "learning_rate": 3.9000000000000006e-05,
322
+ "loss": 0.3433,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 1.24,
327
+ "learning_rate": 3.8642857142857146e-05,
328
+ "loss": 0.3398,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 1.26,
333
+ "learning_rate": 3.8285714285714286e-05,
334
+ "loss": 0.32,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 1.29,
339
+ "learning_rate": 3.792857142857143e-05,
340
+ "loss": 0.3482,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 1.31,
345
+ "learning_rate": 3.757142857142857e-05,
346
+ "loss": 0.3189,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 1.34,
351
+ "learning_rate": 3.721428571428572e-05,
352
+ "loss": 0.3682,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 1.36,
357
+ "learning_rate": 3.685714285714286e-05,
358
+ "loss": 0.3379,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 1.38,
363
+ "learning_rate": 3.65e-05,
364
+ "loss": 0.3116,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 1.41,
369
+ "learning_rate": 3.6142857142857146e-05,
370
+ "loss": 0.354,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 1.43,
375
+ "learning_rate": 3.5785714285714286e-05,
376
+ "loss": 0.3559,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 1.46,
381
+ "learning_rate": 3.5428571428571426e-05,
382
+ "loss": 0.3395,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 1.48,
387
+ "learning_rate": 3.507142857142857e-05,
388
+ "loss": 0.3693,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 1.5,
393
+ "learning_rate": 3.471428571428571e-05,
394
+ "loss": 0.3307,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 1.53,
399
+ "learning_rate": 3.435714285714286e-05,
400
+ "loss": 0.3259,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 1.55,
405
+ "learning_rate": 3.4000000000000007e-05,
406
+ "loss": 0.3576,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 1.57,
411
+ "learning_rate": 3.364285714285714e-05,
412
+ "loss": 0.3083,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 1.6,
417
+ "learning_rate": 3.3285714285714286e-05,
418
+ "loss": 0.3413,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 1.62,
423
+ "learning_rate": 3.292857142857143e-05,
424
+ "loss": 0.3178,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 1.65,
429
+ "learning_rate": 3.257142857142857e-05,
430
+ "loss": 0.3331,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 1.67,
435
+ "learning_rate": 3.221428571428571e-05,
436
+ "loss": 0.3208,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 1.69,
441
+ "learning_rate": 3.185714285714286e-05,
442
+ "loss": 0.3275,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 1.72,
447
+ "learning_rate": 3.15e-05,
448
+ "loss": 0.3368,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 1.74,
453
+ "learning_rate": 3.114285714285715e-05,
454
+ "loss": 0.3335,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 1.77,
459
+ "learning_rate": 3.078571428571429e-05,
460
+ "loss": 0.2934,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 1.79,
465
+ "learning_rate": 3.042857142857143e-05,
466
+ "loss": 0.3081,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 1.81,
471
+ "learning_rate": 3.0071428571428573e-05,
472
+ "loss": 0.3298,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 1.84,
477
+ "learning_rate": 2.9714285714285717e-05,
478
+ "loss": 0.3052,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 1.86,
483
+ "learning_rate": 2.935714285714286e-05,
484
+ "loss": 0.3281,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 1.88,
489
+ "learning_rate": 2.9e-05,
490
+ "loss": 0.3194,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 1.91,
495
+ "learning_rate": 2.8642857142857144e-05,
496
+ "loss": 0.3181,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 1.91,
501
+ "eval_loss": 0.7393131256103516,
502
+ "eval_runtime": 356.4776,
503
+ "eval_samples_per_second": 4.208,
504
+ "eval_steps_per_second": 0.527,
505
+ "eval_wer": 44.81419855795895,
506
+ "step": 2000
507
+ },
508
+ {
509
+ "epoch": 1.93,
510
+ "learning_rate": 2.8285714285714287e-05,
511
+ "loss": 0.3146,
512
+ "step": 2025
513
+ },
514
+ {
515
+ "epoch": 1.96,
516
+ "learning_rate": 2.792857142857143e-05,
517
+ "loss": 0.3081,
518
+ "step": 2050
519
+ },
520
+ {
521
+ "epoch": 1.98,
522
+ "learning_rate": 2.757142857142857e-05,
523
+ "loss": 0.3204,
524
+ "step": 2075
525
+ },
526
+ {
527
+ "epoch": 2.0,
528
+ "learning_rate": 2.7214285714285714e-05,
529
+ "loss": 0.2845,
530
+ "step": 2100
531
+ },
532
+ {
533
+ "epoch": 2.03,
534
+ "learning_rate": 2.6857142857142857e-05,
535
+ "loss": 0.1397,
536
+ "step": 2125
537
+ },
538
+ {
539
+ "epoch": 2.05,
540
+ "learning_rate": 2.6500000000000004e-05,
541
+ "loss": 0.1343,
542
+ "step": 2150
543
+ },
544
+ {
545
+ "epoch": 2.08,
546
+ "learning_rate": 2.6142857142857147e-05,
547
+ "loss": 0.146,
548
+ "step": 2175
549
+ },
550
+ {
551
+ "epoch": 2.1,
552
+ "learning_rate": 2.5785714285714284e-05,
553
+ "loss": 0.1424,
554
+ "step": 2200
555
+ },
556
+ {
557
+ "epoch": 2.12,
558
+ "learning_rate": 2.542857142857143e-05,
559
+ "loss": 0.1407,
560
+ "step": 2225
561
+ },
562
+ {
563
+ "epoch": 2.15,
564
+ "learning_rate": 2.5071428571428574e-05,
565
+ "loss": 0.14,
566
+ "step": 2250
567
+ },
568
+ {
569
+ "epoch": 2.17,
570
+ "learning_rate": 2.4714285714285714e-05,
571
+ "loss": 0.1545,
572
+ "step": 2275
573
+ },
574
+ {
575
+ "epoch": 2.19,
576
+ "learning_rate": 2.4357142857142857e-05,
577
+ "loss": 0.1481,
578
+ "step": 2300
579
+ },
580
+ {
581
+ "epoch": 2.22,
582
+ "learning_rate": 2.4e-05,
583
+ "loss": 0.1475,
584
+ "step": 2325
585
+ },
586
+ {
587
+ "epoch": 2.24,
588
+ "learning_rate": 2.3642857142857144e-05,
589
+ "loss": 0.1641,
590
+ "step": 2350
591
+ },
592
+ {
593
+ "epoch": 2.27,
594
+ "learning_rate": 2.3285714285714287e-05,
595
+ "loss": 0.1418,
596
+ "step": 2375
597
+ },
598
+ {
599
+ "epoch": 2.29,
600
+ "learning_rate": 2.292857142857143e-05,
601
+ "loss": 0.1476,
602
+ "step": 2400
603
+ },
604
+ {
605
+ "epoch": 2.31,
606
+ "learning_rate": 2.257142857142857e-05,
607
+ "loss": 0.1299,
608
+ "step": 2425
609
+ },
610
+ {
611
+ "epoch": 2.34,
612
+ "learning_rate": 2.2214285714285714e-05,
613
+ "loss": 0.1354,
614
+ "step": 2450
615
+ },
616
+ {
617
+ "epoch": 2.36,
618
+ "learning_rate": 2.185714285714286e-05,
619
+ "loss": 0.1379,
620
+ "step": 2475
621
+ },
622
+ {
623
+ "epoch": 2.39,
624
+ "learning_rate": 2.15e-05,
625
+ "loss": 0.1493,
626
+ "step": 2500
627
+ },
628
+ {
629
+ "epoch": 2.41,
630
+ "learning_rate": 2.1142857142857144e-05,
631
+ "loss": 0.1474,
632
+ "step": 2525
633
+ },
634
+ {
635
+ "epoch": 2.43,
636
+ "learning_rate": 2.0785714285714288e-05,
637
+ "loss": 0.147,
638
+ "step": 2550
639
+ },
640
+ {
641
+ "epoch": 2.46,
642
+ "learning_rate": 2.042857142857143e-05,
643
+ "loss": 0.1495,
644
+ "step": 2575
645
+ },
646
+ {
647
+ "epoch": 2.48,
648
+ "learning_rate": 2.007142857142857e-05,
649
+ "loss": 0.1502,
650
+ "step": 2600
651
+ },
652
+ {
653
+ "epoch": 2.5,
654
+ "learning_rate": 1.9714285714285714e-05,
655
+ "loss": 0.147,
656
+ "step": 2625
657
+ },
658
+ {
659
+ "epoch": 2.53,
660
+ "learning_rate": 1.9357142857142858e-05,
661
+ "loss": 0.148,
662
+ "step": 2650
663
+ },
664
+ {
665
+ "epoch": 2.55,
666
+ "learning_rate": 1.9e-05,
667
+ "loss": 0.1493,
668
+ "step": 2675
669
+ },
670
+ {
671
+ "epoch": 2.58,
672
+ "learning_rate": 1.864285714285714e-05,
673
+ "loss": 0.1469,
674
+ "step": 2700
675
+ },
676
+ {
677
+ "epoch": 2.6,
678
+ "learning_rate": 1.8285714285714288e-05,
679
+ "loss": 0.1372,
680
+ "step": 2725
681
+ },
682
+ {
683
+ "epoch": 2.62,
684
+ "learning_rate": 1.792857142857143e-05,
685
+ "loss": 0.1468,
686
+ "step": 2750
687
+ },
688
+ {
689
+ "epoch": 2.65,
690
+ "learning_rate": 1.757142857142857e-05,
691
+ "loss": 0.1418,
692
+ "step": 2775
693
+ },
694
+ {
695
+ "epoch": 2.67,
696
+ "learning_rate": 1.7214285714285715e-05,
697
+ "loss": 0.1389,
698
+ "step": 2800
699
+ },
700
+ {
701
+ "epoch": 2.7,
702
+ "learning_rate": 1.6857142857142858e-05,
703
+ "loss": 0.1347,
704
+ "step": 2825
705
+ },
706
+ {
707
+ "epoch": 2.72,
708
+ "learning_rate": 1.65e-05,
709
+ "loss": 0.1508,
710
+ "step": 2850
711
+ },
712
+ {
713
+ "epoch": 2.74,
714
+ "learning_rate": 1.614285714285714e-05,
715
+ "loss": 0.1347,
716
+ "step": 2875
717
+ },
718
+ {
719
+ "epoch": 2.77,
720
+ "learning_rate": 1.5785714285714288e-05,
721
+ "loss": 0.143,
722
+ "step": 2900
723
+ },
724
+ {
725
+ "epoch": 2.79,
726
+ "learning_rate": 1.5428571428571428e-05,
727
+ "loss": 0.1433,
728
+ "step": 2925
729
+ },
730
+ {
731
+ "epoch": 2.81,
732
+ "learning_rate": 1.5071428571428573e-05,
733
+ "loss": 0.1532,
734
+ "step": 2950
735
+ },
736
+ {
737
+ "epoch": 2.84,
738
+ "learning_rate": 1.4714285714285713e-05,
739
+ "loss": 0.1472,
740
+ "step": 2975
741
+ },
742
+ {
743
+ "epoch": 2.86,
744
+ "learning_rate": 1.4357142857142858e-05,
745
+ "loss": 0.1417,
746
+ "step": 3000
747
+ },
748
+ {
749
+ "epoch": 2.86,
750
+ "eval_loss": 0.7067434787750244,
751
+ "eval_runtime": 360.1108,
752
+ "eval_samples_per_second": 4.165,
753
+ "eval_steps_per_second": 0.522,
754
+ "eval_wer": 42.748197448696615,
755
+ "step": 3000
756
+ },
757
+ {
758
+ "epoch": 2.89,
759
+ "learning_rate": 1.4000000000000001e-05,
760
+ "loss": 0.1477,
761
+ "step": 3025
762
+ },
763
+ {
764
+ "epoch": 2.91,
765
+ "learning_rate": 1.3642857142857143e-05,
766
+ "loss": 0.1466,
767
+ "step": 3050
768
+ },
769
+ {
770
+ "epoch": 2.93,
771
+ "learning_rate": 1.3285714285714288e-05,
772
+ "loss": 0.1388,
773
+ "step": 3075
774
+ },
775
+ {
776
+ "epoch": 2.96,
777
+ "learning_rate": 1.2928571428571428e-05,
778
+ "loss": 0.1417,
779
+ "step": 3100
780
+ },
781
+ {
782
+ "epoch": 2.98,
783
+ "learning_rate": 1.2571428571428573e-05,
784
+ "loss": 0.1316,
785
+ "step": 3125
786
+ },
787
+ {
788
+ "epoch": 3.01,
789
+ "learning_rate": 1.2214285714285715e-05,
790
+ "loss": 0.1185,
791
+ "step": 3150
792
+ },
793
+ {
794
+ "epoch": 3.03,
795
+ "learning_rate": 1.1857142857142858e-05,
796
+ "loss": 0.061,
797
+ "step": 3175
798
+ },
799
+ {
800
+ "epoch": 3.05,
801
+ "learning_rate": 1.1500000000000002e-05,
802
+ "loss": 0.0632,
803
+ "step": 3200
804
+ },
805
+ {
806
+ "epoch": 3.08,
807
+ "learning_rate": 1.1142857142857143e-05,
808
+ "loss": 0.0589,
809
+ "step": 3225
810
+ },
811
+ {
812
+ "epoch": 3.1,
813
+ "learning_rate": 1.0785714285714287e-05,
814
+ "loss": 0.0645,
815
+ "step": 3250
816
+ },
817
+ {
818
+ "epoch": 3.12,
819
+ "learning_rate": 1.0428571428571428e-05,
820
+ "loss": 0.0623,
821
+ "step": 3275
822
+ },
823
+ {
824
+ "epoch": 3.15,
825
+ "learning_rate": 1.0071428571428572e-05,
826
+ "loss": 0.064,
827
+ "step": 3300
828
+ },
829
+ {
830
+ "epoch": 3.17,
831
+ "learning_rate": 9.714285714285715e-06,
832
+ "loss": 0.0623,
833
+ "step": 3325
834
+ },
835
+ {
836
+ "epoch": 3.2,
837
+ "learning_rate": 9.357142857142857e-06,
838
+ "loss": 0.0576,
839
+ "step": 3350
840
+ },
841
+ {
842
+ "epoch": 3.22,
843
+ "learning_rate": 9e-06,
844
+ "loss": 0.0654,
845
+ "step": 3375
846
+ },
847
+ {
848
+ "epoch": 3.24,
849
+ "learning_rate": 8.642857142857144e-06,
850
+ "loss": 0.0652,
851
+ "step": 3400
852
+ },
853
+ {
854
+ "epoch": 3.27,
855
+ "learning_rate": 8.285714285714285e-06,
856
+ "loss": 0.0575,
857
+ "step": 3425
858
+ },
859
+ {
860
+ "epoch": 3.29,
861
+ "learning_rate": 7.928571428571429e-06,
862
+ "loss": 0.062,
863
+ "step": 3450
864
+ },
865
+ {
866
+ "epoch": 3.32,
867
+ "learning_rate": 7.571428571428572e-06,
868
+ "loss": 0.0594,
869
+ "step": 3475
870
+ },
871
+ {
872
+ "epoch": 3.34,
873
+ "learning_rate": 7.214285714285715e-06,
874
+ "loss": 0.063,
875
+ "step": 3500
876
+ },
877
+ {
878
+ "epoch": 3.36,
879
+ "learning_rate": 6.857142857142858e-06,
880
+ "loss": 0.0638,
881
+ "step": 3525
882
+ },
883
+ {
884
+ "epoch": 3.39,
885
+ "learning_rate": 6.5000000000000004e-06,
886
+ "loss": 0.0612,
887
+ "step": 3550
888
+ },
889
+ {
890
+ "epoch": 3.41,
891
+ "learning_rate": 6.142857142857143e-06,
892
+ "loss": 0.0644,
893
+ "step": 3575
894
+ },
895
+ {
896
+ "epoch": 3.44,
897
+ "learning_rate": 5.785714285714286e-06,
898
+ "loss": 0.0592,
899
+ "step": 3600
900
+ },
901
+ {
902
+ "epoch": 3.46,
903
+ "learning_rate": 5.428571428571429e-06,
904
+ "loss": 0.06,
905
+ "step": 3625
906
+ },
907
+ {
908
+ "epoch": 3.48,
909
+ "learning_rate": 5.071428571428571e-06,
910
+ "loss": 0.0552,
911
+ "step": 3650
912
+ },
913
+ {
914
+ "epoch": 3.51,
915
+ "learning_rate": 4.714285714285715e-06,
916
+ "loss": 0.0607,
917
+ "step": 3675
918
+ },
919
+ {
920
+ "epoch": 3.53,
921
+ "learning_rate": 4.357142857142857e-06,
922
+ "loss": 0.0609,
923
+ "step": 3700
924
+ },
925
+ {
926
+ "epoch": 3.55,
927
+ "learning_rate": 4.000000000000001e-06,
928
+ "loss": 0.0638,
929
+ "step": 3725
930
+ },
931
+ {
932
+ "epoch": 3.58,
933
+ "learning_rate": 3.642857142857143e-06,
934
+ "loss": 0.0594,
935
+ "step": 3750
936
+ },
937
+ {
938
+ "epoch": 3.6,
939
+ "learning_rate": 3.2857142857142857e-06,
940
+ "loss": 0.0666,
941
+ "step": 3775
942
+ },
943
+ {
944
+ "epoch": 3.63,
945
+ "learning_rate": 2.9285714285714287e-06,
946
+ "loss": 0.0539,
947
+ "step": 3800
948
+ },
949
+ {
950
+ "epoch": 3.65,
951
+ "learning_rate": 2.5714285714285716e-06,
952
+ "loss": 0.0612,
953
+ "step": 3825
954
+ },
955
+ {
956
+ "epoch": 3.67,
957
+ "learning_rate": 2.214285714285714e-06,
958
+ "loss": 0.0581,
959
+ "step": 3850
960
+ },
961
+ {
962
+ "epoch": 3.7,
963
+ "learning_rate": 1.8571428571428573e-06,
964
+ "loss": 0.0649,
965
+ "step": 3875
966
+ },
967
+ {
968
+ "epoch": 3.72,
969
+ "learning_rate": 1.5e-06,
970
+ "loss": 0.0571,
971
+ "step": 3900
972
+ },
973
+ {
974
+ "epoch": 3.75,
975
+ "learning_rate": 1.142857142857143e-06,
976
+ "loss": 0.0582,
977
+ "step": 3925
978
+ },
979
+ {
980
+ "epoch": 3.77,
981
+ "learning_rate": 7.857142857142858e-07,
982
+ "loss": 0.0597,
983
+ "step": 3950
984
+ },
985
+ {
986
+ "epoch": 3.79,
987
+ "learning_rate": 4.285714285714286e-07,
988
+ "loss": 0.0547,
989
+ "step": 3975
990
+ },
991
+ {
992
+ "epoch": 3.82,
993
+ "learning_rate": 7.142857142857144e-08,
994
+ "loss": 0.0627,
995
+ "step": 4000
996
+ },
997
+ {
998
+ "epoch": 3.82,
999
+ "eval_loss": 0.7125627994537354,
1000
+ "eval_runtime": 353.3354,
1001
+ "eval_samples_per_second": 4.245,
1002
+ "eval_steps_per_second": 0.532,
1003
+ "eval_wer": 41.3546866333888,
1004
+ "step": 4000
1005
+ }
1006
+ ],
1007
+ "max_steps": 4000,
1008
+ "num_train_epochs": 4,
1009
+ "total_flos": 1.57509165109248e+18,
1010
+ "trial_name": null,
1011
+ "trial_params": null
1012
+ }
checkpoint-4000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ef0b51f0fd5d42552e2301c410bc79d1d62b8b879d179c2d243b38241ad5b64
3
+ size 3503