lucio commited on
Commit
c13cb97
1 Parent(s): e0e16ab

init model - 20 epochs training on validation

Browse files
config.json ADDED
@@ -0,0 +1,76 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
3
+ "activation_dropout": 0.05,
4
+ "apply_spec_augment": true,
5
+ "architectures": [
6
+ "Wav2Vec2ForCTC"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "bos_token_id": 1,
10
+ "conv_bias": true,
11
+ "conv_dim": [
12
+ 512,
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512
19
+ ],
20
+ "conv_kernel": [
21
+ 10,
22
+ 3,
23
+ 3,
24
+ 3,
25
+ 3,
26
+ 2,
27
+ 2
28
+ ],
29
+ "conv_stride": [
30
+ 5,
31
+ 2,
32
+ 2,
33
+ 2,
34
+ 2,
35
+ 2,
36
+ 2
37
+ ],
38
+ "ctc_loss_reduction": "mean",
39
+ "ctc_zero_infinity": false,
40
+ "do_stable_layer_norm": true,
41
+ "eos_token_id": 2,
42
+ "feat_extract_activation": "gelu",
43
+ "feat_extract_dropout": 0.0,
44
+ "feat_extract_norm": "layer",
45
+ "feat_proj_dropout": 0.05,
46
+ "final_dropout": 0.0,
47
+ "gradient_checkpointing": true,
48
+ "hidden_act": "gelu",
49
+ "hidden_dropout": 0.05,
50
+ "hidden_size": 1024,
51
+ "initializer_range": 0.02,
52
+ "intermediate_size": 4096,
53
+ "layer_norm_eps": 1e-05,
54
+ "layerdrop": 0.05,
55
+ "mask_channel_length": 10,
56
+ "mask_channel_min_space": 1,
57
+ "mask_channel_other": 0.0,
58
+ "mask_channel_prob": 0.0,
59
+ "mask_channel_selection": "static",
60
+ "mask_feature_length": 10,
61
+ "mask_feature_prob": 0.0,
62
+ "mask_time_length": 10,
63
+ "mask_time_min_space": 1,
64
+ "mask_time_other": 0.0,
65
+ "mask_time_prob": 0.05,
66
+ "mask_time_selection": "static",
67
+ "model_type": "wav2vec2",
68
+ "num_attention_heads": 16,
69
+ "num_conv_pos_embedding_groups": 16,
70
+ "num_conv_pos_embeddings": 128,
71
+ "num_feat_extract_layers": 7,
72
+ "num_hidden_layers": 24,
73
+ "pad_token_id": 62,
74
+ "transformers_version": "4.5.0.dev0",
75
+ "vocab_size": 63
76
+ }
preprocessor_config.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_size": 1,
4
+ "padding_side": "right",
5
+ "padding_value": 0.0,
6
+ "return_attention_mask": true,
7
+ "sampling_rate": 16000
8
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:91c5d0f15305b61780b33114710f3a1f57b704f35927d16f80e9900a148d51bb
3
+ size 1262192151
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]"}
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"unk_token": "[UNK]", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "[PAD]", "do_lower_case": false, "word_delimiter_token": "|"}
trainer_state.json ADDED
@@ -0,0 +1,2067 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 20.0,
5
+ "global_step": 9400,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.14,
12
+ "learning_rate": 2.5e-05,
13
+ "loss": 15.9584,
14
+ "step": 64
15
+ },
16
+ {
17
+ "epoch": 0.14,
18
+ "eval_loss": 14.658380508422852,
19
+ "eval_runtime": 249.5052,
20
+ "eval_samples_per_second": 8.208,
21
+ "eval_wer": 1.0,
22
+ "step": 64
23
+ },
24
+ {
25
+ "epoch": 0.27,
26
+ "learning_rate": 5e-05,
27
+ "loss": 6.3907,
28
+ "step": 128
29
+ },
30
+ {
31
+ "epoch": 0.27,
32
+ "eval_loss": 4.082584857940674,
33
+ "eval_runtime": 247.9481,
34
+ "eval_samples_per_second": 8.26,
35
+ "eval_wer": 1.0,
36
+ "step": 128
37
+ },
38
+ {
39
+ "epoch": 0.41,
40
+ "learning_rate": 7.500000000000001e-05,
41
+ "loss": 3.4412,
42
+ "step": 192
43
+ },
44
+ {
45
+ "epoch": 0.41,
46
+ "eval_loss": 3.1572635173797607,
47
+ "eval_runtime": 251.7843,
48
+ "eval_samples_per_second": 8.134,
49
+ "eval_wer": 1.0,
50
+ "step": 192
51
+ },
52
+ {
53
+ "epoch": 0.54,
54
+ "learning_rate": 0.0001,
55
+ "loss": 3.005,
56
+ "step": 256
57
+ },
58
+ {
59
+ "epoch": 0.54,
60
+ "eval_loss": 3.0201895236968994,
61
+ "eval_runtime": 248.8764,
62
+ "eval_samples_per_second": 8.229,
63
+ "eval_wer": 1.0,
64
+ "step": 256
65
+ },
66
+ {
67
+ "epoch": 0.68,
68
+ "learning_rate": 0.000125,
69
+ "loss": 2.9507,
70
+ "step": 320
71
+ },
72
+ {
73
+ "epoch": 0.68,
74
+ "eval_loss": 3.0097341537475586,
75
+ "eval_runtime": 250.2451,
76
+ "eval_samples_per_second": 8.184,
77
+ "eval_wer": 1.0,
78
+ "step": 320
79
+ },
80
+ {
81
+ "epoch": 0.82,
82
+ "learning_rate": 0.00015000000000000001,
83
+ "loss": 2.9285,
84
+ "step": 384
85
+ },
86
+ {
87
+ "epoch": 0.82,
88
+ "eval_loss": 3.0145537853240967,
89
+ "eval_runtime": 249.391,
90
+ "eval_samples_per_second": 8.212,
91
+ "eval_wer": 1.0,
92
+ "step": 384
93
+ },
94
+ {
95
+ "epoch": 0.95,
96
+ "learning_rate": 0.000175,
97
+ "loss": 2.9081,
98
+ "step": 448
99
+ },
100
+ {
101
+ "epoch": 0.95,
102
+ "eval_loss": 2.889194965362549,
103
+ "eval_runtime": 252.4706,
104
+ "eval_samples_per_second": 8.112,
105
+ "eval_wer": 1.0,
106
+ "step": 448
107
+ },
108
+ {
109
+ "epoch": 1.09,
110
+ "learning_rate": 0.0002,
111
+ "loss": 2.8141,
112
+ "step": 512
113
+ },
114
+ {
115
+ "epoch": 1.09,
116
+ "eval_loss": 3.1128151416778564,
117
+ "eval_runtime": 249.3837,
118
+ "eval_samples_per_second": 8.212,
119
+ "eval_wer": 0.9997997196074504,
120
+ "step": 512
121
+ },
122
+ {
123
+ "epoch": 1.23,
124
+ "learning_rate": 0.00019855985598559857,
125
+ "loss": 1.9764,
126
+ "step": 576
127
+ },
128
+ {
129
+ "epoch": 1.23,
130
+ "eval_loss": 1.4347355365753174,
131
+ "eval_runtime": 253.6868,
132
+ "eval_samples_per_second": 8.073,
133
+ "eval_wer": 0.9202884037652714,
134
+ "step": 576
135
+ },
136
+ {
137
+ "epoch": 1.36,
138
+ "learning_rate": 0.0001971197119711971,
139
+ "loss": 1.171,
140
+ "step": 640
141
+ },
142
+ {
143
+ "epoch": 1.36,
144
+ "eval_loss": 1.1342754364013672,
145
+ "eval_runtime": 252.7977,
146
+ "eval_samples_per_second": 8.101,
147
+ "eval_wer": 0.8585352827291541,
148
+ "step": 640
149
+ },
150
+ {
151
+ "epoch": 1.5,
152
+ "learning_rate": 0.0001956795679567957,
153
+ "loss": 0.9986,
154
+ "step": 704
155
+ },
156
+ {
157
+ "epoch": 1.5,
158
+ "eval_loss": 1.0098124742507935,
159
+ "eval_runtime": 254.0896,
160
+ "eval_samples_per_second": 8.06,
161
+ "eval_wer": 0.8405768075305428,
162
+ "step": 704
163
+ },
164
+ {
165
+ "epoch": 1.63,
166
+ "learning_rate": 0.00019423942394239426,
167
+ "loss": 0.8642,
168
+ "step": 768
169
+ },
170
+ {
171
+ "epoch": 1.63,
172
+ "eval_loss": 0.9640593528747559,
173
+ "eval_runtime": 254.2048,
174
+ "eval_samples_per_second": 8.056,
175
+ "eval_wer": 0.8503905467654717,
176
+ "step": 768
177
+ },
178
+ {
179
+ "epoch": 1.77,
180
+ "learning_rate": 0.0001927992799279928,
181
+ "loss": 0.7951,
182
+ "step": 832
183
+ },
184
+ {
185
+ "epoch": 1.77,
186
+ "eval_loss": 0.8799329996109009,
187
+ "eval_runtime": 255.0842,
188
+ "eval_samples_per_second": 8.029,
189
+ "eval_wer": 0.787903064290006,
190
+ "step": 832
191
+ },
192
+ {
193
+ "epoch": 1.91,
194
+ "learning_rate": 0.00019135913591359136,
195
+ "loss": 0.7317,
196
+ "step": 896
197
+ },
198
+ {
199
+ "epoch": 1.91,
200
+ "eval_loss": 0.8695712089538574,
201
+ "eval_runtime": 253.1871,
202
+ "eval_samples_per_second": 8.089,
203
+ "eval_wer": 0.7686093864743975,
204
+ "step": 896
205
+ },
206
+ {
207
+ "epoch": 2.04,
208
+ "learning_rate": 0.00018991899189918995,
209
+ "loss": 0.6835,
210
+ "step": 960
211
+ },
212
+ {
213
+ "epoch": 2.04,
214
+ "eval_loss": 0.8140479326248169,
215
+ "eval_runtime": 253.7263,
216
+ "eval_samples_per_second": 8.072,
217
+ "eval_wer": 0.7338941184324721,
218
+ "step": 960
219
+ },
220
+ {
221
+ "epoch": 2.18,
222
+ "learning_rate": 0.00018847884788478848,
223
+ "loss": 0.6188,
224
+ "step": 1024
225
+ },
226
+ {
227
+ "epoch": 2.18,
228
+ "eval_loss": 0.8309915065765381,
229
+ "eval_runtime": 254.0037,
230
+ "eval_samples_per_second": 8.063,
231
+ "eval_wer": 0.7473129047332933,
232
+ "step": 1024
233
+ },
234
+ {
235
+ "epoch": 2.31,
236
+ "learning_rate": 0.00018703870387038705,
237
+ "loss": 0.6116,
238
+ "step": 1088
239
+ },
240
+ {
241
+ "epoch": 2.31,
242
+ "eval_loss": 0.8404783010482788,
243
+ "eval_runtime": 255.6119,
244
+ "eval_samples_per_second": 8.012,
245
+ "eval_wer": 0.7376326857600641,
246
+ "step": 1088
247
+ },
248
+ {
249
+ "epoch": 2.45,
250
+ "learning_rate": 0.0001855985598559856,
251
+ "loss": 0.5654,
252
+ "step": 1152
253
+ },
254
+ {
255
+ "epoch": 2.45,
256
+ "eval_loss": 0.7973554134368896,
257
+ "eval_runtime": 253.93,
258
+ "eval_samples_per_second": 8.065,
259
+ "eval_wer": 0.7304893517591294,
260
+ "step": 1152
261
+ },
262
+ {
263
+ "epoch": 2.59,
264
+ "learning_rate": 0.00018415841584158417,
265
+ "loss": 0.5697,
266
+ "step": 1216
267
+ },
268
+ {
269
+ "epoch": 2.59,
270
+ "eval_loss": 0.7916511297225952,
271
+ "eval_runtime": 255.1499,
272
+ "eval_samples_per_second": 8.027,
273
+ "eval_wer": 0.7420388543961546,
274
+ "step": 1216
275
+ },
276
+ {
277
+ "epoch": 2.72,
278
+ "learning_rate": 0.00018271827182718273,
279
+ "loss": 0.5621,
280
+ "step": 1280
281
+ },
282
+ {
283
+ "epoch": 2.72,
284
+ "eval_loss": 0.773869514465332,
285
+ "eval_runtime": 253.9689,
286
+ "eval_samples_per_second": 8.064,
287
+ "eval_wer": 0.7266840243006877,
288
+ "step": 1280
289
+ },
290
+ {
291
+ "epoch": 2.86,
292
+ "learning_rate": 0.0001812781278127813,
293
+ "loss": 0.5646,
294
+ "step": 1344
295
+ },
296
+ {
297
+ "epoch": 2.86,
298
+ "eval_loss": 0.7937338352203369,
299
+ "eval_runtime": 254.2977,
300
+ "eval_samples_per_second": 8.054,
301
+ "eval_wer": 0.70698978569998,
302
+ "step": 1344
303
+ },
304
+ {
305
+ "epoch": 3.0,
306
+ "learning_rate": 0.00017983798379837983,
307
+ "loss": 0.5266,
308
+ "step": 1408
309
+ },
310
+ {
311
+ "epoch": 3.0,
312
+ "eval_loss": 0.738484263420105,
313
+ "eval_runtime": 251.7487,
314
+ "eval_samples_per_second": 8.135,
315
+ "eval_wer": 0.6933707190066093,
316
+ "step": 1408
317
+ },
318
+ {
319
+ "epoch": 3.13,
320
+ "learning_rate": 0.00017839783978397842,
321
+ "loss": 0.4719,
322
+ "step": 1472
323
+ },
324
+ {
325
+ "epoch": 3.13,
326
+ "eval_loss": 0.7587714195251465,
327
+ "eval_runtime": 253.339,
328
+ "eval_samples_per_second": 8.084,
329
+ "eval_wer": 0.6889645503705187,
330
+ "step": 1472
331
+ },
332
+ {
333
+ "epoch": 3.27,
334
+ "learning_rate": 0.00017695769576957698,
335
+ "loss": 0.4479,
336
+ "step": 1536
337
+ },
338
+ {
339
+ "epoch": 3.27,
340
+ "eval_loss": 0.7555954456329346,
341
+ "eval_runtime": 252.6303,
342
+ "eval_samples_per_second": 8.107,
343
+ "eval_wer": 0.6807530542759864,
344
+ "step": 1536
345
+ },
346
+ {
347
+ "epoch": 3.4,
348
+ "learning_rate": 0.00017551755175517552,
349
+ "loss": 0.4614,
350
+ "step": 1600
351
+ },
352
+ {
353
+ "epoch": 3.4,
354
+ "eval_loss": 0.7171066999435425,
355
+ "eval_runtime": 252.5327,
356
+ "eval_samples_per_second": 8.11,
357
+ "eval_wer": 0.6792843313972895,
358
+ "step": 1600
359
+ },
360
+ {
361
+ "epoch": 3.54,
362
+ "learning_rate": 0.00017407740774077408,
363
+ "loss": 0.4491,
364
+ "step": 1664
365
+ },
366
+ {
367
+ "epoch": 3.54,
368
+ "eval_loss": 0.7290804386138916,
369
+ "eval_runtime": 251.92,
370
+ "eval_samples_per_second": 8.13,
371
+ "eval_wer": 0.6693370719006609,
372
+ "step": 1664
373
+ },
374
+ {
375
+ "epoch": 3.68,
376
+ "learning_rate": 0.00017263726372637264,
377
+ "loss": 0.4355,
378
+ "step": 1728
379
+ },
380
+ {
381
+ "epoch": 3.68,
382
+ "eval_loss": 0.7258065938949585,
383
+ "eval_runtime": 251.8895,
384
+ "eval_samples_per_second": 8.131,
385
+ "eval_wer": 0.6802857333600374,
386
+ "step": 1728
387
+ },
388
+ {
389
+ "epoch": 3.81,
390
+ "learning_rate": 0.0001711971197119712,
391
+ "loss": 0.4217,
392
+ "step": 1792
393
+ },
394
+ {
395
+ "epoch": 3.81,
396
+ "eval_loss": 0.7212890982627869,
397
+ "eval_runtime": 251.4573,
398
+ "eval_samples_per_second": 8.145,
399
+ "eval_wer": 0.674878162761199,
400
+ "step": 1792
401
+ },
402
+ {
403
+ "epoch": 3.95,
404
+ "learning_rate": 0.00016975697569756977,
405
+ "loss": 0.4466,
406
+ "step": 1856
407
+ },
408
+ {
409
+ "epoch": 3.95,
410
+ "eval_loss": 0.7866759300231934,
411
+ "eval_runtime": 252.4271,
412
+ "eval_samples_per_second": 8.113,
413
+ "eval_wer": 0.7039855798117365,
414
+ "step": 1856
415
+ },
416
+ {
417
+ "epoch": 4.09,
418
+ "learning_rate": 0.00016831683168316833,
419
+ "loss": 0.4019,
420
+ "step": 1920
421
+ },
422
+ {
423
+ "epoch": 4.09,
424
+ "eval_loss": 0.724478006362915,
425
+ "eval_runtime": 252.8043,
426
+ "eval_samples_per_second": 8.101,
427
+ "eval_wer": 0.6720074771346551,
428
+ "step": 1920
429
+ },
430
+ {
431
+ "epoch": 4.22,
432
+ "learning_rate": 0.0001668766876687669,
433
+ "loss": 0.3812,
434
+ "step": 1984
435
+ },
436
+ {
437
+ "epoch": 4.22,
438
+ "eval_loss": 0.7306143045425415,
439
+ "eval_runtime": 252.5346,
440
+ "eval_samples_per_second": 8.11,
441
+ "eval_wer": 0.671206355564457,
442
+ "step": 1984
443
+ },
444
+ {
445
+ "epoch": 4.36,
446
+ "learning_rate": 0.00016543654365436546,
447
+ "loss": 0.3662,
448
+ "step": 2048
449
+ },
450
+ {
451
+ "epoch": 4.36,
452
+ "eval_loss": 0.7334802150726318,
453
+ "eval_runtime": 251.9822,
454
+ "eval_samples_per_second": 8.128,
455
+ "eval_wer": 0.6639295013018226,
456
+ "step": 2048
457
+ },
458
+ {
459
+ "epoch": 4.49,
460
+ "learning_rate": 0.000163996399639964,
461
+ "loss": 0.3684,
462
+ "step": 2112
463
+ },
464
+ {
465
+ "epoch": 4.49,
466
+ "eval_loss": 0.7250552177429199,
467
+ "eval_runtime": 251.8402,
468
+ "eval_samples_per_second": 8.132,
469
+ "eval_wer": 0.6481073502904066,
470
+ "step": 2112
471
+ },
472
+ {
473
+ "epoch": 4.63,
474
+ "learning_rate": 0.00016255625562556255,
475
+ "loss": 0.3809,
476
+ "step": 2176
477
+ },
478
+ {
479
+ "epoch": 4.63,
480
+ "eval_loss": 0.7295255064964294,
481
+ "eval_runtime": 251.8173,
482
+ "eval_samples_per_second": 8.133,
483
+ "eval_wer": 0.666533146404967,
484
+ "step": 2176
485
+ },
486
+ {
487
+ "epoch": 4.77,
488
+ "learning_rate": 0.00016111611161116112,
489
+ "loss": 0.3618,
490
+ "step": 2240
491
+ },
492
+ {
493
+ "epoch": 4.77,
494
+ "eval_loss": 0.725806474685669,
495
+ "eval_runtime": 251.7567,
496
+ "eval_samples_per_second": 8.135,
497
+ "eval_wer": 0.6467721476734094,
498
+ "step": 2240
499
+ },
500
+ {
501
+ "epoch": 4.9,
502
+ "learning_rate": 0.00015967596759675968,
503
+ "loss": 0.3896,
504
+ "step": 2304
505
+ },
506
+ {
507
+ "epoch": 4.9,
508
+ "eval_loss": 0.6914154291152954,
509
+ "eval_runtime": 253.233,
510
+ "eval_samples_per_second": 8.087,
511
+ "eval_wer": 0.6378262901395286,
512
+ "step": 2304
513
+ },
514
+ {
515
+ "epoch": 5.04,
516
+ "learning_rate": 0.00015823582358235824,
517
+ "loss": 0.3381,
518
+ "step": 2368
519
+ },
520
+ {
521
+ "epoch": 5.04,
522
+ "eval_loss": 0.7252967357635498,
523
+ "eval_runtime": 251.9208,
524
+ "eval_samples_per_second": 8.13,
525
+ "eval_wer": 0.6556512450764403,
526
+ "step": 2368
527
+ },
528
+ {
529
+ "epoch": 5.17,
530
+ "learning_rate": 0.0001567956795679568,
531
+ "loss": 0.3385,
532
+ "step": 2432
533
+ },
534
+ {
535
+ "epoch": 5.17,
536
+ "eval_loss": 0.739501953125,
537
+ "eval_runtime": 251.6871,
538
+ "eval_samples_per_second": 8.137,
539
+ "eval_wer": 0.6486414313372054,
540
+ "step": 2432
541
+ },
542
+ {
543
+ "epoch": 5.31,
544
+ "learning_rate": 0.00015535553555355537,
545
+ "loss": 0.2991,
546
+ "step": 2496
547
+ },
548
+ {
549
+ "epoch": 5.31,
550
+ "eval_loss": 0.718348503112793,
551
+ "eval_runtime": 252.6947,
552
+ "eval_samples_per_second": 8.105,
553
+ "eval_wer": 0.6476400293744575,
554
+ "step": 2496
555
+ },
556
+ {
557
+ "epoch": 5.45,
558
+ "learning_rate": 0.00015391539153915393,
559
+ "loss": 0.3219,
560
+ "step": 2560
561
+ },
562
+ {
563
+ "epoch": 5.45,
564
+ "eval_loss": 0.6939400434494019,
565
+ "eval_runtime": 252.3364,
566
+ "eval_samples_per_second": 8.116,
567
+ "eval_wer": 0.6454369450564124,
568
+ "step": 2560
569
+ },
570
+ {
571
+ "epoch": 5.58,
572
+ "learning_rate": 0.0001524752475247525,
573
+ "loss": 0.3212,
574
+ "step": 2624
575
+ },
576
+ {
577
+ "epoch": 5.58,
578
+ "eval_loss": 0.723796010017395,
579
+ "eval_runtime": 252.2722,
580
+ "eval_samples_per_second": 8.118,
581
+ "eval_wer": 0.6450363842713132,
582
+ "step": 2624
583
+ },
584
+ {
585
+ "epoch": 5.72,
586
+ "learning_rate": 0.00015103510351035103,
587
+ "loss": 0.3152,
588
+ "step": 2688
589
+ },
590
+ {
591
+ "epoch": 5.72,
592
+ "eval_loss": 0.7430602312088013,
593
+ "eval_runtime": 252.488,
594
+ "eval_samples_per_second": 8.111,
595
+ "eval_wer": 0.6561853261232392,
596
+ "step": 2688
597
+ },
598
+ {
599
+ "epoch": 5.86,
600
+ "learning_rate": 0.0001495949594959496,
601
+ "loss": 0.3359,
602
+ "step": 2752
603
+ },
604
+ {
605
+ "epoch": 5.86,
606
+ "eval_loss": 0.7211728096008301,
607
+ "eval_runtime": 252.04,
608
+ "eval_samples_per_second": 8.126,
609
+ "eval_wer": 0.6389612123639762,
610
+ "step": 2752
611
+ },
612
+ {
613
+ "epoch": 5.99,
614
+ "learning_rate": 0.00014815481548154818,
615
+ "loss": 0.3171,
616
+ "step": 2816
617
+ },
618
+ {
619
+ "epoch": 5.99,
620
+ "eval_loss": 0.707861602306366,
621
+ "eval_runtime": 253.0372,
622
+ "eval_samples_per_second": 8.094,
623
+ "eval_wer": 0.6495093130382535,
624
+ "step": 2816
625
+ },
626
+ {
627
+ "epoch": 6.13,
628
+ "learning_rate": 0.00014671467146714671,
629
+ "loss": 0.2806,
630
+ "step": 2880
631
+ },
632
+ {
633
+ "epoch": 6.13,
634
+ "eval_loss": 0.6907961368560791,
635
+ "eval_runtime": 254.422,
636
+ "eval_samples_per_second": 8.05,
637
+ "eval_wer": 0.6292142332598972,
638
+ "step": 2880
639
+ },
640
+ {
641
+ "epoch": 6.26,
642
+ "learning_rate": 0.00014527452745274528,
643
+ "loss": 0.2765,
644
+ "step": 2944
645
+ },
646
+ {
647
+ "epoch": 6.26,
648
+ "eval_loss": 0.7019878029823303,
649
+ "eval_runtime": 252.9804,
650
+ "eval_samples_per_second": 8.095,
651
+ "eval_wer": 0.6294145136524467,
652
+ "step": 2944
653
+ },
654
+ {
655
+ "epoch": 6.4,
656
+ "learning_rate": 0.00014383438343834384,
657
+ "loss": 0.2854,
658
+ "step": 3008
659
+ },
660
+ {
661
+ "epoch": 6.4,
662
+ "eval_loss": 0.811292290687561,
663
+ "eval_runtime": 253.017,
664
+ "eval_samples_per_second": 8.094,
665
+ "eval_wer": 0.6774818078643434,
666
+ "step": 3008
667
+ },
668
+ {
669
+ "epoch": 6.54,
670
+ "learning_rate": 0.0001423942394239424,
671
+ "loss": 0.2805,
672
+ "step": 3072
673
+ },
674
+ {
675
+ "epoch": 6.54,
676
+ "eval_loss": 0.7887662649154663,
677
+ "eval_runtime": 251.6217,
678
+ "eval_samples_per_second": 8.139,
679
+ "eval_wer": 0.6378930502703786,
680
+ "step": 3072
681
+ },
682
+ {
683
+ "epoch": 6.67,
684
+ "learning_rate": 0.00014095409540954096,
685
+ "loss": 0.2861,
686
+ "step": 3136
687
+ },
688
+ {
689
+ "epoch": 6.67,
690
+ "eval_loss": 0.7010805606842041,
691
+ "eval_runtime": 252.2814,
692
+ "eval_samples_per_second": 8.118,
693
+ "eval_wer": 0.635356165298084,
694
+ "step": 3136
695
+ },
696
+ {
697
+ "epoch": 6.81,
698
+ "learning_rate": 0.00013951395139513953,
699
+ "loss": 0.2903,
700
+ "step": 3200
701
+ },
702
+ {
703
+ "epoch": 6.81,
704
+ "eval_loss": 0.7183523774147034,
705
+ "eval_runtime": 253.427,
706
+ "eval_samples_per_second": 8.081,
707
+ "eval_wer": 0.6266105881567527,
708
+ "step": 3200
709
+ },
710
+ {
711
+ "epoch": 6.94,
712
+ "learning_rate": 0.00013807380738073806,
713
+ "loss": 0.2798,
714
+ "step": 3264
715
+ },
716
+ {
717
+ "epoch": 6.94,
718
+ "eval_loss": 0.7161312699317932,
719
+ "eval_runtime": 253.1259,
720
+ "eval_samples_per_second": 8.091,
721
+ "eval_wer": 0.6217370986047133,
722
+ "step": 3264
723
+ },
724
+ {
725
+ "epoch": 7.08,
726
+ "learning_rate": 0.00013663366336633665,
727
+ "loss": 0.2528,
728
+ "step": 3328
729
+ },
730
+ {
731
+ "epoch": 7.08,
732
+ "eval_loss": 0.7560293674468994,
733
+ "eval_runtime": 252.846,
734
+ "eval_samples_per_second": 8.1,
735
+ "eval_wer": 0.6506442352627011,
736
+ "step": 3328
737
+ },
738
+ {
739
+ "epoch": 7.22,
740
+ "learning_rate": 0.00013519351935193521,
741
+ "loss": 0.2549,
742
+ "step": 3392
743
+ },
744
+ {
745
+ "epoch": 7.22,
746
+ "eval_loss": 0.7602236270904541,
747
+ "eval_runtime": 253.2296,
748
+ "eval_samples_per_second": 8.088,
749
+ "eval_wer": 0.6505774751318513,
750
+ "step": 3392
751
+ },
752
+ {
753
+ "epoch": 7.35,
754
+ "learning_rate": 0.00013375337533753375,
755
+ "loss": 0.2431,
756
+ "step": 3456
757
+ },
758
+ {
759
+ "epoch": 7.35,
760
+ "eval_loss": 0.6875869035720825,
761
+ "eval_runtime": 252.6805,
762
+ "eval_samples_per_second": 8.105,
763
+ "eval_wer": 0.61859937245477,
764
+ "step": 3456
765
+ },
766
+ {
767
+ "epoch": 7.49,
768
+ "learning_rate": 0.0001323132313231323,
769
+ "loss": 0.2506,
770
+ "step": 3520
771
+ },
772
+ {
773
+ "epoch": 7.49,
774
+ "eval_loss": 0.7143360376358032,
775
+ "eval_runtime": 253.2779,
776
+ "eval_samples_per_second": 8.086,
777
+ "eval_wer": 0.6345550437278857,
778
+ "step": 3520
779
+ },
780
+ {
781
+ "epoch": 7.63,
782
+ "learning_rate": 0.00013087308730873087,
783
+ "loss": 0.2497,
784
+ "step": 3584
785
+ },
786
+ {
787
+ "epoch": 7.63,
788
+ "eval_loss": 0.7228794693946838,
789
+ "eval_runtime": 252.2511,
790
+ "eval_samples_per_second": 8.119,
791
+ "eval_wer": 0.6364910875225316,
792
+ "step": 3584
793
+ },
794
+ {
795
+ "epoch": 7.76,
796
+ "learning_rate": 0.00012943294329432944,
797
+ "loss": 0.2492,
798
+ "step": 3648
799
+ },
800
+ {
801
+ "epoch": 7.76,
802
+ "eval_loss": 0.7050414085388184,
803
+ "eval_runtime": 253.5665,
804
+ "eval_samples_per_second": 8.077,
805
+ "eval_wer": 0.6172641698377729,
806
+ "step": 3648
807
+ },
808
+ {
809
+ "epoch": 7.9,
810
+ "learning_rate": 0.000127992799279928,
811
+ "loss": 0.2444,
812
+ "step": 3712
813
+ },
814
+ {
815
+ "epoch": 7.9,
816
+ "eval_loss": 0.7215788960456848,
817
+ "eval_runtime": 252.8539,
818
+ "eval_samples_per_second": 8.1,
819
+ "eval_wer": 0.6314173175779425,
820
+ "step": 3712
821
+ },
822
+ {
823
+ "epoch": 8.03,
824
+ "learning_rate": 0.00012655265526552656,
825
+ "loss": 0.2438,
826
+ "step": 3776
827
+ },
828
+ {
829
+ "epoch": 8.03,
830
+ "eval_loss": 0.7077196836471558,
831
+ "eval_runtime": 253.6898,
832
+ "eval_samples_per_second": 8.073,
833
+ "eval_wer": 0.6182655718005208,
834
+ "step": 3776
835
+ },
836
+ {
837
+ "epoch": 8.17,
838
+ "learning_rate": 0.0001251125112511251,
839
+ "loss": 0.2083,
840
+ "step": 3840
841
+ },
842
+ {
843
+ "epoch": 8.17,
844
+ "eval_loss": 0.7524451613426208,
845
+ "eval_runtime": 255.0333,
846
+ "eval_samples_per_second": 8.03,
847
+ "eval_wer": 0.6258762267174044,
848
+ "step": 3840
849
+ },
850
+ {
851
+ "epoch": 8.31,
852
+ "learning_rate": 0.0001236723672367237,
853
+ "loss": 0.204,
854
+ "step": 3904
855
+ },
856
+ {
857
+ "epoch": 8.31,
858
+ "eval_loss": 0.7626018524169922,
859
+ "eval_runtime": 253.5087,
860
+ "eval_samples_per_second": 8.079,
861
+ "eval_wer": 0.6330863208491888,
862
+ "step": 3904
863
+ },
864
+ {
865
+ "epoch": 8.44,
866
+ "learning_rate": 0.00012223222322232225,
867
+ "loss": 0.2353,
868
+ "step": 3968
869
+ },
870
+ {
871
+ "epoch": 8.44,
872
+ "eval_loss": 0.7177530527114868,
873
+ "eval_runtime": 253.4757,
874
+ "eval_samples_per_second": 8.08,
875
+ "eval_wer": 0.6127244809399827,
876
+ "step": 3968
877
+ },
878
+ {
879
+ "epoch": 8.58,
880
+ "learning_rate": 0.0001207920792079208,
881
+ "loss": 0.2208,
882
+ "step": 4032
883
+ },
884
+ {
885
+ "epoch": 8.58,
886
+ "eval_loss": 0.7555935382843018,
887
+ "eval_runtime": 253.7282,
888
+ "eval_samples_per_second": 8.072,
889
+ "eval_wer": 0.6294812737832967,
890
+ "step": 4032
891
+ },
892
+ {
893
+ "epoch": 8.71,
894
+ "learning_rate": 0.00011935193519351935,
895
+ "loss": 0.2266,
896
+ "step": 4096
897
+ },
898
+ {
899
+ "epoch": 8.71,
900
+ "eval_loss": 0.7341886162757874,
901
+ "eval_runtime": 253.8255,
902
+ "eval_samples_per_second": 8.069,
903
+ "eval_wer": 0.6294812737832967,
904
+ "step": 4096
905
+ },
906
+ {
907
+ "epoch": 8.85,
908
+ "learning_rate": 0.00011791179117911792,
909
+ "loss": 0.2188,
910
+ "step": 4160
911
+ },
912
+ {
913
+ "epoch": 8.85,
914
+ "eval_loss": 0.7594360709190369,
915
+ "eval_runtime": 253.012,
916
+ "eval_samples_per_second": 8.094,
917
+ "eval_wer": 0.6449028640096135,
918
+ "step": 4160
919
+ },
920
+ {
921
+ "epoch": 8.99,
922
+ "learning_rate": 0.00011647164716471647,
923
+ "loss": 0.231,
924
+ "step": 4224
925
+ },
926
+ {
927
+ "epoch": 8.99,
928
+ "eval_loss": 0.6901054382324219,
929
+ "eval_runtime": 253.458,
930
+ "eval_samples_per_second": 8.08,
931
+ "eval_wer": 0.6077842312570932,
932
+ "step": 4224
933
+ },
934
+ {
935
+ "epoch": 9.12,
936
+ "learning_rate": 0.00011503150315031504,
937
+ "loss": 0.2063,
938
+ "step": 4288
939
+ },
940
+ {
941
+ "epoch": 9.12,
942
+ "eval_loss": 0.7627975344657898,
943
+ "eval_runtime": 253.6008,
944
+ "eval_samples_per_second": 8.076,
945
+ "eval_wer": 0.6116563188463849,
946
+ "step": 4288
947
+ },
948
+ {
949
+ "epoch": 9.26,
950
+ "learning_rate": 0.00011359135913591358,
951
+ "loss": 0.1931,
952
+ "step": 4352
953
+ },
954
+ {
955
+ "epoch": 9.26,
956
+ "eval_loss": 0.7425730228424072,
957
+ "eval_runtime": 254.3179,
958
+ "eval_samples_per_second": 8.053,
959
+ "eval_wer": 0.6149943253888778,
960
+ "step": 4352
961
+ },
962
+ {
963
+ "epoch": 9.4,
964
+ "learning_rate": 0.00011215121512151216,
965
+ "loss": 0.2189,
966
+ "step": 4416
967
+ },
968
+ {
969
+ "epoch": 9.4,
970
+ "eval_loss": 0.7322723269462585,
971
+ "eval_runtime": 254.9695,
972
+ "eval_samples_per_second": 8.032,
973
+ "eval_wer": 0.6066493090326457,
974
+ "step": 4416
975
+ },
976
+ {
977
+ "epoch": 9.53,
978
+ "learning_rate": 0.00011071107110711072,
979
+ "loss": 0.1931,
980
+ "step": 4480
981
+ },
982
+ {
983
+ "epoch": 9.53,
984
+ "eval_loss": 0.7253696918487549,
985
+ "eval_runtime": 253.0608,
986
+ "eval_samples_per_second": 8.093,
987
+ "eval_wer": 0.613525602510181,
988
+ "step": 4480
989
+ },
990
+ {
991
+ "epoch": 9.67,
992
+ "learning_rate": 0.00010927092709270927,
993
+ "loss": 0.2017,
994
+ "step": 4544
995
+ },
996
+ {
997
+ "epoch": 9.67,
998
+ "eval_loss": 0.6984794735908508,
999
+ "eval_runtime": 253.842,
1000
+ "eval_samples_per_second": 8.068,
1001
+ "eval_wer": 0.6072501502102944,
1002
+ "step": 4544
1003
+ },
1004
+ {
1005
+ "epoch": 9.8,
1006
+ "learning_rate": 0.00010783078307830783,
1007
+ "loss": 0.2169,
1008
+ "step": 4608
1009
+ },
1010
+ {
1011
+ "epoch": 9.8,
1012
+ "eval_loss": 0.698683500289917,
1013
+ "eval_runtime": 253.2368,
1014
+ "eval_samples_per_second": 8.087,
1015
+ "eval_wer": 0.6078509913879431,
1016
+ "step": 4608
1017
+ },
1018
+ {
1019
+ "epoch": 9.94,
1020
+ "learning_rate": 0.00010639063906390641,
1021
+ "loss": 0.1956,
1022
+ "step": 4672
1023
+ },
1024
+ {
1025
+ "epoch": 9.94,
1026
+ "eval_loss": 0.7127052545547485,
1027
+ "eval_runtime": 254.4368,
1028
+ "eval_samples_per_second": 8.049,
1029
+ "eval_wer": 0.6073836704719942,
1030
+ "step": 4672
1031
+ },
1032
+ {
1033
+ "epoch": 10.08,
1034
+ "learning_rate": 0.00010495049504950496,
1035
+ "loss": 0.1972,
1036
+ "step": 4736
1037
+ },
1038
+ {
1039
+ "epoch": 10.08,
1040
+ "eval_loss": 0.7555835247039795,
1041
+ "eval_runtime": 253.4953,
1042
+ "eval_samples_per_second": 8.079,
1043
+ "eval_wer": 0.6094532345283397,
1044
+ "step": 4736
1045
+ },
1046
+ {
1047
+ "epoch": 10.21,
1048
+ "learning_rate": 0.00010351035103510351,
1049
+ "loss": 0.1827,
1050
+ "step": 4800
1051
+ },
1052
+ {
1053
+ "epoch": 10.21,
1054
+ "eval_loss": 0.7492347359657288,
1055
+ "eval_runtime": 255.7545,
1056
+ "eval_samples_per_second": 8.008,
1057
+ "eval_wer": 0.6165298083984244,
1058
+ "step": 4800
1059
+ },
1060
+ {
1061
+ "epoch": 10.35,
1062
+ "learning_rate": 0.00010207020702070207,
1063
+ "loss": 0.1798,
1064
+ "step": 4864
1065
+ },
1066
+ {
1067
+ "epoch": 10.35,
1068
+ "eval_loss": 0.7403519749641418,
1069
+ "eval_runtime": 254.7315,
1070
+ "eval_samples_per_second": 8.04,
1071
+ "eval_wer": 0.6058481874624474,
1072
+ "step": 4864
1073
+ },
1074
+ {
1075
+ "epoch": 10.49,
1076
+ "learning_rate": 0.00010063006300630065,
1077
+ "loss": 0.1839,
1078
+ "step": 4928
1079
+ },
1080
+ {
1081
+ "epoch": 10.49,
1082
+ "eval_loss": 0.7588409185409546,
1083
+ "eval_runtime": 255.5749,
1084
+ "eval_samples_per_second": 8.013,
1085
+ "eval_wer": 0.60931971426664,
1086
+ "step": 4928
1087
+ },
1088
+ {
1089
+ "epoch": 10.62,
1090
+ "learning_rate": 9.91899189918992e-05,
1091
+ "loss": 0.1823,
1092
+ "step": 4992
1093
+ },
1094
+ {
1095
+ "epoch": 10.62,
1096
+ "eval_loss": 0.7923687696456909,
1097
+ "eval_runtime": 254.6347,
1098
+ "eval_samples_per_second": 8.043,
1099
+ "eval_wer": 0.621403297950464,
1100
+ "step": 4992
1101
+ },
1102
+ {
1103
+ "epoch": 10.76,
1104
+ "learning_rate": 9.774977497749776e-05,
1105
+ "loss": 0.1749,
1106
+ "step": 5056
1107
+ },
1108
+ {
1109
+ "epoch": 10.76,
1110
+ "eval_loss": 0.7493650913238525,
1111
+ "eval_runtime": 252.6472,
1112
+ "eval_samples_per_second": 8.106,
1113
+ "eval_wer": 0.6101875959676881,
1114
+ "step": 5056
1115
+ },
1116
+ {
1117
+ "epoch": 10.89,
1118
+ "learning_rate": 9.630963096309631e-05,
1119
+ "loss": 0.1876,
1120
+ "step": 5120
1121
+ },
1122
+ {
1123
+ "epoch": 10.89,
1124
+ "eval_loss": 0.7571142911911011,
1125
+ "eval_runtime": 253.4609,
1126
+ "eval_samples_per_second": 8.08,
1127
+ "eval_wer": 0.6097870351825889,
1128
+ "step": 5120
1129
+ },
1130
+ {
1131
+ "epoch": 11.03,
1132
+ "learning_rate": 9.486948694869487e-05,
1133
+ "loss": 0.2027,
1134
+ "step": 5184
1135
+ },
1136
+ {
1137
+ "epoch": 11.03,
1138
+ "eval_loss": 0.7235008478164673,
1139
+ "eval_runtime": 252.9146,
1140
+ "eval_samples_per_second": 8.098,
1141
+ "eval_wer": 0.6036451031444021,
1142
+ "step": 5184
1143
+ },
1144
+ {
1145
+ "epoch": 11.17,
1146
+ "learning_rate": 9.342934293429343e-05,
1147
+ "loss": 0.1699,
1148
+ "step": 5248
1149
+ },
1150
+ {
1151
+ "epoch": 11.17,
1152
+ "eval_loss": 0.7639468908309937,
1153
+ "eval_runtime": 253.3789,
1154
+ "eval_samples_per_second": 8.083,
1155
+ "eval_wer": 0.6093864743974898,
1156
+ "step": 5248
1157
+ },
1158
+ {
1159
+ "epoch": 11.3,
1160
+ "learning_rate": 9.1989198919892e-05,
1161
+ "loss": 0.1556,
1162
+ "step": 5312
1163
+ },
1164
+ {
1165
+ "epoch": 11.3,
1166
+ "eval_loss": 0.7758333683013916,
1167
+ "eval_runtime": 255.9522,
1168
+ "eval_samples_per_second": 8.001,
1169
+ "eval_wer": 0.6128580012016823,
1170
+ "step": 5312
1171
+ },
1172
+ {
1173
+ "epoch": 11.44,
1174
+ "learning_rate": 9.054905490549054e-05,
1175
+ "loss": 0.1679,
1176
+ "step": 5376
1177
+ },
1178
+ {
1179
+ "epoch": 11.44,
1180
+ "eval_loss": 0.7407320737838745,
1181
+ "eval_runtime": 253.1505,
1182
+ "eval_samples_per_second": 8.09,
1183
+ "eval_wer": 0.5980372521530142,
1184
+ "step": 5376
1185
+ },
1186
+ {
1187
+ "epoch": 11.57,
1188
+ "learning_rate": 8.910891089108912e-05,
1189
+ "loss": 0.1681,
1190
+ "step": 5440
1191
+ },
1192
+ {
1193
+ "epoch": 11.57,
1194
+ "eval_loss": 0.7515969276428223,
1195
+ "eval_runtime": 254.0821,
1196
+ "eval_samples_per_second": 8.06,
1197
+ "eval_wer": 0.6076507109953936,
1198
+ "step": 5440
1199
+ },
1200
+ {
1201
+ "epoch": 11.71,
1202
+ "learning_rate": 8.766876687668767e-05,
1203
+ "loss": 0.1814,
1204
+ "step": 5504
1205
+ },
1206
+ {
1207
+ "epoch": 11.71,
1208
+ "eval_loss": 0.7393285036087036,
1209
+ "eval_runtime": 255.9386,
1210
+ "eval_samples_per_second": 8.002,
1211
+ "eval_wer": 0.6063822685092463,
1212
+ "step": 5504
1213
+ },
1214
+ {
1215
+ "epoch": 11.85,
1216
+ "learning_rate": 8.622862286228623e-05,
1217
+ "loss": 0.162,
1218
+ "step": 5568
1219
+ },
1220
+ {
1221
+ "epoch": 11.85,
1222
+ "eval_loss": 0.7688995003700256,
1223
+ "eval_runtime": 255.0601,
1224
+ "eval_samples_per_second": 8.029,
1225
+ "eval_wer": 0.6063155083783964,
1226
+ "step": 5568
1227
+ },
1228
+ {
1229
+ "epoch": 11.98,
1230
+ "learning_rate": 8.47884788478848e-05,
1231
+ "loss": 0.1835,
1232
+ "step": 5632
1233
+ },
1234
+ {
1235
+ "epoch": 11.98,
1236
+ "eval_loss": 0.7229721546173096,
1237
+ "eval_runtime": 256.1311,
1238
+ "eval_samples_per_second": 7.996,
1239
+ "eval_wer": 0.6020428600040056,
1240
+ "step": 5632
1241
+ },
1242
+ {
1243
+ "epoch": 12.12,
1244
+ "learning_rate": 8.334833483348336e-05,
1245
+ "loss": 0.1491,
1246
+ "step": 5696
1247
+ },
1248
+ {
1249
+ "epoch": 12.12,
1250
+ "eval_loss": 0.7780319452285767,
1251
+ "eval_runtime": 253.9764,
1252
+ "eval_samples_per_second": 8.064,
1253
+ "eval_wer": 0.6079845116496428,
1254
+ "step": 5696
1255
+ },
1256
+ {
1257
+ "epoch": 12.26,
1258
+ "learning_rate": 8.19081908190819e-05,
1259
+ "loss": 0.163,
1260
+ "step": 5760
1261
+ },
1262
+ {
1263
+ "epoch": 12.26,
1264
+ "eval_loss": 0.7481300830841064,
1265
+ "eval_runtime": 255.5174,
1266
+ "eval_samples_per_second": 8.015,
1267
+ "eval_wer": 0.6035115828827025,
1268
+ "step": 5760
1269
+ },
1270
+ {
1271
+ "epoch": 12.39,
1272
+ "learning_rate": 8.046804680468047e-05,
1273
+ "loss": 0.1564,
1274
+ "step": 5824
1275
+ },
1276
+ {
1277
+ "epoch": 12.39,
1278
+ "eval_loss": 0.7477182149887085,
1279
+ "eval_runtime": 253.9361,
1280
+ "eval_samples_per_second": 8.065,
1281
+ "eval_wer": 0.6073169103411443,
1282
+ "step": 5824
1283
+ },
1284
+ {
1285
+ "epoch": 12.53,
1286
+ "learning_rate": 7.902790279027903e-05,
1287
+ "loss": 0.1651,
1288
+ "step": 5888
1289
+ },
1290
+ {
1291
+ "epoch": 12.53,
1292
+ "eval_loss": 0.7700252532958984,
1293
+ "eval_runtime": 255.811,
1294
+ "eval_samples_per_second": 8.006,
1295
+ "eval_wer": 0.6119901195006342,
1296
+ "step": 5888
1297
+ },
1298
+ {
1299
+ "epoch": 12.66,
1300
+ "learning_rate": 7.75877587758776e-05,
1301
+ "loss": 0.1406,
1302
+ "step": 5952
1303
+ },
1304
+ {
1305
+ "epoch": 12.66,
1306
+ "eval_loss": 0.7303926348686218,
1307
+ "eval_runtime": 255.3564,
1308
+ "eval_samples_per_second": 8.02,
1309
+ "eval_wer": 0.6010414580412577,
1310
+ "step": 5952
1311
+ },
1312
+ {
1313
+ "epoch": 12.8,
1314
+ "learning_rate": 7.614761476147616e-05,
1315
+ "loss": 0.1454,
1316
+ "step": 6016
1317
+ },
1318
+ {
1319
+ "epoch": 12.8,
1320
+ "eval_loss": 0.7804365158081055,
1321
+ "eval_runtime": 258.3951,
1322
+ "eval_samples_per_second": 7.926,
1323
+ "eval_wer": 0.6108551972761866,
1324
+ "step": 6016
1325
+ },
1326
+ {
1327
+ "epoch": 12.94,
1328
+ "learning_rate": 7.470747074707472e-05,
1329
+ "loss": 0.1405,
1330
+ "step": 6080
1331
+ },
1332
+ {
1333
+ "epoch": 12.94,
1334
+ "eval_loss": 0.7660069465637207,
1335
+ "eval_runtime": 258.3136,
1336
+ "eval_samples_per_second": 7.928,
1337
+ "eval_wer": 0.5965017691434675,
1338
+ "step": 6080
1339
+ },
1340
+ {
1341
+ "epoch": 13.07,
1342
+ "learning_rate": 7.326732673267327e-05,
1343
+ "loss": 0.142,
1344
+ "step": 6144
1345
+ },
1346
+ {
1347
+ "epoch": 13.07,
1348
+ "eval_loss": 0.7802821397781372,
1349
+ "eval_runtime": 255.9638,
1350
+ "eval_samples_per_second": 8.001,
1351
+ "eval_wer": 0.6066493090326457,
1352
+ "step": 6144
1353
+ },
1354
+ {
1355
+ "epoch": 13.21,
1356
+ "learning_rate": 7.182718271827183e-05,
1357
+ "loss": 0.1423,
1358
+ "step": 6208
1359
+ },
1360
+ {
1361
+ "epoch": 13.21,
1362
+ "eval_loss": 0.7940844297409058,
1363
+ "eval_runtime": 257.753,
1364
+ "eval_samples_per_second": 7.946,
1365
+ "eval_wer": 0.6061152279858468,
1366
+ "step": 6208
1367
+ },
1368
+ {
1369
+ "epoch": 13.34,
1370
+ "learning_rate": 7.038703870387039e-05,
1371
+ "loss": 0.1494,
1372
+ "step": 6272
1373
+ },
1374
+ {
1375
+ "epoch": 13.34,
1376
+ "eval_loss": 0.7598519325256348,
1377
+ "eval_runtime": 257.9759,
1378
+ "eval_samples_per_second": 7.939,
1379
+ "eval_wer": 0.6009079377795581,
1380
+ "step": 6272
1381
+ },
1382
+ {
1383
+ "epoch": 13.48,
1384
+ "learning_rate": 6.894689468946895e-05,
1385
+ "loss": 0.1459,
1386
+ "step": 6336
1387
+ },
1388
+ {
1389
+ "epoch": 13.48,
1390
+ "eval_loss": 0.7658703327178955,
1391
+ "eval_runtime": 257.4947,
1392
+ "eval_samples_per_second": 7.954,
1393
+ "eval_wer": 0.6087188730889913,
1394
+ "step": 6336
1395
+ },
1396
+ {
1397
+ "epoch": 13.62,
1398
+ "learning_rate": 6.75067506750675e-05,
1399
+ "loss": 0.1316,
1400
+ "step": 6400
1401
+ },
1402
+ {
1403
+ "epoch": 13.62,
1404
+ "eval_loss": 0.7935535311698914,
1405
+ "eval_runtime": 256.5247,
1406
+ "eval_samples_per_second": 7.984,
1407
+ "eval_wer": 0.6047132652379998,
1408
+ "step": 6400
1409
+ },
1410
+ {
1411
+ "epoch": 13.75,
1412
+ "learning_rate": 6.606660666066608e-05,
1413
+ "loss": 0.1428,
1414
+ "step": 6464
1415
+ },
1416
+ {
1417
+ "epoch": 13.75,
1418
+ "eval_loss": 0.782112181186676,
1419
+ "eval_runtime": 259.1672,
1420
+ "eval_samples_per_second": 7.902,
1421
+ "eval_wer": 0.5998397756859604,
1422
+ "step": 6464
1423
+ },
1424
+ {
1425
+ "epoch": 13.89,
1426
+ "learning_rate": 6.462646264626463e-05,
1427
+ "loss": 0.1391,
1428
+ "step": 6528
1429
+ },
1430
+ {
1431
+ "epoch": 13.89,
1432
+ "eval_loss": 0.7535277605056763,
1433
+ "eval_runtime": 257.3279,
1434
+ "eval_samples_per_second": 7.959,
1435
+ "eval_wer": 0.5976366913679151,
1436
+ "step": 6528
1437
+ },
1438
+ {
1439
+ "epoch": 14.03,
1440
+ "learning_rate": 6.318631863186318e-05,
1441
+ "loss": 0.138,
1442
+ "step": 6592
1443
+ },
1444
+ {
1445
+ "epoch": 14.03,
1446
+ "eval_loss": 0.7846142649650574,
1447
+ "eval_runtime": 257.0571,
1448
+ "eval_samples_per_second": 7.967,
1449
+ "eval_wer": 0.5956338874424194,
1450
+ "step": 6592
1451
+ },
1452
+ {
1453
+ "epoch": 14.16,
1454
+ "learning_rate": 6.174617461746175e-05,
1455
+ "loss": 0.1299,
1456
+ "step": 6656
1457
+ },
1458
+ {
1459
+ "epoch": 14.16,
1460
+ "eval_loss": 0.7645628452301025,
1461
+ "eval_runtime": 259.0072,
1462
+ "eval_samples_per_second": 7.907,
1463
+ "eval_wer": 0.5987716135923626,
1464
+ "step": 6656
1465
+ },
1466
+ {
1467
+ "epoch": 14.3,
1468
+ "learning_rate": 6.03060306030603e-05,
1469
+ "loss": 0.1287,
1470
+ "step": 6720
1471
+ },
1472
+ {
1473
+ "epoch": 14.3,
1474
+ "eval_loss": 0.8099244832992554,
1475
+ "eval_runtime": 258.7408,
1476
+ "eval_samples_per_second": 7.915,
1477
+ "eval_wer": 0.612056879631484,
1478
+ "step": 6720
1479
+ },
1480
+ {
1481
+ "epoch": 14.43,
1482
+ "learning_rate": 5.886588658865887e-05,
1483
+ "loss": 0.1288,
1484
+ "step": 6784
1485
+ },
1486
+ {
1487
+ "epoch": 14.43,
1488
+ "eval_loss": 0.8042709231376648,
1489
+ "eval_runtime": 258.8574,
1490
+ "eval_samples_per_second": 7.912,
1491
+ "eval_wer": 0.6015755390880566,
1492
+ "step": 6784
1493
+ },
1494
+ {
1495
+ "epoch": 14.57,
1496
+ "learning_rate": 5.742574257425742e-05,
1497
+ "loss": 0.131,
1498
+ "step": 6848
1499
+ },
1500
+ {
1501
+ "epoch": 14.57,
1502
+ "eval_loss": 0.7604876756668091,
1503
+ "eval_runtime": 256.5152,
1504
+ "eval_samples_per_second": 7.984,
1505
+ "eval_wer": 0.5951665665264704,
1506
+ "step": 6848
1507
+ },
1508
+ {
1509
+ "epoch": 14.71,
1510
+ "learning_rate": 5.598559855985599e-05,
1511
+ "loss": 0.1353,
1512
+ "step": 6912
1513
+ },
1514
+ {
1515
+ "epoch": 14.71,
1516
+ "eval_loss": 0.784662127494812,
1517
+ "eval_runtime": 256.9735,
1518
+ "eval_samples_per_second": 7.97,
1519
+ "eval_wer": 0.59997329594766,
1520
+ "step": 6912
1521
+ },
1522
+ {
1523
+ "epoch": 14.84,
1524
+ "learning_rate": 5.4545454545454546e-05,
1525
+ "loss": 0.133,
1526
+ "step": 6976
1527
+ },
1528
+ {
1529
+ "epoch": 14.84,
1530
+ "eval_loss": 0.8255736827850342,
1531
+ "eval_runtime": 256.2102,
1532
+ "eval_samples_per_second": 7.993,
1533
+ "eval_wer": 0.6010414580412577,
1534
+ "step": 6976
1535
+ },
1536
+ {
1537
+ "epoch": 14.98,
1538
+ "learning_rate": 5.310531053105311e-05,
1539
+ "loss": 0.1351,
1540
+ "step": 7040
1541
+ },
1542
+ {
1543
+ "epoch": 14.98,
1544
+ "eval_loss": 0.7962599992752075,
1545
+ "eval_runtime": 256.4813,
1546
+ "eval_samples_per_second": 7.985,
1547
+ "eval_wer": 0.5983710528072635,
1548
+ "step": 7040
1549
+ },
1550
+ {
1551
+ "epoch": 15.11,
1552
+ "learning_rate": 5.1665166516651664e-05,
1553
+ "loss": 0.1156,
1554
+ "step": 7104
1555
+ },
1556
+ {
1557
+ "epoch": 15.11,
1558
+ "eval_loss": 0.8220678567886353,
1559
+ "eval_runtime": 255.8959,
1560
+ "eval_samples_per_second": 8.003,
1561
+ "eval_wer": 0.6015087789572068,
1562
+ "step": 7104
1563
+ },
1564
+ {
1565
+ "epoch": 15.25,
1566
+ "learning_rate": 5.022502250225023e-05,
1567
+ "loss": 0.1171,
1568
+ "step": 7168
1569
+ },
1570
+ {
1571
+ "epoch": 15.25,
1572
+ "eval_loss": 0.7746037840843201,
1573
+ "eval_runtime": 255.6563,
1574
+ "eval_samples_per_second": 8.011,
1575
+ "eval_wer": 0.592562921423326,
1576
+ "step": 7168
1577
+ },
1578
+ {
1579
+ "epoch": 15.39,
1580
+ "learning_rate": 4.878487848784879e-05,
1581
+ "loss": 0.1213,
1582
+ "step": 7232
1583
+ },
1584
+ {
1585
+ "epoch": 15.39,
1586
+ "eval_loss": 0.8010141253471375,
1587
+ "eval_runtime": 256.8217,
1588
+ "eval_samples_per_second": 7.974,
1589
+ "eval_wer": 0.6013084985646572,
1590
+ "step": 7232
1591
+ },
1592
+ {
1593
+ "epoch": 15.52,
1594
+ "learning_rate": 4.734473447344735e-05,
1595
+ "loss": 0.1175,
1596
+ "step": 7296
1597
+ },
1598
+ {
1599
+ "epoch": 15.52,
1600
+ "eval_loss": 0.8121392726898193,
1601
+ "eval_runtime": 255.5954,
1602
+ "eval_samples_per_second": 8.013,
1603
+ "eval_wer": 0.6077174711262434,
1604
+ "step": 7296
1605
+ },
1606
+ {
1607
+ "epoch": 15.66,
1608
+ "learning_rate": 4.59045904590459e-05,
1609
+ "loss": 0.12,
1610
+ "step": 7360
1611
+ },
1612
+ {
1613
+ "epoch": 15.66,
1614
+ "eval_loss": 0.7914840579032898,
1615
+ "eval_runtime": 256.4734,
1616
+ "eval_samples_per_second": 7.985,
1617
+ "eval_wer": 0.5952333266573202,
1618
+ "step": 7360
1619
+ },
1620
+ {
1621
+ "epoch": 15.8,
1622
+ "learning_rate": 4.4464446444644464e-05,
1623
+ "loss": 0.1147,
1624
+ "step": 7424
1625
+ },
1626
+ {
1627
+ "epoch": 15.8,
1628
+ "eval_loss": 0.7968710660934448,
1629
+ "eval_runtime": 255.805,
1630
+ "eval_samples_per_second": 8.006,
1631
+ "eval_wer": 0.5996394952934108,
1632
+ "step": 7424
1633
+ },
1634
+ {
1635
+ "epoch": 15.93,
1636
+ "learning_rate": 4.3024302430243026e-05,
1637
+ "loss": 0.1217,
1638
+ "step": 7488
1639
+ },
1640
+ {
1641
+ "epoch": 15.93,
1642
+ "eval_loss": 0.8240338563919067,
1643
+ "eval_runtime": 254.7962,
1644
+ "eval_samples_per_second": 8.038,
1645
+ "eval_wer": 0.6034448227518526,
1646
+ "step": 7488
1647
+ },
1648
+ {
1649
+ "epoch": 16.07,
1650
+ "learning_rate": 4.158415841584158e-05,
1651
+ "loss": 0.1088,
1652
+ "step": 7552
1653
+ },
1654
+ {
1655
+ "epoch": 16.07,
1656
+ "eval_loss": 0.8148845434188843,
1657
+ "eval_runtime": 255.3742,
1658
+ "eval_samples_per_second": 8.02,
1659
+ "eval_wer": 0.6006408972561587,
1660
+ "step": 7552
1661
+ },
1662
+ {
1663
+ "epoch": 16.2,
1664
+ "learning_rate": 4.0144014401440144e-05,
1665
+ "loss": 0.1086,
1666
+ "step": 7616
1667
+ },
1668
+ {
1669
+ "epoch": 16.2,
1670
+ "eval_loss": 0.8434123992919922,
1671
+ "eval_runtime": 256.199,
1672
+ "eval_samples_per_second": 7.994,
1673
+ "eval_wer": 0.6020428600040056,
1674
+ "step": 7616
1675
+ },
1676
+ {
1677
+ "epoch": 16.34,
1678
+ "learning_rate": 3.870387038703871e-05,
1679
+ "loss": 0.122,
1680
+ "step": 7680
1681
+ },
1682
+ {
1683
+ "epoch": 16.34,
1684
+ "eval_loss": 0.8277792930603027,
1685
+ "eval_runtime": 255.246,
1686
+ "eval_samples_per_second": 8.024,
1687
+ "eval_wer": 0.5944989652179719,
1688
+ "step": 7680
1689
+ },
1690
+ {
1691
+ "epoch": 16.48,
1692
+ "learning_rate": 3.726372637263726e-05,
1693
+ "loss": 0.109,
1694
+ "step": 7744
1695
+ },
1696
+ {
1697
+ "epoch": 16.48,
1698
+ "eval_loss": 0.8119443655014038,
1699
+ "eval_runtime": 255.9109,
1700
+ "eval_samples_per_second": 8.003,
1701
+ "eval_wer": 0.5994392149008612,
1702
+ "step": 7744
1703
+ },
1704
+ {
1705
+ "epoch": 16.61,
1706
+ "learning_rate": 3.5823582358235825e-05,
1707
+ "loss": 0.1159,
1708
+ "step": 7808
1709
+ },
1710
+ {
1711
+ "epoch": 16.61,
1712
+ "eval_loss": 0.8458046913146973,
1713
+ "eval_runtime": 255.3949,
1714
+ "eval_samples_per_second": 8.019,
1715
+ "eval_wer": 0.6047800253688497,
1716
+ "step": 7808
1717
+ },
1718
+ {
1719
+ "epoch": 16.75,
1720
+ "learning_rate": 3.438343834383439e-05,
1721
+ "loss": 0.0992,
1722
+ "step": 7872
1723
+ },
1724
+ {
1725
+ "epoch": 16.75,
1726
+ "eval_loss": 0.8495743870735168,
1727
+ "eval_runtime": 255.6285,
1728
+ "eval_samples_per_second": 8.012,
1729
+ "eval_wer": 0.6048467854996996,
1730
+ "step": 7872
1731
+ },
1732
+ {
1733
+ "epoch": 16.89,
1734
+ "learning_rate": 3.294329432943294e-05,
1735
+ "loss": 0.1078,
1736
+ "step": 7936
1737
+ },
1738
+ {
1739
+ "epoch": 16.89,
1740
+ "eval_loss": 0.8393277525901794,
1741
+ "eval_runtime": 255.6117,
1742
+ "eval_samples_per_second": 8.012,
1743
+ "eval_wer": 0.5990386541157621,
1744
+ "step": 7936
1745
+ },
1746
+ {
1747
+ "epoch": 17.02,
1748
+ "learning_rate": 3.1503150315031506e-05,
1749
+ "loss": 0.1096,
1750
+ "step": 8000
1751
+ },
1752
+ {
1753
+ "epoch": 17.02,
1754
+ "eval_loss": 0.8454439640045166,
1755
+ "eval_runtime": 255.5979,
1756
+ "eval_samples_per_second": 8.013,
1757
+ "eval_wer": 0.5995059750317111,
1758
+ "step": 8000
1759
+ },
1760
+ {
1761
+ "epoch": 17.16,
1762
+ "learning_rate": 3.0063006300630065e-05,
1763
+ "loss": 0.0986,
1764
+ "step": 8064
1765
+ },
1766
+ {
1767
+ "epoch": 17.16,
1768
+ "eval_loss": 0.8286423683166504,
1769
+ "eval_runtime": 257.011,
1770
+ "eval_samples_per_second": 7.969,
1771
+ "eval_wer": 0.6011082181721076,
1772
+ "step": 8064
1773
+ },
1774
+ {
1775
+ "epoch": 17.29,
1776
+ "learning_rate": 2.8622862286228624e-05,
1777
+ "loss": 0.1154,
1778
+ "step": 8128
1779
+ },
1780
+ {
1781
+ "epoch": 17.29,
1782
+ "eval_loss": 0.8368263244628906,
1783
+ "eval_runtime": 258.2275,
1784
+ "eval_samples_per_second": 7.931,
1785
+ "eval_wer": 0.5988383737232125,
1786
+ "step": 8128
1787
+ },
1788
+ {
1789
+ "epoch": 17.43,
1790
+ "learning_rate": 2.7182718271827183e-05,
1791
+ "loss": 0.0927,
1792
+ "step": 8192
1793
+ },
1794
+ {
1795
+ "epoch": 17.43,
1796
+ "eval_loss": 0.861258864402771,
1797
+ "eval_runtime": 257.1554,
1798
+ "eval_samples_per_second": 7.964,
1799
+ "eval_wer": 0.6089191534815408,
1800
+ "step": 8192
1801
+ },
1802
+ {
1803
+ "epoch": 17.57,
1804
+ "learning_rate": 2.5742574257425746e-05,
1805
+ "loss": 0.1056,
1806
+ "step": 8256
1807
+ },
1808
+ {
1809
+ "epoch": 17.57,
1810
+ "eval_loss": 0.8197841644287109,
1811
+ "eval_runtime": 256.4042,
1812
+ "eval_samples_per_second": 7.987,
1813
+ "eval_wer": 0.592562921423326,
1814
+ "step": 8256
1815
+ },
1816
+ {
1817
+ "epoch": 17.7,
1818
+ "learning_rate": 2.4302430243024305e-05,
1819
+ "loss": 0.1112,
1820
+ "step": 8320
1821
+ },
1822
+ {
1823
+ "epoch": 17.7,
1824
+ "eval_loss": 0.8646882772445679,
1825
+ "eval_runtime": 257.1754,
1826
+ "eval_samples_per_second": 7.963,
1827
+ "eval_wer": 0.5978369717604647,
1828
+ "step": 8320
1829
+ },
1830
+ {
1831
+ "epoch": 17.84,
1832
+ "learning_rate": 2.2862286228622864e-05,
1833
+ "loss": 0.0962,
1834
+ "step": 8384
1835
+ },
1836
+ {
1837
+ "epoch": 17.84,
1838
+ "eval_loss": 0.8387396335601807,
1839
+ "eval_runtime": 255.7406,
1840
+ "eval_samples_per_second": 8.008,
1841
+ "eval_wer": 0.5963014887509179,
1842
+ "step": 8384
1843
+ },
1844
+ {
1845
+ "epoch": 17.97,
1846
+ "learning_rate": 2.1422142214221423e-05,
1847
+ "loss": 0.0973,
1848
+ "step": 8448
1849
+ },
1850
+ {
1851
+ "epoch": 17.97,
1852
+ "eval_loss": 0.8584703207015991,
1853
+ "eval_runtime": 256.391,
1854
+ "eval_samples_per_second": 7.988,
1855
+ "eval_wer": 0.5963682488817678,
1856
+ "step": 8448
1857
+ },
1858
+ {
1859
+ "epoch": 18.11,
1860
+ "learning_rate": 1.9981998199819982e-05,
1861
+ "loss": 0.1013,
1862
+ "step": 8512
1863
+ },
1864
+ {
1865
+ "epoch": 18.11,
1866
+ "eval_loss": 0.8741297721862793,
1867
+ "eval_runtime": 257.1245,
1868
+ "eval_samples_per_second": 7.965,
1869
+ "eval_wer": 0.6001735763402096,
1870
+ "step": 8512
1871
+ },
1872
+ {
1873
+ "epoch": 18.25,
1874
+ "learning_rate": 1.854185418541854e-05,
1875
+ "loss": 0.1037,
1876
+ "step": 8576
1877
+ },
1878
+ {
1879
+ "epoch": 18.25,
1880
+ "eval_loss": 0.8630872368812561,
1881
+ "eval_runtime": 257.6611,
1882
+ "eval_samples_per_second": 7.948,
1883
+ "eval_wer": 0.5943654449562721,
1884
+ "step": 8576
1885
+ },
1886
+ {
1887
+ "epoch": 18.38,
1888
+ "learning_rate": 1.71017101710171e-05,
1889
+ "loss": 0.0947,
1890
+ "step": 8640
1891
+ },
1892
+ {
1893
+ "epoch": 18.38,
1894
+ "eval_loss": 0.86741703748703,
1895
+ "eval_runtime": 257.8968,
1896
+ "eval_samples_per_second": 7.941,
1897
+ "eval_wer": 0.596768809666867,
1898
+ "step": 8640
1899
+ },
1900
+ {
1901
+ "epoch": 18.52,
1902
+ "learning_rate": 1.5661566156615663e-05,
1903
+ "loss": 0.0884,
1904
+ "step": 8704
1905
+ },
1906
+ {
1907
+ "epoch": 18.52,
1908
+ "eval_loss": 0.8764230608940125,
1909
+ "eval_runtime": 256.705,
1910
+ "eval_samples_per_second": 7.978,
1911
+ "eval_wer": 0.5943654449562721,
1912
+ "step": 8704
1913
+ },
1914
+ {
1915
+ "epoch": 18.66,
1916
+ "learning_rate": 1.4221422142214222e-05,
1917
+ "loss": 0.0997,
1918
+ "step": 8768
1919
+ },
1920
+ {
1921
+ "epoch": 18.66,
1922
+ "eval_loss": 0.8616538047790527,
1923
+ "eval_runtime": 256.2987,
1924
+ "eval_samples_per_second": 7.991,
1925
+ "eval_wer": 0.593030242339275,
1926
+ "step": 8768
1927
+ },
1928
+ {
1929
+ "epoch": 18.79,
1930
+ "learning_rate": 1.2781278127812781e-05,
1931
+ "loss": 0.1066,
1932
+ "step": 8832
1933
+ },
1934
+ {
1935
+ "epoch": 18.79,
1936
+ "eval_loss": 0.8461120128631592,
1937
+ "eval_runtime": 255.9062,
1938
+ "eval_samples_per_second": 8.003,
1939
+ "eval_wer": 0.5889578743574337,
1940
+ "step": 8832
1941
+ },
1942
+ {
1943
+ "epoch": 18.93,
1944
+ "learning_rate": 1.1341134113411342e-05,
1945
+ "loss": 0.0952,
1946
+ "step": 8896
1947
+ },
1948
+ {
1949
+ "epoch": 18.93,
1950
+ "eval_loss": 0.8545425534248352,
1951
+ "eval_runtime": 254.9463,
1952
+ "eval_samples_per_second": 8.033,
1953
+ "eval_wer": 0.5970358501902664,
1954
+ "step": 8896
1955
+ },
1956
+ {
1957
+ "epoch": 19.06,
1958
+ "learning_rate": 9.900990099009901e-06,
1959
+ "loss": 0.1094,
1960
+ "step": 8960
1961
+ },
1962
+ {
1963
+ "epoch": 19.06,
1964
+ "eval_loss": 0.8487824201583862,
1965
+ "eval_runtime": 255.9015,
1966
+ "eval_samples_per_second": 8.003,
1967
+ "eval_wer": 0.5928967220775753,
1968
+ "step": 8960
1969
+ },
1970
+ {
1971
+ "epoch": 19.2,
1972
+ "learning_rate": 8.46084608460846e-06,
1973
+ "loss": 0.1023,
1974
+ "step": 9024
1975
+ },
1976
+ {
1977
+ "epoch": 19.2,
1978
+ "eval_loss": 0.8481884002685547,
1979
+ "eval_runtime": 255.378,
1980
+ "eval_samples_per_second": 8.019,
1981
+ "eval_wer": 0.5924961612924762,
1982
+ "step": 9024
1983
+ },
1984
+ {
1985
+ "epoch": 19.34,
1986
+ "learning_rate": 7.02070207020702e-06,
1987
+ "loss": 0.0896,
1988
+ "step": 9088
1989
+ },
1990
+ {
1991
+ "epoch": 19.34,
1992
+ "eval_loss": 0.8508026003837585,
1993
+ "eval_runtime": 258.1213,
1994
+ "eval_samples_per_second": 7.934,
1995
+ "eval_wer": 0.5942319246945724,
1996
+ "step": 9088
1997
+ },
1998
+ {
1999
+ "epoch": 19.47,
2000
+ "learning_rate": 5.58055805580558e-06,
2001
+ "loss": 0.1055,
2002
+ "step": 9152
2003
+ },
2004
+ {
2005
+ "epoch": 19.47,
2006
+ "eval_loss": 0.8545140027999878,
2007
+ "eval_runtime": 255.3836,
2008
+ "eval_samples_per_second": 8.019,
2009
+ "eval_wer": 0.5957674077041191,
2010
+ "step": 9152
2011
+ },
2012
+ {
2013
+ "epoch": 19.61,
2014
+ "learning_rate": 4.14041404140414e-06,
2015
+ "loss": 0.1102,
2016
+ "step": 9216
2017
+ },
2018
+ {
2019
+ "epoch": 19.61,
2020
+ "eval_loss": 0.8519406318664551,
2021
+ "eval_runtime": 255.7976,
2022
+ "eval_samples_per_second": 8.006,
2023
+ "eval_wer": 0.5952333266573202,
2024
+ "step": 9216
2025
+ },
2026
+ {
2027
+ "epoch": 19.74,
2028
+ "learning_rate": 2.7002700270027004e-06,
2029
+ "loss": 0.1021,
2030
+ "step": 9280
2031
+ },
2032
+ {
2033
+ "epoch": 19.74,
2034
+ "eval_loss": 0.8542845249176025,
2035
+ "eval_runtime": 255.9288,
2036
+ "eval_samples_per_second": 8.002,
2037
+ "eval_wer": 0.5940984044328727,
2038
+ "step": 9280
2039
+ },
2040
+ {
2041
+ "epoch": 19.88,
2042
+ "learning_rate": 1.2601260126012601e-06,
2043
+ "loss": 0.0934,
2044
+ "step": 9344
2045
+ },
2046
+ {
2047
+ "epoch": 19.88,
2048
+ "eval_loss": 0.8573585748672485,
2049
+ "eval_runtime": 256.1564,
2050
+ "eval_samples_per_second": 7.995,
2051
+ "eval_wer": 0.5932305227318245,
2052
+ "step": 9344
2053
+ },
2054
+ {
2055
+ "epoch": 20.0,
2056
+ "step": 9400,
2057
+ "total_flos": 5.3055511592602894e+19,
2058
+ "train_runtime": 80232.1889,
2059
+ "train_samples_per_second": 0.117
2060
+ }
2061
+ ],
2062
+ "max_steps": 9400,
2063
+ "num_train_epochs": 20,
2064
+ "total_flos": 5.3055511592602894e+19,
2065
+ "trial_name": null,
2066
+ "trial_params": null
2067
+ }
vocab.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"q": 0, "ê": 1, "d": 2, "ç": 3, "е": 4, "o": 5, "ö": 6, "©": 7, "u": 8, "&": 9, "y": 10, "ā": 11, "í": 12, "ô": 13, "l": 14, "с": 15, "k": 16, "â": 17, "f": 18, "a": 19, "z": 20, "½": 21, "s": 22, "è": 23, "p": 24, "á": 25, "w": 26, "c": 27, "t": 28, "i": 29, "™": 30, "а": 31, "і": 32, "x": 33, "à": 34, "r": 35, "e": 36, "å": 37, "€": 38, "b": 39, "$": 40, "ñ": 41, "v": 42, "ã": 43, "é": 44, "ü": 45, "ï": 46, "в": 47, "h": 48, "g": 50, "j": 51, "ú": 52, "m": 53, "ë": 54, "¼": 55, "ğ": 56, "ō": 57, "+": 58, "°": 59, "n": 60, "|": 49, "[UNK]": 61, "[PAD]": 62}