Hugging Face's logo

mlengineer-ai
/
whisper-small-fa-specaug

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
whisper
Inference Endpoints
Model card Files Metrics Community
mehran commited on
Commit
93b868d
·
1 Parent(s): b9d6f85

Uploading check points

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. checkpoint-1000/config.json +51 -0
  2. checkpoint-1000/generation_config.json +221 -0
  3. checkpoint-1000/optimizer.pt +3 -0
  4. checkpoint-1000/preprocessor_config.json +14 -0
  5. checkpoint-1000/pytorch_model.bin +3 -0
  6. checkpoint-1000/rng_state.pth +3 -0
  7. checkpoint-1000/scaler.pt +3 -0
  8. checkpoint-1000/scheduler.pt +3 -0
  9. checkpoint-1000/trainer_state.json +265 -0
  10. checkpoint-1000/training_args.bin +3 -0
  11. checkpoint-10000/config.json +51 -0
  12. checkpoint-10000/generation_config.json +221 -0
  13. checkpoint-10000/optimizer.pt +3 -0
  14. checkpoint-10000/preprocessor_config.json +14 -0
  15. checkpoint-10000/pytorch_model.bin +3 -0
  16. checkpoint-10000/rng_state.pth +3 -0
  17. checkpoint-10000/scaler.pt +3 -0
  18. checkpoint-10000/scheduler.pt +3 -0
  19. checkpoint-10000/trainer_state.json +2506 -0
  20. checkpoint-10000/training_args.bin +3 -0
  21. checkpoint-11000/config.json +51 -0
  22. checkpoint-11000/generation_config.json +221 -0
  23. checkpoint-11000/optimizer.pt +3 -0
  24. checkpoint-11000/preprocessor_config.json +14 -0
  25. checkpoint-11000/pytorch_model.bin +3 -0
  26. checkpoint-11000/rng_state.pth +3 -0
  27. checkpoint-11000/scaler.pt +3 -0
  28. checkpoint-11000/scheduler.pt +3 -0
  29. checkpoint-11000/trainer_state.json +2755 -0
  30. checkpoint-11000/training_args.bin +3 -0
  31. checkpoint-12000/config.json +51 -0
  32. checkpoint-12000/generation_config.json +221 -0
  33. checkpoint-12000/optimizer.pt +3 -0
  34. checkpoint-12000/preprocessor_config.json +14 -0
  35. checkpoint-12000/pytorch_model.bin +3 -0
  36. checkpoint-12000/rng_state.pth +3 -0
  37. checkpoint-12000/scaler.pt +3 -0
  38. checkpoint-12000/scheduler.pt +3 -0
  39. checkpoint-12000/trainer_state.json +3004 -0
  40. checkpoint-12000/training_args.bin +3 -0
  41. checkpoint-13000/config.json +51 -0
  42. checkpoint-13000/generation_config.json +221 -0
  43. checkpoint-13000/optimizer.pt +3 -0
  44. checkpoint-13000/preprocessor_config.json +14 -0
  45. checkpoint-13000/pytorch_model.bin +3 -0
  46. checkpoint-13000/rng_state.pth +3 -0
  47. checkpoint-13000/scaler.pt +3 -0
  48. checkpoint-13000/scheduler.pt +3 -0
  49. checkpoint-13000/trainer_state.json +3253 -0
  50. checkpoint-13000/training_args.bin +3 -0
checkpoint-1000/config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-small",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": true,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 768,
17
+ "decoder_attention_heads": 12,
18
+ "decoder_ffn_dim": 3072,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 12,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 12,
24
+ "encoder_ffn_dim": 3072,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 12,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": null,
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "mask_feature_length": 10,
32
+ "mask_feature_min_masks": 0,
33
+ "mask_feature_prob": 0.05,
34
+ "mask_time_length": 10,
35
+ "mask_time_min_masks": 2,
36
+ "mask_time_prob": 0.05,
37
+ "max_length": 448,
38
+ "max_source_positions": 1500,
39
+ "max_target_positions": 448,
40
+ "model_type": "whisper",
41
+ "num_hidden_layers": 12,
42
+ "num_mel_bins": 80,
43
+ "pad_token_id": 50257,
44
+ "scale_embedding": false,
45
+ "suppress_tokens": [],
46
+ "torch_dtype": "float32",
47
+ "transformers_version": "4.30.0.dev0",
48
+ "use_cache": true,
49
+ "use_weighted_layer_sum": false,
50
+ "vocab_size": 51865
51
+ }
checkpoint-1000/generation_config.json ADDED
@@ -0,0 +1,221 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "begin_suppress_tokens": [
3
+ 220,
4
+ 50257
5
+ ],
6
+ "bos_token_id": 50257,
7
+ "decoder_start_token_id": 50258,
8
+ "eos_token_id": 50257,
9
+ "forced_decoder_ids": [
10
+ [
11
+ 1,
12
+ null
13
+ ],
14
+ [
15
+ 2,
16
+ 50359
17
+ ]
18
+ ],
19
+ "is_multilingual": true,
20
+ "lang_to_id": {
21
+ "<|af|>": 50327,
22
+ "<|am|>": 50334,
23
+ "<|ar|>": 50272,
24
+ "<|as|>": 50350,
25
+ "<|az|>": 50304,
26
+ "<|ba|>": 50355,
27
+ "<|be|>": 50330,
28
+ "<|bg|>": 50292,
29
+ "<|bn|>": 50302,
30
+ "<|bo|>": 50347,
31
+ "<|br|>": 50309,
32
+ "<|bs|>": 50315,
33
+ "<|ca|>": 50270,
34
+ "<|cs|>": 50283,
35
+ "<|cy|>": 50297,
36
+ "<|da|>": 50285,
37
+ "<|de|>": 50261,
38
+ "<|el|>": 50281,
39
+ "<|en|>": 50259,
40
+ "<|es|>": 50262,
41
+ "<|et|>": 50307,
42
+ "<|eu|>": 50310,
43
+ "<|fa|>": 50300,
44
+ "<|fi|>": 50277,
45
+ "<|fo|>": 50338,
46
+ "<|fr|>": 50265,
47
+ "<|gl|>": 50319,
48
+ "<|gu|>": 50333,
49
+ "<|haw|>": 50352,
50
+ "<|ha|>": 50354,
51
+ "<|he|>": 50279,
52
+ "<|hi|>": 50276,
53
+ "<|hr|>": 50291,
54
+ "<|ht|>": 50339,
55
+ "<|hu|>": 50286,
56
+ "<|hy|>": 50312,
57
+ "<|id|>": 50275,
58
+ "<|is|>": 50311,
59
+ "<|it|>": 50274,
60
+ "<|ja|>": 50266,
61
+ "<|jw|>": 50356,
62
+ "<|ka|>": 50329,
63
+ "<|kk|>": 50316,
64
+ "<|km|>": 50323,
65
+ "<|kn|>": 50306,
66
+ "<|ko|>": 50264,
67
+ "<|la|>": 50294,
68
+ "<|lb|>": 50345,
69
+ "<|ln|>": 50353,
70
+ "<|lo|>": 50336,
71
+ "<|lt|>": 50293,
72
+ "<|lv|>": 50301,
73
+ "<|mg|>": 50349,
74
+ "<|mi|>": 50295,
75
+ "<|mk|>": 50308,
76
+ "<|ml|>": 50296,
77
+ "<|mn|>": 50314,
78
+ "<|mr|>": 50320,
79
+ "<|ms|>": 50282,
80
+ "<|mt|>": 50343,
81
+ "<|my|>": 50346,
82
+ "<|ne|>": 50313,
83
+ "<|nl|>": 50271,
84
+ "<|nn|>": 50342,
85
+ "<|no|>": 50288,
86
+ "<|oc|>": 50328,
87
+ "<|pa|>": 50321,
88
+ "<|pl|>": 50269,
89
+ "<|ps|>": 50340,
90
+ "<|pt|>": 50267,
91
+ "<|ro|>": 50284,
92
+ "<|ru|>": 50263,
93
+ "<|sa|>": 50344,
94
+ "<|sd|>": 50332,
95
+ "<|si|>": 50322,
96
+ "<|sk|>": 50298,
97
+ "<|sl|>": 50305,
98
+ "<|sn|>": 50324,
99
+ "<|so|>": 50326,
100
+ "<|sq|>": 50317,
101
+ "<|sr|>": 50303,
102
+ "<|su|>": 50357,
103
+ "<|sv|>": 50273,
104
+ "<|sw|>": 50318,
105
+ "<|ta|>": 50287,
106
+ "<|te|>": 50299,
107
+ "<|tg|>": 50331,
108
+ "<|th|>": 50289,
109
+ "<|tk|>": 50341,
110
+ "<|tl|>": 50348,
111
+ "<|tr|>": 50268,
112
+ "<|tt|>": 50351,
113
+ "<|uk|>": 50280,
114
+ "<|ur|>": 50290,
115
+ "<|uz|>": 50337,
116
+ "<|vi|>": 50278,
117
+ "<|yi|>": 50335,
118
+ "<|yo|>": 50325,
119
+ "<|zh|>": 50260
120
+ },
121
+ "max_initial_timestamp_index": 1,
122
+ "max_length": 448,
123
+ "no_timestamps_token_id": 50363,
124
+ "pad_token_id": 50257,
125
+ "return_timestamps": false,
126
+ "suppress_tokens": [
127
+ 1,
128
+ 2,
129
+ 7,
130
+ 8,
131
+ 9,
132
+ 10,
133
+ 14,
134
+ 25,
135
+ 26,
136
+ 27,
137
+ 28,
138
+ 29,
139
+ 31,
140
+ 58,
141
+ 59,
142
+ 60,
143
+ 61,
144
+ 62,
145
+ 63,
146
+ 90,
147
+ 91,
148
+ 92,
149
+ 93,
150
+ 359,
151
+ 503,
152
+ 522,
153
+ 542,
154
+ 873,
155
+ 893,
156
+ 902,
157
+ 918,
158
+ 922,
159
+ 931,
160
+ 1350,
161
+ 1853,
162
+ 1982,
163
+ 2460,
164
+ 2627,
165
+ 3246,
166
+ 3253,
167
+ 3268,
168
+ 3536,
169
+ 3846,
170
+ 3961,
171
+ 4183,
172
+ 4667,
173
+ 6585,
174
+ 6647,
175
+ 7273,
176
+ 9061,
177
+ 9383,
178
+ 10428,
179
+ 10929,
180
+ 11938,
181
+ 12033,
182
+ 12331,
183
+ 12562,
184
+ 13793,
185
+ 14157,
186
+ 14635,
187
+ 15265,
188
+ 15618,
189
+ 16553,
190
+ 16604,
191
+ 18362,
192
+ 18956,
193
+ 20075,
194
+ 21675,
195
+ 22520,
196
+ 26130,
197
+ 26161,
198
+ 26435,
199
+ 28279,
200
+ 29464,
201
+ 31650,
202
+ 32302,
203
+ 32470,
204
+ 36865,
205
+ 42863,
206
+ 47425,
207
+ 49870,
208
+ 50254,
209
+ 50258,
210
+ 50358,
211
+ 50359,
212
+ 50360,
213
+ 50361,
214
+ 50362
215
+ ],
216
+ "task_to_id": {
217
+ "transcribe": 50359,
218
+ "translate": 50358
219
+ },
220
+ "transformers_version": "4.30.0.dev0"
221
+ }
checkpoint-1000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2da35eae1fa4d0f4d64dfaf13611139c7fc5071e4c3c68541849d845f43d47c2
3
+ size 1934161093
checkpoint-1000/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-1000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6412ab0b63845c764c6eb8a9c86bd92518cd9f98695d342206dea155a7b8d658
3
+ size 967102601
checkpoint-1000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0dc62a67522cc3c0c64b59a755907ab97d94448fa5d895120dc520ded0092da5
3
+ size 14639
checkpoint-1000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2292b0497ddb0554f3fc4518d7cc9b046879e509117234d9fc7434b7c01df20c
3
+ size 557
checkpoint-1000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b0888ebd79f5d0cf62353bb8847946cd9a34fe2467936dd8536d23eebe9b51c
3
+ size 627
checkpoint-1000/trainer_state.json ADDED
@@ -0,0 +1,265 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 40.657571667832,
3
+ "best_model_checkpoint": "./whisper-small-fa-aug/checkpoint-1000",
4
+ "epoch": 0.025,
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": 0.0,
12
+ "learning_rate": 4.2000000000000006e-07,
13
+ "loss": 3.8,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 9.200000000000001e-07,
19
+ "loss": 3.0497,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.0,
24
+ "learning_rate": 1.42e-06,
25
+ "loss": 2.0102,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.0,
30
+ "learning_rate": 1.9200000000000003e-06,
31
+ "loss": 1.5948,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.0,
36
+ "learning_rate": 2.42e-06,
37
+ "loss": 1.4094,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.0,
42
+ "learning_rate": 2.92e-06,
43
+ "loss": 1.2611,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.0,
48
+ "learning_rate": 3.4200000000000007e-06,
49
+ "loss": 1.1366,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.01,
54
+ "learning_rate": 3.920000000000001e-06,
55
+ "loss": 1.0817,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.01,
60
+ "learning_rate": 4.42e-06,
61
+ "loss": 1.0324,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.01,
66
+ "learning_rate": 4.92e-06,
67
+ "loss": 0.9176,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.01,
72
+ "learning_rate": 5.420000000000001e-06,
73
+ "loss": 0.7975,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.01,
78
+ "learning_rate": 5.92e-06,
79
+ "loss": 0.7079,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.01,
84
+ "learning_rate": 6.42e-06,
85
+ "loss": 0.5832,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.01,
90
+ "learning_rate": 6.92e-06,
91
+ "loss": 0.4469,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.01,
96
+ "learning_rate": 7.420000000000001e-06,
97
+ "loss": 0.3913,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.01,
102
+ "learning_rate": 7.92e-06,
103
+ "loss": 0.4189,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.01,
108
+ "learning_rate": 8.42e-06,
109
+ "loss": 0.3756,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.01,
114
+ "learning_rate": 8.920000000000001e-06,
115
+ "loss": 0.3876,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.01,
120
+ "learning_rate": 9.42e-06,
121
+ "loss": 0.3802,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.01,
126
+ "learning_rate": 9.920000000000002e-06,
127
+ "loss": 0.4061,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.01,
132
+ "learning_rate": 9.994683544303798e-06,
133
+ "loss": 0.3885,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.01,
138
+ "learning_rate": 9.988354430379748e-06,
139
+ "loss": 0.3449,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.01,
144
+ "learning_rate": 9.982025316455697e-06,
145
+ "loss": 0.3292,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.01,
150
+ "learning_rate": 9.975696202531647e-06,
151
+ "loss": 0.3439,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.02,
156
+ "learning_rate": 9.969367088607596e-06,
157
+ "loss": 0.3031,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.02,
162
+ "learning_rate": 9.963037974683545e-06,
163
+ "loss": 0.3159,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.02,
168
+ "learning_rate": 9.956708860759495e-06,
169
+ "loss": 0.3304,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.02,
174
+ "learning_rate": 9.950379746835444e-06,
175
+ "loss": 0.2918,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.02,
180
+ "learning_rate": 9.944050632911392e-06,
181
+ "loss": 0.2785,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.02,
186
+ "learning_rate": 9.937721518987341e-06,
187
+ "loss": 0.2618,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.02,
192
+ "learning_rate": 9.931392405063291e-06,
193
+ "loss": 0.2862,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.02,
198
+ "learning_rate": 9.92506329113924e-06,
199
+ "loss": 0.2962,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.02,
204
+ "learning_rate": 9.91873417721519e-06,
205
+ "loss": 0.2935,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.02,
210
+ "learning_rate": 9.912405063291141e-06,
211
+ "loss": 0.2827,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.02,
216
+ "learning_rate": 9.90607594936709e-06,
217
+ "loss": 0.2702,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 0.02,
222
+ "learning_rate": 9.89974683544304e-06,
223
+ "loss": 0.2987,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 0.02,
228
+ "learning_rate": 9.89341772151899e-06,
229
+ "loss": 0.3067,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 0.02,
234
+ "learning_rate": 9.887088607594939e-06,
235
+ "loss": 0.2962,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 0.02,
240
+ "learning_rate": 9.880759493670888e-06,
241
+ "loss": 0.2807,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 0.03,
246
+ "learning_rate": 9.874430379746836e-06,
247
+ "loss": 0.2533,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 0.03,
252
+ "eval_loss": 0.3274388611316681,
253
+ "eval_runtime": 1481.9106,
254
+ "eval_samples_per_second": 7.025,
255
+ "eval_steps_per_second": 0.439,
256
+ "eval_wer": 40.657571667832,
257
+ "step": 1000
258
+ }
259
+ ],
260
+ "max_steps": 40000,
261
+ "num_train_epochs": 9223372036854775807,
262
+ "total_flos": 4.61736640512e+18,
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 sha256:0289e1bf8fd1a33f7c3c9d3e15cc9bdd13d10e9774e0ada934c0b9d71501a993
3
+ size 4091
checkpoint-10000/config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-small",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": true,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 768,
17
+ "decoder_attention_heads": 12,
18
+ "decoder_ffn_dim": 3072,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 12,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 12,
24
+ "encoder_ffn_dim": 3072,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 12,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": null,
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "mask_feature_length": 10,
32
+ "mask_feature_min_masks": 0,
33
+ "mask_feature_prob": 0.05,
34
+ "mask_time_length": 10,
35
+ "mask_time_min_masks": 2,
36
+ "mask_time_prob": 0.05,
37
+ "max_length": 448,
38
+ "max_source_positions": 1500,
39
+ "max_target_positions": 448,
40
+ "model_type": "whisper",
41
+ "num_hidden_layers": 12,
42
+ "num_mel_bins": 80,
43
+ "pad_token_id": 50257,
44
+ "scale_embedding": false,
45
+ "suppress_tokens": [],
46
+ "torch_dtype": "float32",
47
+ "transformers_version": "4.30.0.dev0",
48
+ "use_cache": true,
49
+ "use_weighted_layer_sum": false,
50
+ "vocab_size": 51865
51
+ }
checkpoint-10000/generation_config.json ADDED
@@ -0,0 +1,221 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "begin_suppress_tokens": [
3
+ 220,
4
+ 50257
5
+ ],
6
+ "bos_token_id": 50257,
7
+ "decoder_start_token_id": 50258,
8
+ "eos_token_id": 50257,
9
+ "forced_decoder_ids": [
10
+ [
11
+ 1,
12
+ null
13
+ ],
14
+ [
15
+ 2,
16
+ 50359
17
+ ]
18
+ ],
19
+ "is_multilingual": true,
20
+ "lang_to_id": {
21
+ "<|af|>": 50327,
22
+ "<|am|>": 50334,
23
+ "<|ar|>": 50272,
24
+ "<|as|>": 50350,
25
+ "<|az|>": 50304,
26
+ "<|ba|>": 50355,
27
+ "<|be|>": 50330,
28
+ "<|bg|>": 50292,
29
+ "<|bn|>": 50302,
30
+ "<|bo|>": 50347,
31
+ "<|br|>": 50309,
32
+ "<|bs|>": 50315,
33
+ "<|ca|>": 50270,
34
+ "<|cs|>": 50283,
35
+ "<|cy|>": 50297,
36
+ "<|da|>": 50285,
37
+ "<|de|>": 50261,
38
+ "<|el|>": 50281,
39
+ "<|en|>": 50259,
40
+ "<|es|>": 50262,
41
+ "<|et|>": 50307,
42
+ "<|eu|>": 50310,
43
+ "<|fa|>": 50300,
44
+ "<|fi|>": 50277,
45
+ "<|fo|>": 50338,
46
+ "<|fr|>": 50265,
47
+ "<|gl|>": 50319,
48
+ "<|gu|>": 50333,
49
+ "<|haw|>": 50352,
50
+ "<|ha|>": 50354,
51
+ "<|he|>": 50279,
52
+ "<|hi|>": 50276,
53
+ "<|hr|>": 50291,
54
+ "<|ht|>": 50339,
55
+ "<|hu|>": 50286,
56
+ "<|hy|>": 50312,
57
+ "<|id|>": 50275,
58
+ "<|is|>": 50311,
59
+ "<|it|>": 50274,
60
+ "<|ja|>": 50266,
61
+ "<|jw|>": 50356,
62
+ "<|ka|>": 50329,
63
+ "<|kk|>": 50316,
64
+ "<|km|>": 50323,
65
+ "<|kn|>": 50306,
66
+ "<|ko|>": 50264,
67
+ "<|la|>": 50294,
68
+ "<|lb|>": 50345,
69
+ "<|ln|>": 50353,
70
+ "<|lo|>": 50336,
71
+ "<|lt|>": 50293,
72
+ "<|lv|>": 50301,
73
+ "<|mg|>": 50349,
74
+ "<|mi|>": 50295,
75
+ "<|mk|>": 50308,
76
+ "<|ml|>": 50296,
77
+ "<|mn|>": 50314,
78
+ "<|mr|>": 50320,
79
+ "<|ms|>": 50282,
80
+ "<|mt|>": 50343,
81
+ "<|my|>": 50346,
82
+ "<|ne|>": 50313,
83
+ "<|nl|>": 50271,
84
+ "<|nn|>": 50342,
85
+ "<|no|>": 50288,
86
+ "<|oc|>": 50328,
87
+ "<|pa|>": 50321,
88
+ "<|pl|>": 50269,
89
+ "<|ps|>": 50340,
90
+ "<|pt|>": 50267,
91
+ "<|ro|>": 50284,
92
+ "<|ru|>": 50263,
93
+ "<|sa|>": 50344,
94
+ "<|sd|>": 50332,
95
+ "<|si|>": 50322,
96
+ "<|sk|>": 50298,
97
+ "<|sl|>": 50305,
98
+ "<|sn|>": 50324,
99
+ "<|so|>": 50326,
100
+ "<|sq|>": 50317,
101
+ "<|sr|>": 50303,
102
+ "<|su|>": 50357,
103
+ "<|sv|>": 50273,
104
+ "<|sw|>": 50318,
105
+ "<|ta|>": 50287,
106
+ "<|te|>": 50299,
107
+ "<|tg|>": 50331,
108
+ "<|th|>": 50289,
109
+ "<|tk|>": 50341,
110
+ "<|tl|>": 50348,
111
+ "<|tr|>": 50268,
112
+ "<|tt|>": 50351,
113
+ "<|uk|>": 50280,
114
+ "<|ur|>": 50290,
115
+ "<|uz|>": 50337,
116
+ "<|vi|>": 50278,
117
+ "<|yi|>": 50335,
118
+ "<|yo|>": 50325,
119
+ "<|zh|>": 50260
120
+ },
121
+ "max_initial_timestamp_index": 1,
122
+ "max_length": 448,
123
+ "no_timestamps_token_id": 50363,
124
+ "pad_token_id": 50257,
125
+ "return_timestamps": false,
126
+ "suppress_tokens": [
127
+ 1,
128
+ 2,
129
+ 7,
130
+ 8,
131
+ 9,
132
+ 10,
133
+ 14,
134
+ 25,
135
+ 26,
136
+ 27,
137
+ 28,
138
+ 29,
139
+ 31,
140
+ 58,
141
+ 59,
142
+ 60,
143
+ 61,
144
+ 62,
145
+ 63,
146
+ 90,
147
+ 91,
148
+ 92,
149
+ 93,
150
+ 359,
151
+ 503,
152
+ 522,
153
+ 542,
154
+ 873,
155
+ 893,
156
+ 902,
157
+ 918,
158
+ 922,
159
+ 931,
160
+ 1350,
161
+ 1853,
162
+ 1982,
163
+ 2460,
164
+ 2627,
165
+ 3246,
166
+ 3253,
167
+ 3268,
168
+ 3536,
169
+ 3846,
170
+ 3961,
171
+ 4183,
172
+ 4667,
173
+ 6585,
174
+ 6647,
175
+ 7273,
176
+ 9061,
177
+ 9383,
178
+ 10428,
179
+ 10929,
180
+ 11938,
181
+ 12033,
182
+ 12331,
183
+ 12562,
184
+ 13793,
185
+ 14157,
186
+ 14635,
187
+ 15265,
188
+ 15618,
189
+ 16553,
190
+ 16604,
191
+ 18362,
192
+ 18956,
193
+ 20075,
194
+ 21675,
195
+ 22520,
196
+ 26130,
197
+ 26161,
198
+ 26435,
199
+ 28279,
200
+ 29464,
201
+ 31650,
202
+ 32302,
203
+ 32470,
204
+ 36865,
205
+ 42863,
206
+ 47425,
207
+ 49870,
208
+ 50254,
209
+ 50258,
210
+ 50358,
211
+ 50359,
212
+ 50360,
213
+ 50361,
214
+ 50362
215
+ ],
216
+ "task_to_id": {
217
+ "transcribe": 50359,
218
+ "translate": 50358
219
+ },
220
+ "transformers_version": "4.30.0.dev0"
221
+ }
checkpoint-10000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a32d9805c59193708ca3457928f4e0338c3798a0997a5d0140a2681e805eed2
3
+ size 1934161093
checkpoint-10000/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-10000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94254e2ff2be8fa9da3df84401ad9f9c5890f1a4ddc50b64f8354ec4c73f0f37
3
+ size 967102601
checkpoint-10000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49cf56c88b088b236492e7c5ae93416b8f881a5577a4e18dd7faa6b701450b61
3
+ size 14575
checkpoint-10000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54dbacd4cb4bdfc1c8e244d48f82af5f3eeb9c7ff475cb8e857c534db9234089
3
+ size 557
checkpoint-10000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f77fc232ee012a7636d6474825911122e451aa1757106e94044784b07ca6c9bf
3
+ size 627
checkpoint-10000/trainer_state.json ADDED
@@ -0,0 +1,2506 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 31.58262735979447,
3
+ "best_model_checkpoint": "./whisper-small-fa-aug/checkpoint-9000",
4
+ "epoch": 5.031125,
5
+ "global_step": 10000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 4.2000000000000006e-07,
13
+ "loss": 3.8,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 9.200000000000001e-07,
19
+ "loss": 3.0497,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.0,
24
+ "learning_rate": 1.42e-06,
25
+ "loss": 2.0102,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.0,
30
+ "learning_rate": 1.9200000000000003e-06,
31
+ "loss": 1.5948,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.0,
36
+ "learning_rate": 2.42e-06,
37
+ "loss": 1.4094,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.0,
42
+ "learning_rate": 2.92e-06,
43
+ "loss": 1.2611,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.0,
48
+ "learning_rate": 3.4200000000000007e-06,
49
+ "loss": 1.1366,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.01,
54
+ "learning_rate": 3.920000000000001e-06,
55
+ "loss": 1.0817,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.01,
60
+ "learning_rate": 4.42e-06,
61
+ "loss": 1.0324,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.01,
66
+ "learning_rate": 4.92e-06,
67
+ "loss": 0.9176,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.01,
72
+ "learning_rate": 5.420000000000001e-06,
73
+ "loss": 0.7975,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.01,
78
+ "learning_rate": 5.92e-06,
79
+ "loss": 0.7079,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.01,
84
+ "learning_rate": 6.42e-06,
85
+ "loss": 0.5832,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.01,
90
+ "learning_rate": 6.92e-06,
91
+ "loss": 0.4469,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.01,
96
+ "learning_rate": 7.420000000000001e-06,
97
+ "loss": 0.3913,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.01,
102
+ "learning_rate": 7.92e-06,
103
+ "loss": 0.4189,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.01,
108
+ "learning_rate": 8.42e-06,
109
+ "loss": 0.3756,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.01,
114
+ "learning_rate": 8.920000000000001e-06,
115
+ "loss": 0.3876,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.01,
120
+ "learning_rate": 9.42e-06,
121
+ "loss": 0.3802,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.01,
126
+ "learning_rate": 9.920000000000002e-06,
127
+ "loss": 0.4061,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.01,
132
+ "learning_rate": 9.994683544303798e-06,
133
+ "loss": 0.3885,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.01,
138
+ "learning_rate": 9.988354430379748e-06,
139
+ "loss": 0.3449,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.01,
144
+ "learning_rate": 9.982025316455697e-06,
145
+ "loss": 0.3292,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.01,
150
+ "learning_rate": 9.975696202531647e-06,
151
+ "loss": 0.3439,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.02,
156
+ "learning_rate": 9.969367088607596e-06,
157
+ "loss": 0.3031,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.02,
162
+ "learning_rate": 9.963037974683545e-06,
163
+ "loss": 0.3159,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.02,
168
+ "learning_rate": 9.956708860759495e-06,
169
+ "loss": 0.3304,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.02,
174
+ "learning_rate": 9.950379746835444e-06,
175
+ "loss": 0.2918,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.02,
180
+ "learning_rate": 9.944050632911392e-06,
181
+ "loss": 0.2785,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.02,
186
+ "learning_rate": 9.937721518987341e-06,
187
+ "loss": 0.2618,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.02,
192
+ "learning_rate": 9.931392405063291e-06,
193
+ "loss": 0.2862,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.02,
198
+ "learning_rate": 9.92506329113924e-06,
199
+ "loss": 0.2962,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.02,
204
+ "learning_rate": 9.91873417721519e-06,
205
+ "loss": 0.2935,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.02,
210
+ "learning_rate": 9.912405063291141e-06,
211
+ "loss": 0.2827,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.02,
216
+ "learning_rate": 9.90607594936709e-06,
217
+ "loss": 0.2702,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 0.02,
222
+ "learning_rate": 9.89974683544304e-06,
223
+ "loss": 0.2987,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 0.02,
228
+ "learning_rate": 9.89341772151899e-06,
229
+ "loss": 0.3067,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 0.02,
234
+ "learning_rate": 9.887088607594939e-06,
235
+ "loss": 0.2962,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 0.02,
240
+ "learning_rate": 9.880759493670888e-06,
241
+ "loss": 0.2807,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 0.03,
246
+ "learning_rate": 9.874430379746836e-06,
247
+ "loss": 0.2533,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 0.03,
252
+ "eval_loss": 0.3274388611316681,
253
+ "eval_runtime": 1481.9106,
254
+ "eval_samples_per_second": 7.025,
255
+ "eval_steps_per_second": 0.439,
256
+ "eval_wer": 40.657571667832,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 0.03,
261
+ "learning_rate": 9.868101265822785e-06,
262
+ "loss": 0.21,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 0.03,
267
+ "learning_rate": 9.861772151898735e-06,
268
+ "loss": 0.2463,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 0.03,
273
+ "learning_rate": 9.855443037974684e-06,
274
+ "loss": 0.2688,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 0.03,
279
+ "learning_rate": 9.849113924050633e-06,
280
+ "loss": 0.2293,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 0.03,
285
+ "learning_rate": 9.842784810126583e-06,
286
+ "loss": 0.2283,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 0.03,
291
+ "learning_rate": 9.836455696202532e-06,
292
+ "loss": 0.228,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 0.03,
297
+ "learning_rate": 9.830126582278482e-06,
298
+ "loss": 0.1919,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 0.03,
303
+ "learning_rate": 9.823797468354431e-06,
304
+ "loss": 0.2173,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 0.03,
309
+ "learning_rate": 9.81746835443038e-06,
310
+ "loss": 0.2066,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 0.03,
315
+ "learning_rate": 9.81113924050633e-06,
316
+ "loss": 0.1826,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 0.03,
321
+ "learning_rate": 9.80481012658228e-06,
322
+ "loss": 0.1855,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 0.03,
327
+ "learning_rate": 9.798481012658229e-06,
328
+ "loss": 0.1702,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 0.03,
333
+ "learning_rate": 9.792151898734178e-06,
334
+ "loss": 0.1714,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 0.03,
339
+ "learning_rate": 9.785822784810128e-06,
340
+ "loss": 0.1686,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 0.03,
345
+ "learning_rate": 9.779493670886077e-06,
346
+ "loss": 0.1762,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 0.04,
351
+ "learning_rate": 9.773164556962027e-06,
352
+ "loss": 0.1722,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 0.04,
357
+ "learning_rate": 9.766835443037976e-06,
358
+ "loss": 0.2051,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 0.04,
363
+ "learning_rate": 9.760506329113924e-06,
364
+ "loss": 0.2071,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 0.04,
369
+ "learning_rate": 9.754177215189873e-06,
370
+ "loss": 0.1664,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 0.04,
375
+ "learning_rate": 9.747848101265823e-06,
376
+ "loss": 0.1871,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 0.04,
381
+ "learning_rate": 9.741518987341772e-06,
382
+ "loss": 0.1893,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 0.04,
387
+ "learning_rate": 9.735189873417721e-06,
388
+ "loss": 0.1756,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 0.04,
393
+ "learning_rate": 9.728860759493671e-06,
394
+ "loss": 0.1687,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 0.04,
399
+ "learning_rate": 9.72253164556962e-06,
400
+ "loss": 0.1685,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 0.04,
405
+ "learning_rate": 9.71620253164557e-06,
406
+ "loss": 0.1805,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 0.04,
411
+ "learning_rate": 9.70987341772152e-06,
412
+ "loss": 0.1868,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 0.04,
417
+ "learning_rate": 9.70354430379747e-06,
418
+ "loss": 0.212,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 0.04,
423
+ "learning_rate": 9.69721518987342e-06,
424
+ "loss": 0.2019,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 0.04,
429
+ "learning_rate": 9.69088607594937e-06,
430
+ "loss": 0.2076,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 0.04,
435
+ "learning_rate": 9.684556962025317e-06,
436
+ "loss": 0.1971,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 1.0,
441
+ "learning_rate": 9.678227848101266e-06,
442
+ "loss": 0.2424,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 1.0,
447
+ "learning_rate": 9.671898734177216e-06,
448
+ "loss": 0.2346,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 1.0,
453
+ "learning_rate": 9.665569620253165e-06,
454
+ "loss": 0.2436,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 1.0,
459
+ "learning_rate": 9.659240506329115e-06,
460
+ "loss": 0.2613,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 1.0,
465
+ "learning_rate": 9.652911392405064e-06,
466
+ "loss": 0.2781,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 1.0,
471
+ "learning_rate": 9.646582278481013e-06,
472
+ "loss": 0.2465,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 1.0,
477
+ "learning_rate": 9.640253164556963e-06,
478
+ "loss": 0.2386,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 1.0,
483
+ "learning_rate": 9.633924050632912e-06,
484
+ "loss": 0.2298,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 1.01,
489
+ "learning_rate": 9.627594936708862e-06,
490
+ "loss": 0.264,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 1.01,
495
+ "learning_rate": 9.621265822784811e-06,
496
+ "loss": 0.2077,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 1.01,
501
+ "eval_loss": 0.25397753715515137,
502
+ "eval_runtime": 1477.9319,
503
+ "eval_samples_per_second": 7.044,
504
+ "eval_steps_per_second": 0.44,
505
+ "eval_wer": 35.12252231743606,
506
+ "step": 2000
507
+ },
508
+ {
509
+ "epoch": 1.01,
510
+ "learning_rate": 9.61493670886076e-06,
511
+ "loss": 0.192,
512
+ "step": 2025
513
+ },
514
+ {
515
+ "epoch": 1.01,
516
+ "learning_rate": 9.60860759493671e-06,
517
+ "loss": 0.1903,
518
+ "step": 2050
519
+ },
520
+ {
521
+ "epoch": 1.01,
522
+ "learning_rate": 9.60227848101266e-06,
523
+ "loss": 0.2135,
524
+ "step": 2075
525
+ },
526
+ {
527
+ "epoch": 1.01,
528
+ "learning_rate": 9.595949367088609e-06,
529
+ "loss": 0.1917,
530
+ "step": 2100
531
+ },
532
+ {
533
+ "epoch": 1.01,
534
+ "learning_rate": 9.589620253164558e-06,
535
+ "loss": 0.1579,
536
+ "step": 2125
537
+ },
538
+ {
539
+ "epoch": 1.01,
540
+ "learning_rate": 9.583291139240508e-06,
541
+ "loss": 0.1779,
542
+ "step": 2150
543
+ },
544
+ {
545
+ "epoch": 1.01,
546
+ "learning_rate": 9.576962025316457e-06,
547
+ "loss": 0.1504,
548
+ "step": 2175
549
+ },
550
+ {
551
+ "epoch": 1.01,
552
+ "learning_rate": 9.570632911392405e-06,
553
+ "loss": 0.1574,
554
+ "step": 2200
555
+ },
556
+ {
557
+ "epoch": 1.01,
558
+ "learning_rate": 9.564303797468354e-06,
559
+ "loss": 0.1757,
560
+ "step": 2225
561
+ },
562
+ {
563
+ "epoch": 1.01,
564
+ "learning_rate": 9.557974683544304e-06,
565
+ "loss": 0.1996,
566
+ "step": 2250
567
+ },
568
+ {
569
+ "epoch": 1.01,
570
+ "learning_rate": 9.551645569620253e-06,
571
+ "loss": 0.1716,
572
+ "step": 2275
573
+ },
574
+ {
575
+ "epoch": 1.01,
576
+ "learning_rate": 9.545316455696203e-06,
577
+ "loss": 0.1509,
578
+ "step": 2300
579
+ },
580
+ {
581
+ "epoch": 1.01,
582
+ "learning_rate": 9.538987341772152e-06,
583
+ "loss": 0.1402,
584
+ "step": 2325
585
+ },
586
+ {
587
+ "epoch": 1.01,
588
+ "learning_rate": 9.532658227848101e-06,
589
+ "loss": 0.1612,
590
+ "step": 2350
591
+ },
592
+ {
593
+ "epoch": 1.02,
594
+ "learning_rate": 9.526329113924051e-06,
595
+ "loss": 0.1357,
596
+ "step": 2375
597
+ },
598
+ {
599
+ "epoch": 1.02,
600
+ "learning_rate": 9.52e-06,
601
+ "loss": 0.1426,
602
+ "step": 2400
603
+ },
604
+ {
605
+ "epoch": 1.02,
606
+ "learning_rate": 9.51367088607595e-06,
607
+ "loss": 0.1644,
608
+ "step": 2425
609
+ },
610
+ {
611
+ "epoch": 1.02,
612
+ "learning_rate": 9.5073417721519e-06,
613
+ "loss": 0.1196,
614
+ "step": 2450
615
+ },
616
+ {
617
+ "epoch": 1.02,
618
+ "learning_rate": 9.501012658227849e-06,
619
+ "loss": 0.1332,
620
+ "step": 2475
621
+ },
622
+ {
623
+ "epoch": 1.02,
624
+ "learning_rate": 9.494683544303798e-06,
625
+ "loss": 0.1096,
626
+ "step": 2500
627
+ },
628
+ {
629
+ "epoch": 1.02,
630
+ "learning_rate": 9.488354430379747e-06,
631
+ "loss": 0.1263,
632
+ "step": 2525
633
+ },
634
+ {
635
+ "epoch": 1.02,
636
+ "learning_rate": 9.482025316455697e-06,
637
+ "loss": 0.1467,
638
+ "step": 2550
639
+ },
640
+ {
641
+ "epoch": 1.02,
642
+ "learning_rate": 9.475696202531646e-06,
643
+ "loss": 0.1372,
644
+ "step": 2575
645
+ },
646
+ {
647
+ "epoch": 1.02,
648
+ "learning_rate": 9.469367088607596e-06,
649
+ "loss": 0.1256,
650
+ "step": 2600
651
+ },
652
+ {
653
+ "epoch": 1.02,
654
+ "learning_rate": 9.463037974683545e-06,
655
+ "loss": 0.1319,
656
+ "step": 2625
657
+ },
658
+ {
659
+ "epoch": 1.02,
660
+ "learning_rate": 9.456708860759495e-06,
661
+ "loss": 0.1385,
662
+ "step": 2650
663
+ },
664
+ {
665
+ "epoch": 1.02,
666
+ "learning_rate": 9.450379746835444e-06,
667
+ "loss": 0.147,
668
+ "step": 2675
669
+ },
670
+ {
671
+ "epoch": 1.02,
672
+ "learning_rate": 9.444050632911393e-06,
673
+ "loss": 0.1527,
674
+ "step": 2700
675
+ },
676
+ {
677
+ "epoch": 1.02,
678
+ "learning_rate": 9.437721518987343e-06,
679
+ "loss": 0.1357,
680
+ "step": 2725
681
+ },
682
+ {
683
+ "epoch": 1.02,
684
+ "learning_rate": 9.431392405063292e-06,
685
+ "loss": 0.122,
686
+ "step": 2750
687
+ },
688
+ {
689
+ "epoch": 1.03,
690
+ "learning_rate": 9.425063291139242e-06,
691
+ "loss": 0.0918,
692
+ "step": 2775
693
+ },
694
+ {
695
+ "epoch": 1.03,
696
+ "learning_rate": 9.418734177215191e-06,
697
+ "loss": 0.1218,
698
+ "step": 2800
699
+ },
700
+ {
701
+ "epoch": 1.03,
702
+ "learning_rate": 9.41240506329114e-06,
703
+ "loss": 0.139,
704
+ "step": 2825
705
+ },
706
+ {
707
+ "epoch": 1.03,
708
+ "learning_rate": 9.40607594936709e-06,
709
+ "loss": 0.1205,
710
+ "step": 2850
711
+ },
712
+ {
713
+ "epoch": 1.03,
714
+ "learning_rate": 9.39974683544304e-06,
715
+ "loss": 0.1158,
716
+ "step": 2875
717
+ },
718
+ {
719
+ "epoch": 1.03,
720
+ "learning_rate": 9.393417721518989e-06,
721
+ "loss": 0.1156,
722
+ "step": 2900
723
+ },
724
+ {
725
+ "epoch": 1.03,
726
+ "learning_rate": 9.387088607594937e-06,
727
+ "loss": 0.1044,
728
+ "step": 2925
729
+ },
730
+ {
731
+ "epoch": 1.03,
732
+ "learning_rate": 9.380759493670886e-06,
733
+ "loss": 0.112,
734
+ "step": 2950
735
+ },
736
+ {
737
+ "epoch": 1.03,
738
+ "learning_rate": 9.374430379746835e-06,
739
+ "loss": 0.1023,
740
+ "step": 2975
741
+ },
742
+ {
743
+ "epoch": 1.03,
744
+ "learning_rate": 9.368101265822785e-06,
745
+ "loss": 0.0966,
746
+ "step": 3000
747
+ },
748
+ {
749
+ "epoch": 1.03,
750
+ "eval_loss": 0.2638459801673889,
751
+ "eval_runtime": 1500.936,
752
+ "eval_samples_per_second": 6.936,
753
+ "eval_steps_per_second": 0.434,
754
+ "eval_wer": 37.93069806013106,
755
+ "step": 3000
756
+ },
757
+ {
758
+ "epoch": 1.03,
759
+ "learning_rate": 9.361772151898734e-06,
760
+ "loss": 0.0942,
761
+ "step": 3025
762
+ },
763
+ {
764
+ "epoch": 1.03,
765
+ "learning_rate": 9.355443037974684e-06,
766
+ "loss": 0.0826,
767
+ "step": 3050
768
+ },
769
+ {
770
+ "epoch": 1.03,
771
+ "learning_rate": 9.349113924050633e-06,
772
+ "loss": 0.086,
773
+ "step": 3075
774
+ },
775
+ {
776
+ "epoch": 1.03,
777
+ "learning_rate": 9.342784810126583e-06,
778
+ "loss": 0.0956,
779
+ "step": 3100
780
+ },
781
+ {
782
+ "epoch": 1.03,
783
+ "learning_rate": 9.336455696202532e-06,
784
+ "loss": 0.0935,
785
+ "step": 3125
786
+ },
787
+ {
788
+ "epoch": 1.03,
789
+ "learning_rate": 9.330126582278481e-06,
790
+ "loss": 0.0923,
791
+ "step": 3150
792
+ },
793
+ {
794
+ "epoch": 1.04,
795
+ "learning_rate": 9.323797468354431e-06,
796
+ "loss": 0.1163,
797
+ "step": 3175
798
+ },
799
+ {
800
+ "epoch": 1.04,
801
+ "learning_rate": 9.31746835443038e-06,
802
+ "loss": 0.1099,
803
+ "step": 3200
804
+ },
805
+ {
806
+ "epoch": 1.04,
807
+ "learning_rate": 9.31113924050633e-06,
808
+ "loss": 0.0879,
809
+ "step": 3225
810
+ },
811
+ {
812
+ "epoch": 1.04,
813
+ "learning_rate": 9.30481012658228e-06,
814
+ "loss": 0.0994,
815
+ "step": 3250
816
+ },
817
+ {
818
+ "epoch": 1.04,
819
+ "learning_rate": 9.298481012658229e-06,
820
+ "loss": 0.1069,
821
+ "step": 3275
822
+ },
823
+ {
824
+ "epoch": 1.04,
825
+ "learning_rate": 9.292151898734178e-06,
826
+ "loss": 0.0957,
827
+ "step": 3300
828
+ },
829
+ {
830
+ "epoch": 1.04,
831
+ "learning_rate": 9.285822784810127e-06,
832
+ "loss": 0.0866,
833
+ "step": 3325
834
+ },
835
+ {
836
+ "epoch": 1.04,
837
+ "learning_rate": 9.279493670886077e-06,
838
+ "loss": 0.0965,
839
+ "step": 3350
840
+ },
841
+ {
842
+ "epoch": 1.04,
843
+ "learning_rate": 9.273164556962026e-06,
844
+ "loss": 0.0952,
845
+ "step": 3375
846
+ },
847
+ {
848
+ "epoch": 1.04,
849
+ "learning_rate": 9.266835443037976e-06,
850
+ "loss": 0.1178,
851
+ "step": 3400
852
+ },
853
+ {
854
+ "epoch": 1.04,
855
+ "learning_rate": 9.260506329113925e-06,
856
+ "loss": 0.1154,
857
+ "step": 3425
858
+ },
859
+ {
860
+ "epoch": 1.04,
861
+ "learning_rate": 9.254177215189875e-06,
862
+ "loss": 0.1075,
863
+ "step": 3450
864
+ },
865
+ {
866
+ "epoch": 1.04,
867
+ "learning_rate": 9.247848101265824e-06,
868
+ "loss": 0.1161,
869
+ "step": 3475
870
+ },
871
+ {
872
+ "epoch": 1.04,
873
+ "learning_rate": 9.241518987341773e-06,
874
+ "loss": 0.1153,
875
+ "step": 3500
876
+ },
877
+ {
878
+ "epoch": 2.0,
879
+ "learning_rate": 9.235189873417723e-06,
880
+ "loss": 0.1343,
881
+ "step": 3525
882
+ },
883
+ {
884
+ "epoch": 2.0,
885
+ "learning_rate": 9.228860759493672e-06,
886
+ "loss": 0.1374,
887
+ "step": 3550
888
+ },
889
+ {
890
+ "epoch": 2.0,
891
+ "learning_rate": 9.222531645569622e-06,
892
+ "loss": 0.1436,
893
+ "step": 3575
894
+ },
895
+ {
896
+ "epoch": 2.0,
897
+ "learning_rate": 9.216202531645571e-06,
898
+ "loss": 0.1561,
899
+ "step": 3600
900
+ },
901
+ {
902
+ "epoch": 2.0,
903
+ "learning_rate": 9.20987341772152e-06,
904
+ "loss": 0.1709,
905
+ "step": 3625
906
+ },
907
+ {
908
+ "epoch": 2.0,
909
+ "learning_rate": 9.20354430379747e-06,
910
+ "loss": 0.142,
911
+ "step": 3650
912
+ },
913
+ {
914
+ "epoch": 2.0,
915
+ "learning_rate": 9.197215189873418e-06,
916
+ "loss": 0.1389,
917
+ "step": 3675
918
+ },
919
+ {
920
+ "epoch": 2.0,
921
+ "learning_rate": 9.190886075949367e-06,
922
+ "loss": 0.1398,
923
+ "step": 3700
924
+ },
925
+ {
926
+ "epoch": 2.01,
927
+ "learning_rate": 9.184556962025317e-06,
928
+ "loss": 0.1526,
929
+ "step": 3725
930
+ },
931
+ {
932
+ "epoch": 2.01,
933
+ "learning_rate": 9.178227848101266e-06,
934
+ "loss": 0.1273,
935
+ "step": 3750
936
+ },
937
+ {
938
+ "epoch": 2.01,
939
+ "learning_rate": 9.171898734177215e-06,
940
+ "loss": 0.1119,
941
+ "step": 3775
942
+ },
943
+ {
944
+ "epoch": 2.01,
945
+ "learning_rate": 9.165569620253165e-06,
946
+ "loss": 0.1088,
947
+ "step": 3800
948
+ },
949
+ {
950
+ "epoch": 2.01,
951
+ "learning_rate": 9.159493670886076e-06,
952
+ "loss": 0.1362,
953
+ "step": 3825
954
+ },
955
+ {
956
+ "epoch": 2.01,
957
+ "learning_rate": 9.153164556962026e-06,
958
+ "loss": 0.1247,
959
+ "step": 3850
960
+ },
961
+ {
962
+ "epoch": 2.01,
963
+ "learning_rate": 9.146835443037975e-06,
964
+ "loss": 0.0974,
965
+ "step": 3875
966
+ },
967
+ {
968
+ "epoch": 2.01,
969
+ "learning_rate": 9.140506329113925e-06,
970
+ "loss": 0.1218,
971
+ "step": 3900
972
+ },
973
+ {
974
+ "epoch": 2.01,
975
+ "learning_rate": 9.134177215189874e-06,
976
+ "loss": 0.0846,
977
+ "step": 3925
978
+ },
979
+ {
980
+ "epoch": 2.01,
981
+ "learning_rate": 9.127848101265823e-06,
982
+ "loss": 0.0964,
983
+ "step": 3950
984
+ },
985
+ {
986
+ "epoch": 2.01,
987
+ "learning_rate": 9.121518987341773e-06,
988
+ "loss": 0.1166,
989
+ "step": 3975
990
+ },
991
+ {
992
+ "epoch": 2.01,
993
+ "learning_rate": 9.115189873417722e-06,
994
+ "loss": 0.1212,
995
+ "step": 4000
996
+ },
997
+ {
998
+ "epoch": 2.01,
999
+ "eval_loss": 0.22517943382263184,
1000
+ "eval_runtime": 1537.556,
1001
+ "eval_samples_per_second": 6.771,
1002
+ "eval_steps_per_second": 0.423,
1003
+ "eval_wer": 69.04990325045935,
1004
+ "step": 4000
1005
+ },
1006
+ {
1007
+ "epoch": 2.01,
1008
+ "learning_rate": 9.108860759493672e-06,
1009
+ "loss": 0.1076,
1010
+ "step": 4025
1011
+ },
1012
+ {
1013
+ "epoch": 2.01,
1014
+ "learning_rate": 9.102531645569621e-06,
1015
+ "loss": 0.086,
1016
+ "step": 4050
1017
+ },
1018
+ {
1019
+ "epoch": 2.01,
1020
+ "learning_rate": 9.09620253164557e-06,
1021
+ "loss": 0.0773,
1022
+ "step": 4075
1023
+ },
1024
+ {
1025
+ "epoch": 2.01,
1026
+ "learning_rate": 9.08987341772152e-06,
1027
+ "loss": 0.098,
1028
+ "step": 4100
1029
+ },
1030
+ {
1031
+ "epoch": 2.02,
1032
+ "learning_rate": 9.08354430379747e-06,
1033
+ "loss": 0.0747,
1034
+ "step": 4125
1035
+ },
1036
+ {
1037
+ "epoch": 2.02,
1038
+ "learning_rate": 9.077215189873419e-06,
1039
+ "loss": 0.0895,
1040
+ "step": 4150
1041
+ },
1042
+ {
1043
+ "epoch": 2.02,
1044
+ "learning_rate": 9.070886075949368e-06,
1045
+ "loss": 0.0957,
1046
+ "step": 4175
1047
+ },
1048
+ {
1049
+ "epoch": 2.02,
1050
+ "learning_rate": 9.064556962025318e-06,
1051
+ "loss": 0.0789,
1052
+ "step": 4200
1053
+ },
1054
+ {
1055
+ "epoch": 2.02,
1056
+ "learning_rate": 9.058227848101267e-06,
1057
+ "loss": 0.0718,
1058
+ "step": 4225
1059
+ },
1060
+ {
1061
+ "epoch": 2.02,
1062
+ "learning_rate": 9.051898734177215e-06,
1063
+ "loss": 0.0737,
1064
+ "step": 4250
1065
+ },
1066
+ {
1067
+ "epoch": 2.02,
1068
+ "learning_rate": 9.045569620253164e-06,
1069
+ "loss": 0.0731,
1070
+ "step": 4275
1071
+ },
1072
+ {
1073
+ "epoch": 2.02,
1074
+ "learning_rate": 9.039240506329114e-06,
1075
+ "loss": 0.087,
1076
+ "step": 4300
1077
+ },
1078
+ {
1079
+ "epoch": 2.02,
1080
+ "learning_rate": 9.032911392405063e-06,
1081
+ "loss": 0.0875,
1082
+ "step": 4325
1083
+ },
1084
+ {
1085
+ "epoch": 2.02,
1086
+ "learning_rate": 9.026582278481013e-06,
1087
+ "loss": 0.0768,
1088
+ "step": 4350
1089
+ },
1090
+ {
1091
+ "epoch": 2.02,
1092
+ "learning_rate": 9.020253164556962e-06,
1093
+ "loss": 0.0842,
1094
+ "step": 4375
1095
+ },
1096
+ {
1097
+ "epoch": 2.02,
1098
+ "learning_rate": 9.013924050632911e-06,
1099
+ "loss": 0.092,
1100
+ "step": 4400
1101
+ },
1102
+ {
1103
+ "epoch": 2.02,
1104
+ "learning_rate": 9.007594936708861e-06,
1105
+ "loss": 0.0949,
1106
+ "step": 4425
1107
+ },
1108
+ {
1109
+ "epoch": 2.02,
1110
+ "learning_rate": 9.00126582278481e-06,
1111
+ "loss": 0.0989,
1112
+ "step": 4450
1113
+ },
1114
+ {
1115
+ "epoch": 2.02,
1116
+ "learning_rate": 8.99493670886076e-06,
1117
+ "loss": 0.0908,
1118
+ "step": 4475
1119
+ },
1120
+ {
1121
+ "epoch": 2.02,
1122
+ "learning_rate": 8.98860759493671e-06,
1123
+ "loss": 0.0783,
1124
+ "step": 4500
1125
+ },
1126
+ {
1127
+ "epoch": 2.03,
1128
+ "learning_rate": 8.982278481012659e-06,
1129
+ "loss": 0.0651,
1130
+ "step": 4525
1131
+ },
1132
+ {
1133
+ "epoch": 2.03,
1134
+ "learning_rate": 8.975949367088608e-06,
1135
+ "loss": 0.0791,
1136
+ "step": 4550
1137
+ },
1138
+ {
1139
+ "epoch": 2.03,
1140
+ "learning_rate": 8.969620253164557e-06,
1141
+ "loss": 0.0829,
1142
+ "step": 4575
1143
+ },
1144
+ {
1145
+ "epoch": 2.03,
1146
+ "learning_rate": 8.963291139240507e-06,
1147
+ "loss": 0.0797,
1148
+ "step": 4600
1149
+ },
1150
+ {
1151
+ "epoch": 2.03,
1152
+ "learning_rate": 8.956962025316456e-06,
1153
+ "loss": 0.0743,
1154
+ "step": 4625
1155
+ },
1156
+ {
1157
+ "epoch": 2.03,
1158
+ "learning_rate": 8.950632911392406e-06,
1159
+ "loss": 0.0814,
1160
+ "step": 4650
1161
+ },
1162
+ {
1163
+ "epoch": 2.03,
1164
+ "learning_rate": 8.944303797468355e-06,
1165
+ "loss": 0.0594,
1166
+ "step": 4675
1167
+ },
1168
+ {
1169
+ "epoch": 2.03,
1170
+ "learning_rate": 8.937974683544305e-06,
1171
+ "loss": 0.0709,
1172
+ "step": 4700
1173
+ },
1174
+ {
1175
+ "epoch": 2.03,
1176
+ "learning_rate": 8.931645569620254e-06,
1177
+ "loss": 0.0621,
1178
+ "step": 4725
1179
+ },
1180
+ {
1181
+ "epoch": 2.03,
1182
+ "learning_rate": 8.925316455696203e-06,
1183
+ "loss": 0.0599,
1184
+ "step": 4750
1185
+ },
1186
+ {
1187
+ "epoch": 2.03,
1188
+ "learning_rate": 8.918987341772153e-06,
1189
+ "loss": 0.0625,
1190
+ "step": 4775
1191
+ },
1192
+ {
1193
+ "epoch": 2.03,
1194
+ "learning_rate": 8.912658227848102e-06,
1195
+ "loss": 0.0541,
1196
+ "step": 4800
1197
+ },
1198
+ {
1199
+ "epoch": 2.03,
1200
+ "learning_rate": 8.906329113924052e-06,
1201
+ "loss": 0.059,
1202
+ "step": 4825
1203
+ },
1204
+ {
1205
+ "epoch": 2.03,
1206
+ "learning_rate": 8.900000000000001e-06,
1207
+ "loss": 0.0576,
1208
+ "step": 4850
1209
+ },
1210
+ {
1211
+ "epoch": 2.03,
1212
+ "learning_rate": 8.89367088607595e-06,
1213
+ "loss": 0.0623,
1214
+ "step": 4875
1215
+ },
1216
+ {
1217
+ "epoch": 2.03,
1218
+ "learning_rate": 8.8873417721519e-06,
1219
+ "loss": 0.0592,
1220
+ "step": 4900
1221
+ },
1222
+ {
1223
+ "epoch": 2.04,
1224
+ "learning_rate": 8.88101265822785e-06,
1225
+ "loss": 0.0692,
1226
+ "step": 4925
1227
+ },
1228
+ {
1229
+ "epoch": 2.04,
1230
+ "learning_rate": 8.874683544303799e-06,
1231
+ "loss": 0.077,
1232
+ "step": 4950
1233
+ },
1234
+ {
1235
+ "epoch": 2.04,
1236
+ "learning_rate": 8.868354430379748e-06,
1237
+ "loss": 0.0662,
1238
+ "step": 4975
1239
+ },
1240
+ {
1241
+ "epoch": 2.04,
1242
+ "learning_rate": 8.862025316455696e-06,
1243
+ "loss": 0.0605,
1244
+ "step": 5000
1245
+ },
1246
+ {
1247
+ "epoch": 2.04,
1248
+ "eval_loss": 0.2468421906232834,
1249
+ "eval_runtime": 1442.9116,
1250
+ "eval_samples_per_second": 7.215,
1251
+ "eval_steps_per_second": 0.451,
1252
+ "eval_wer": 33.46883689165677,
1253
+ "step": 5000
1254
+ },
1255
+ {
1256
+ "epoch": 2.04,
1257
+ "learning_rate": 8.855696202531645e-06,
1258
+ "loss": 0.0769,
1259
+ "step": 5025
1260
+ },
1261
+ {
1262
+ "epoch": 2.04,
1263
+ "learning_rate": 8.849367088607595e-06,
1264
+ "loss": 0.0601,
1265
+ "step": 5050
1266
+ },
1267
+ {
1268
+ "epoch": 2.04,
1269
+ "learning_rate": 8.843037974683544e-06,
1270
+ "loss": 0.0698,
1271
+ "step": 5075
1272
+ },
1273
+ {
1274
+ "epoch": 2.04,
1275
+ "learning_rate": 8.836708860759494e-06,
1276
+ "loss": 0.0593,
1277
+ "step": 5100
1278
+ },
1279
+ {
1280
+ "epoch": 2.04,
1281
+ "learning_rate": 8.830379746835443e-06,
1282
+ "loss": 0.0521,
1283
+ "step": 5125
1284
+ },
1285
+ {
1286
+ "epoch": 2.04,
1287
+ "learning_rate": 8.824050632911393e-06,
1288
+ "loss": 0.0687,
1289
+ "step": 5150
1290
+ },
1291
+ {
1292
+ "epoch": 2.04,
1293
+ "learning_rate": 8.817721518987342e-06,
1294
+ "loss": 0.0818,
1295
+ "step": 5175
1296
+ },
1297
+ {
1298
+ "epoch": 2.04,
1299
+ "learning_rate": 8.811392405063291e-06,
1300
+ "loss": 0.0706,
1301
+ "step": 5200
1302
+ },
1303
+ {
1304
+ "epoch": 2.04,
1305
+ "learning_rate": 8.805063291139241e-06,
1306
+ "loss": 0.0734,
1307
+ "step": 5225
1308
+ },
1309
+ {
1310
+ "epoch": 2.04,
1311
+ "learning_rate": 8.79873417721519e-06,
1312
+ "loss": 0.0795,
1313
+ "step": 5250
1314
+ },
1315
+ {
1316
+ "epoch": 3.0,
1317
+ "learning_rate": 8.79240506329114e-06,
1318
+ "loss": 0.0905,
1319
+ "step": 5275
1320
+ },
1321
+ {
1322
+ "epoch": 3.0,
1323
+ "learning_rate": 8.78607594936709e-06,
1324
+ "loss": 0.086,
1325
+ "step": 5300
1326
+ },
1327
+ {
1328
+ "epoch": 3.0,
1329
+ "learning_rate": 8.779746835443039e-06,
1330
+ "loss": 0.105,
1331
+ "step": 5325
1332
+ },
1333
+ {
1334
+ "epoch": 3.0,
1335
+ "learning_rate": 8.773417721518988e-06,
1336
+ "loss": 0.0981,
1337
+ "step": 5350
1338
+ },
1339
+ {
1340
+ "epoch": 3.0,
1341
+ "learning_rate": 8.767088607594937e-06,
1342
+ "loss": 0.1131,
1343
+ "step": 5375
1344
+ },
1345
+ {
1346
+ "epoch": 3.0,
1347
+ "learning_rate": 8.760759493670887e-06,
1348
+ "loss": 0.0916,
1349
+ "step": 5400
1350
+ },
1351
+ {
1352
+ "epoch": 3.0,
1353
+ "learning_rate": 8.754430379746836e-06,
1354
+ "loss": 0.0974,
1355
+ "step": 5425
1356
+ },
1357
+ {
1358
+ "epoch": 3.0,
1359
+ "learning_rate": 8.748101265822786e-06,
1360
+ "loss": 0.0899,
1361
+ "step": 5450
1362
+ },
1363
+ {
1364
+ "epoch": 3.01,
1365
+ "learning_rate": 8.741772151898735e-06,
1366
+ "loss": 0.1104,
1367
+ "step": 5475
1368
+ },
1369
+ {
1370
+ "epoch": 3.01,
1371
+ "learning_rate": 8.735443037974685e-06,
1372
+ "loss": 0.0818,
1373
+ "step": 5500
1374
+ },
1375
+ {
1376
+ "epoch": 3.01,
1377
+ "learning_rate": 8.729113924050634e-06,
1378
+ "loss": 0.0764,
1379
+ "step": 5525
1380
+ },
1381
+ {
1382
+ "epoch": 3.01,
1383
+ "learning_rate": 8.722784810126583e-06,
1384
+ "loss": 0.0709,
1385
+ "step": 5550
1386
+ },
1387
+ {
1388
+ "epoch": 3.01,
1389
+ "learning_rate": 8.716455696202533e-06,
1390
+ "loss": 0.0891,
1391
+ "step": 5575
1392
+ },
1393
+ {
1394
+ "epoch": 3.01,
1395
+ "learning_rate": 8.710126582278482e-06,
1396
+ "loss": 0.0869,
1397
+ "step": 5600
1398
+ },
1399
+ {
1400
+ "epoch": 3.01,
1401
+ "learning_rate": 8.703797468354432e-06,
1402
+ "loss": 0.0631,
1403
+ "step": 5625
1404
+ },
1405
+ {
1406
+ "epoch": 3.01,
1407
+ "learning_rate": 8.697468354430381e-06,
1408
+ "loss": 0.0795,
1409
+ "step": 5650
1410
+ },
1411
+ {
1412
+ "epoch": 3.01,
1413
+ "learning_rate": 8.69113924050633e-06,
1414
+ "loss": 0.0603,
1415
+ "step": 5675
1416
+ },
1417
+ {
1418
+ "epoch": 3.01,
1419
+ "learning_rate": 8.68481012658228e-06,
1420
+ "loss": 0.064,
1421
+ "step": 5700
1422
+ },
1423
+ {
1424
+ "epoch": 3.01,
1425
+ "learning_rate": 8.67848101265823e-06,
1426
+ "loss": 0.0738,
1427
+ "step": 5725
1428
+ },
1429
+ {
1430
+ "epoch": 3.01,
1431
+ "learning_rate": 8.672151898734177e-06,
1432
+ "loss": 0.0796,
1433
+ "step": 5750
1434
+ },
1435
+ {
1436
+ "epoch": 3.01,
1437
+ "learning_rate": 8.665822784810127e-06,
1438
+ "loss": 0.0723,
1439
+ "step": 5775
1440
+ },
1441
+ {
1442
+ "epoch": 3.01,
1443
+ "learning_rate": 8.659493670886076e-06,
1444
+ "loss": 0.0592,
1445
+ "step": 5800
1446
+ },
1447
+ {
1448
+ "epoch": 3.01,
1449
+ "learning_rate": 8.653164556962025e-06,
1450
+ "loss": 0.051,
1451
+ "step": 5825
1452
+ },
1453
+ {
1454
+ "epoch": 3.01,
1455
+ "learning_rate": 8.646835443037975e-06,
1456
+ "loss": 0.0716,
1457
+ "step": 5850
1458
+ },
1459
+ {
1460
+ "epoch": 3.02,
1461
+ "learning_rate": 8.640506329113924e-06,
1462
+ "loss": 0.0598,
1463
+ "step": 5875
1464
+ },
1465
+ {
1466
+ "epoch": 3.02,
1467
+ "learning_rate": 8.634177215189874e-06,
1468
+ "loss": 0.0524,
1469
+ "step": 5900
1470
+ },
1471
+ {
1472
+ "epoch": 3.02,
1473
+ "learning_rate": 8.627848101265823e-06,
1474
+ "loss": 0.0623,
1475
+ "step": 5925
1476
+ },
1477
+ {
1478
+ "epoch": 3.02,
1479
+ "learning_rate": 8.621518987341773e-06,
1480
+ "loss": 0.054,
1481
+ "step": 5950
1482
+ },
1483
+ {
1484
+ "epoch": 3.02,
1485
+ "learning_rate": 8.615189873417722e-06,
1486
+ "loss": 0.057,
1487
+ "step": 5975
1488
+ },
1489
+ {
1490
+ "epoch": 3.02,
1491
+ "learning_rate": 8.608860759493671e-06,
1492
+ "loss": 0.0479,
1493
+ "step": 6000
1494
+ },
1495
+ {
1496
+ "epoch": 3.02,
1497
+ "eval_loss": 0.24428367614746094,
1498
+ "eval_runtime": 1457.5568,
1499
+ "eval_samples_per_second": 7.143,
1500
+ "eval_steps_per_second": 0.447,
1501
+ "eval_wer": 36.51929299663409,
1502
+ "step": 6000
1503
+ },
1504
+ {
1505
+ "epoch": 3.02,
1506
+ "learning_rate": 8.602531645569621e-06,
1507
+ "loss": 0.0501,
1508
+ "step": 6025
1509
+ },
1510
+ {
1511
+ "epoch": 3.02,
1512
+ "learning_rate": 8.59620253164557e-06,
1513
+ "loss": 0.0527,
1514
+ "step": 6050
1515
+ },
1516
+ {
1517
+ "epoch": 3.02,
1518
+ "learning_rate": 8.58987341772152e-06,
1519
+ "loss": 0.0558,
1520
+ "step": 6075
1521
+ },
1522
+ {
1523
+ "epoch": 3.02,
1524
+ "learning_rate": 8.58354430379747e-06,
1525
+ "loss": 0.0548,
1526
+ "step": 6100
1527
+ },
1528
+ {
1529
+ "epoch": 3.02,
1530
+ "learning_rate": 8.577215189873419e-06,
1531
+ "loss": 0.0503,
1532
+ "step": 6125
1533
+ },
1534
+ {
1535
+ "epoch": 3.02,
1536
+ "learning_rate": 8.570886075949368e-06,
1537
+ "loss": 0.0641,
1538
+ "step": 6150
1539
+ },
1540
+ {
1541
+ "epoch": 3.02,
1542
+ "learning_rate": 8.564556962025316e-06,
1543
+ "loss": 0.0748,
1544
+ "step": 6175
1545
+ },
1546
+ {
1547
+ "epoch": 3.02,
1548
+ "learning_rate": 8.558227848101267e-06,
1549
+ "loss": 0.0775,
1550
+ "step": 6200
1551
+ },
1552
+ {
1553
+ "epoch": 3.02,
1554
+ "learning_rate": 8.551898734177216e-06,
1555
+ "loss": 0.0662,
1556
+ "step": 6225
1557
+ },
1558
+ {
1559
+ "epoch": 3.02,
1560
+ "learning_rate": 8.545569620253166e-06,
1561
+ "loss": 0.0591,
1562
+ "step": 6250
1563
+ },
1564
+ {
1565
+ "epoch": 3.03,
1566
+ "learning_rate": 8.539240506329115e-06,
1567
+ "loss": 0.0421,
1568
+ "step": 6275
1569
+ },
1570
+ {
1571
+ "epoch": 3.03,
1572
+ "learning_rate": 8.532911392405065e-06,
1573
+ "loss": 0.0565,
1574
+ "step": 6300
1575
+ },
1576
+ {
1577
+ "epoch": 3.03,
1578
+ "learning_rate": 8.526582278481014e-06,
1579
+ "loss": 0.0574,
1580
+ "step": 6325
1581
+ },
1582
+ {
1583
+ "epoch": 3.03,
1584
+ "learning_rate": 8.520253164556963e-06,
1585
+ "loss": 0.0466,
1586
+ "step": 6350
1587
+ },
1588
+ {
1589
+ "epoch": 3.03,
1590
+ "learning_rate": 8.513924050632913e-06,
1591
+ "loss": 0.0486,
1592
+ "step": 6375
1593
+ },
1594
+ {
1595
+ "epoch": 3.03,
1596
+ "learning_rate": 8.507594936708862e-06,
1597
+ "loss": 0.0502,
1598
+ "step": 6400
1599
+ },
1600
+ {
1601
+ "epoch": 3.03,
1602
+ "learning_rate": 8.501265822784812e-06,
1603
+ "loss": 0.0415,
1604
+ "step": 6425
1605
+ },
1606
+ {
1607
+ "epoch": 3.03,
1608
+ "learning_rate": 8.494936708860761e-06,
1609
+ "loss": 0.0449,
1610
+ "step": 6450
1611
+ },
1612
+ {
1613
+ "epoch": 3.03,
1614
+ "learning_rate": 8.488607594936709e-06,
1615
+ "loss": 0.0525,
1616
+ "step": 6475
1617
+ },
1618
+ {
1619
+ "epoch": 3.03,
1620
+ "learning_rate": 8.482278481012658e-06,
1621
+ "loss": 0.0401,
1622
+ "step": 6500
1623
+ },
1624
+ {
1625
+ "epoch": 3.03,
1626
+ "learning_rate": 8.475949367088608e-06,
1627
+ "loss": 0.0502,
1628
+ "step": 6525
1629
+ },
1630
+ {
1631
+ "epoch": 3.03,
1632
+ "learning_rate": 8.469620253164557e-06,
1633
+ "loss": 0.0453,
1634
+ "step": 6550
1635
+ },
1636
+ {
1637
+ "epoch": 3.03,
1638
+ "learning_rate": 8.463291139240507e-06,
1639
+ "loss": 0.0442,
1640
+ "step": 6575
1641
+ },
1642
+ {
1643
+ "epoch": 3.03,
1644
+ "learning_rate": 8.456962025316456e-06,
1645
+ "loss": 0.0423,
1646
+ "step": 6600
1647
+ },
1648
+ {
1649
+ "epoch": 3.03,
1650
+ "learning_rate": 8.450632911392405e-06,
1651
+ "loss": 0.0518,
1652
+ "step": 6625
1653
+ },
1654
+ {
1655
+ "epoch": 3.03,
1656
+ "learning_rate": 8.444303797468355e-06,
1657
+ "loss": 0.0479,
1658
+ "step": 6650
1659
+ },
1660
+ {
1661
+ "epoch": 3.04,
1662
+ "learning_rate": 8.437974683544304e-06,
1663
+ "loss": 0.0521,
1664
+ "step": 6675
1665
+ },
1666
+ {
1667
+ "epoch": 3.04,
1668
+ "learning_rate": 8.431645569620254e-06,
1669
+ "loss": 0.0436,
1670
+ "step": 6700
1671
+ },
1672
+ {
1673
+ "epoch": 3.04,
1674
+ "learning_rate": 8.425316455696203e-06,
1675
+ "loss": 0.0539,
1676
+ "step": 6725
1677
+ },
1678
+ {
1679
+ "epoch": 3.04,
1680
+ "learning_rate": 8.418987341772153e-06,
1681
+ "loss": 0.0446,
1682
+ "step": 6750
1683
+ },
1684
+ {
1685
+ "epoch": 3.04,
1686
+ "learning_rate": 8.412658227848102e-06,
1687
+ "loss": 0.0521,
1688
+ "step": 6775
1689
+ },
1690
+ {
1691
+ "epoch": 3.04,
1692
+ "learning_rate": 8.406329113924051e-06,
1693
+ "loss": 0.0391,
1694
+ "step": 6800
1695
+ },
1696
+ {
1697
+ "epoch": 3.04,
1698
+ "learning_rate": 8.400000000000001e-06,
1699
+ "loss": 0.0513,
1700
+ "step": 6825
1701
+ },
1702
+ {
1703
+ "epoch": 3.04,
1704
+ "learning_rate": 8.39367088607595e-06,
1705
+ "loss": 0.0449,
1706
+ "step": 6850
1707
+ },
1708
+ {
1709
+ "epoch": 3.04,
1710
+ "learning_rate": 8.3873417721519e-06,
1711
+ "loss": 0.0463,
1712
+ "step": 6875
1713
+ },
1714
+ {
1715
+ "epoch": 3.04,
1716
+ "learning_rate": 8.38101265822785e-06,
1717
+ "loss": 0.0495,
1718
+ "step": 6900
1719
+ },
1720
+ {
1721
+ "epoch": 3.04,
1722
+ "learning_rate": 8.374683544303797e-06,
1723
+ "loss": 0.0628,
1724
+ "step": 6925
1725
+ },
1726
+ {
1727
+ "epoch": 3.04,
1728
+ "learning_rate": 8.368354430379746e-06,
1729
+ "loss": 0.0594,
1730
+ "step": 6950
1731
+ },
1732
+ {
1733
+ "epoch": 3.04,
1734
+ "learning_rate": 8.362025316455696e-06,
1735
+ "loss": 0.0573,
1736
+ "step": 6975
1737
+ },
1738
+ {
1739
+ "epoch": 3.04,
1740
+ "learning_rate": 8.355696202531645e-06,
1741
+ "loss": 0.057,
1742
+ "step": 7000
1743
+ },
1744
+ {
1745
+ "epoch": 3.04,
1746
+ "eval_loss": 0.25129273533821106,
1747
+ "eval_runtime": 1479.2043,
1748
+ "eval_samples_per_second": 7.038,
1749
+ "eval_steps_per_second": 0.44,
1750
+ "eval_wer": 57.010683100538216,
1751
+ "step": 7000
1752
+ },
1753
+ {
1754
+ "epoch": 4.0,
1755
+ "learning_rate": 8.349367088607596e-06,
1756
+ "loss": 0.0534,
1757
+ "step": 7025
1758
+ },
1759
+ {
1760
+ "epoch": 4.0,
1761
+ "learning_rate": 8.343037974683546e-06,
1762
+ "loss": 0.0663,
1763
+ "step": 7050
1764
+ },
1765
+ {
1766
+ "epoch": 4.0,
1767
+ "learning_rate": 8.336708860759495e-06,
1768
+ "loss": 0.0698,
1769
+ "step": 7075
1770
+ },
1771
+ {
1772
+ "epoch": 4.0,
1773
+ "learning_rate": 8.330379746835445e-06,
1774
+ "loss": 0.0633,
1775
+ "step": 7100
1776
+ },
1777
+ {
1778
+ "epoch": 4.0,
1779
+ "learning_rate": 8.324050632911394e-06,
1780
+ "loss": 0.0808,
1781
+ "step": 7125
1782
+ },
1783
+ {
1784
+ "epoch": 4.0,
1785
+ "learning_rate": 8.317721518987343e-06,
1786
+ "loss": 0.0699,
1787
+ "step": 7150
1788
+ },
1789
+ {
1790
+ "epoch": 4.0,
1791
+ "learning_rate": 8.311392405063293e-06,
1792
+ "loss": 0.0736,
1793
+ "step": 7175
1794
+ },
1795
+ {
1796
+ "epoch": 4.0,
1797
+ "learning_rate": 8.305063291139242e-06,
1798
+ "loss": 0.069,
1799
+ "step": 7200
1800
+ },
1801
+ {
1802
+ "epoch": 4.01,
1803
+ "learning_rate": 8.29873417721519e-06,
1804
+ "loss": 0.0718,
1805
+ "step": 7225
1806
+ },
1807
+ {
1808
+ "epoch": 4.01,
1809
+ "learning_rate": 8.29240506329114e-06,
1810
+ "loss": 0.0674,
1811
+ "step": 7250
1812
+ },
1813
+ {
1814
+ "epoch": 4.01,
1815
+ "learning_rate": 8.286075949367089e-06,
1816
+ "loss": 0.0568,
1817
+ "step": 7275
1818
+ },
1819
+ {
1820
+ "epoch": 4.01,
1821
+ "learning_rate": 8.279746835443038e-06,
1822
+ "loss": 0.0497,
1823
+ "step": 7300
1824
+ },
1825
+ {
1826
+ "epoch": 4.01,
1827
+ "learning_rate": 8.273417721518988e-06,
1828
+ "loss": 0.0563,
1829
+ "step": 7325
1830
+ },
1831
+ {
1832
+ "epoch": 4.01,
1833
+ "learning_rate": 8.267088607594937e-06,
1834
+ "loss": 0.0649,
1835
+ "step": 7350
1836
+ },
1837
+ {
1838
+ "epoch": 4.01,
1839
+ "learning_rate": 8.260759493670887e-06,
1840
+ "loss": 0.042,
1841
+ "step": 7375
1842
+ },
1843
+ {
1844
+ "epoch": 4.01,
1845
+ "learning_rate": 8.254430379746836e-06,
1846
+ "loss": 0.0559,
1847
+ "step": 7400
1848
+ },
1849
+ {
1850
+ "epoch": 4.01,
1851
+ "learning_rate": 8.248101265822785e-06,
1852
+ "loss": 0.0476,
1853
+ "step": 7425
1854
+ },
1855
+ {
1856
+ "epoch": 4.01,
1857
+ "learning_rate": 8.241772151898735e-06,
1858
+ "loss": 0.0427,
1859
+ "step": 7450
1860
+ },
1861
+ {
1862
+ "epoch": 4.01,
1863
+ "learning_rate": 8.235443037974684e-06,
1864
+ "loss": 0.0533,
1865
+ "step": 7475
1866
+ },
1867
+ {
1868
+ "epoch": 4.01,
1869
+ "learning_rate": 8.229113924050634e-06,
1870
+ "loss": 0.0633,
1871
+ "step": 7500
1872
+ },
1873
+ {
1874
+ "epoch": 4.01,
1875
+ "learning_rate": 8.222784810126583e-06,
1876
+ "loss": 0.055,
1877
+ "step": 7525
1878
+ },
1879
+ {
1880
+ "epoch": 4.01,
1881
+ "learning_rate": 8.216455696202533e-06,
1882
+ "loss": 0.043,
1883
+ "step": 7550
1884
+ },
1885
+ {
1886
+ "epoch": 4.01,
1887
+ "learning_rate": 8.210126582278482e-06,
1888
+ "loss": 0.0421,
1889
+ "step": 7575
1890
+ },
1891
+ {
1892
+ "epoch": 4.01,
1893
+ "learning_rate": 8.203797468354431e-06,
1894
+ "loss": 0.0519,
1895
+ "step": 7600
1896
+ },
1897
+ {
1898
+ "epoch": 4.02,
1899
+ "learning_rate": 8.197468354430381e-06,
1900
+ "loss": 0.0398,
1901
+ "step": 7625
1902
+ },
1903
+ {
1904
+ "epoch": 4.02,
1905
+ "learning_rate": 8.19113924050633e-06,
1906
+ "loss": 0.041,
1907
+ "step": 7650
1908
+ },
1909
+ {
1910
+ "epoch": 4.02,
1911
+ "learning_rate": 8.184810126582278e-06,
1912
+ "loss": 0.0477,
1913
+ "step": 7675
1914
+ },
1915
+ {
1916
+ "epoch": 4.02,
1917
+ "learning_rate": 8.178481012658227e-06,
1918
+ "loss": 0.046,
1919
+ "step": 7700
1920
+ },
1921
+ {
1922
+ "epoch": 4.02,
1923
+ "learning_rate": 8.172151898734177e-06,
1924
+ "loss": 0.0465,
1925
+ "step": 7725
1926
+ },
1927
+ {
1928
+ "epoch": 4.02,
1929
+ "learning_rate": 8.165822784810126e-06,
1930
+ "loss": 0.0359,
1931
+ "step": 7750
1932
+ },
1933
+ {
1934
+ "epoch": 4.02,
1935
+ "learning_rate": 8.159493670886076e-06,
1936
+ "loss": 0.0362,
1937
+ "step": 7775
1938
+ },
1939
+ {
1940
+ "epoch": 4.02,
1941
+ "learning_rate": 8.153164556962025e-06,
1942
+ "loss": 0.04,
1943
+ "step": 7800
1944
+ },
1945
+ {
1946
+ "epoch": 4.02,
1947
+ "learning_rate": 8.146835443037975e-06,
1948
+ "loss": 0.0433,
1949
+ "step": 7825
1950
+ },
1951
+ {
1952
+ "epoch": 4.02,
1953
+ "learning_rate": 8.140506329113926e-06,
1954
+ "loss": 0.0399,
1955
+ "step": 7850
1956
+ },
1957
+ {
1958
+ "epoch": 4.02,
1959
+ "learning_rate": 8.134177215189875e-06,
1960
+ "loss": 0.0396,
1961
+ "step": 7875
1962
+ },
1963
+ {
1964
+ "epoch": 4.02,
1965
+ "learning_rate": 8.127848101265825e-06,
1966
+ "loss": 0.0472,
1967
+ "step": 7900
1968
+ },
1969
+ {
1970
+ "epoch": 4.02,
1971
+ "learning_rate": 8.121518987341774e-06,
1972
+ "loss": 0.0533,
1973
+ "step": 7925
1974
+ },
1975
+ {
1976
+ "epoch": 4.02,
1977
+ "learning_rate": 8.115189873417722e-06,
1978
+ "loss": 0.0564,
1979
+ "step": 7950
1980
+ },
1981
+ {
1982
+ "epoch": 4.02,
1983
+ "learning_rate": 8.108860759493671e-06,
1984
+ "loss": 0.0508,
1985
+ "step": 7975
1986
+ },
1987
+ {
1988
+ "epoch": 4.02,
1989
+ "learning_rate": 8.10253164556962e-06,
1990
+ "loss": 0.0375,
1991
+ "step": 8000
1992
+ },
1993
+ {
1994
+ "epoch": 4.02,
1995
+ "eval_loss": 0.252152681350708,
1996
+ "eval_runtime": 1456.2607,
1997
+ "eval_samples_per_second": 7.149,
1998
+ "eval_steps_per_second": 0.447,
1999
+ "eval_wer": 44.38120945056017,
2000
+ "step": 8000
2001
+ },
2002
+ {
2003
+ "epoch": 4.03,
2004
+ "learning_rate": 8.09620253164557e-06,
2005
+ "loss": 0.0296,
2006
+ "step": 8025
2007
+ },
2008
+ {
2009
+ "epoch": 4.03,
2010
+ "learning_rate": 8.08987341772152e-06,
2011
+ "loss": 0.0395,
2012
+ "step": 8050
2013
+ },
2014
+ {
2015
+ "epoch": 4.03,
2016
+ "learning_rate": 8.083544303797469e-06,
2017
+ "loss": 0.0409,
2018
+ "step": 8075
2019
+ },
2020
+ {
2021
+ "epoch": 4.03,
2022
+ "learning_rate": 8.077215189873418e-06,
2023
+ "loss": 0.0343,
2024
+ "step": 8100
2025
+ },
2026
+ {
2027
+ "epoch": 4.03,
2028
+ "learning_rate": 8.070886075949368e-06,
2029
+ "loss": 0.0321,
2030
+ "step": 8125
2031
+ },
2032
+ {
2033
+ "epoch": 4.03,
2034
+ "learning_rate": 8.064556962025317e-06,
2035
+ "loss": 0.0333,
2036
+ "step": 8150
2037
+ },
2038
+ {
2039
+ "epoch": 4.03,
2040
+ "learning_rate": 8.058227848101267e-06,
2041
+ "loss": 0.0327,
2042
+ "step": 8175
2043
+ },
2044
+ {
2045
+ "epoch": 4.03,
2046
+ "learning_rate": 8.051898734177216e-06,
2047
+ "loss": 0.0318,
2048
+ "step": 8200
2049
+ },
2050
+ {
2051
+ "epoch": 4.03,
2052
+ "learning_rate": 8.045569620253165e-06,
2053
+ "loss": 0.0405,
2054
+ "step": 8225
2055
+ },
2056
+ {
2057
+ "epoch": 4.03,
2058
+ "learning_rate": 8.039240506329115e-06,
2059
+ "loss": 0.0382,
2060
+ "step": 8250
2061
+ },
2062
+ {
2063
+ "epoch": 4.03,
2064
+ "learning_rate": 8.032911392405064e-06,
2065
+ "loss": 0.0317,
2066
+ "step": 8275
2067
+ },
2068
+ {
2069
+ "epoch": 4.03,
2070
+ "learning_rate": 8.026582278481014e-06,
2071
+ "loss": 0.0285,
2072
+ "step": 8300
2073
+ },
2074
+ {
2075
+ "epoch": 4.03,
2076
+ "learning_rate": 8.020253164556963e-06,
2077
+ "loss": 0.0285,
2078
+ "step": 8325
2079
+ },
2080
+ {
2081
+ "epoch": 4.03,
2082
+ "learning_rate": 8.013924050632913e-06,
2083
+ "loss": 0.0269,
2084
+ "step": 8350
2085
+ },
2086
+ {
2087
+ "epoch": 4.03,
2088
+ "learning_rate": 8.007594936708862e-06,
2089
+ "loss": 0.0334,
2090
+ "step": 8375
2091
+ },
2092
+ {
2093
+ "epoch": 4.03,
2094
+ "learning_rate": 8.00126582278481e-06,
2095
+ "loss": 0.0406,
2096
+ "step": 8400
2097
+ },
2098
+ {
2099
+ "epoch": 4.04,
2100
+ "learning_rate": 7.99493670886076e-06,
2101
+ "loss": 0.038,
2102
+ "step": 8425
2103
+ },
2104
+ {
2105
+ "epoch": 4.04,
2106
+ "learning_rate": 7.988607594936709e-06,
2107
+ "loss": 0.0315,
2108
+ "step": 8450
2109
+ },
2110
+ {
2111
+ "epoch": 4.04,
2112
+ "learning_rate": 7.982278481012658e-06,
2113
+ "loss": 0.0329,
2114
+ "step": 8475
2115
+ },
2116
+ {
2117
+ "epoch": 4.04,
2118
+ "learning_rate": 7.975949367088607e-06,
2119
+ "loss": 0.0373,
2120
+ "step": 8500
2121
+ },
2122
+ {
2123
+ "epoch": 4.04,
2124
+ "learning_rate": 7.969620253164557e-06,
2125
+ "loss": 0.0514,
2126
+ "step": 8525
2127
+ },
2128
+ {
2129
+ "epoch": 4.04,
2130
+ "learning_rate": 7.963291139240506e-06,
2131
+ "loss": 0.0349,
2132
+ "step": 8550
2133
+ },
2134
+ {
2135
+ "epoch": 4.04,
2136
+ "learning_rate": 7.956962025316456e-06,
2137
+ "loss": 0.0385,
2138
+ "step": 8575
2139
+ },
2140
+ {
2141
+ "epoch": 4.04,
2142
+ "learning_rate": 7.950632911392405e-06,
2143
+ "loss": 0.0367,
2144
+ "step": 8600
2145
+ },
2146
+ {
2147
+ "epoch": 4.04,
2148
+ "learning_rate": 7.944303797468355e-06,
2149
+ "loss": 0.0317,
2150
+ "step": 8625
2151
+ },
2152
+ {
2153
+ "epoch": 4.04,
2154
+ "learning_rate": 7.937974683544304e-06,
2155
+ "loss": 0.0368,
2156
+ "step": 8650
2157
+ },
2158
+ {
2159
+ "epoch": 4.04,
2160
+ "learning_rate": 7.931645569620255e-06,
2161
+ "loss": 0.045,
2162
+ "step": 8675
2163
+ },
2164
+ {
2165
+ "epoch": 4.04,
2166
+ "learning_rate": 7.925316455696203e-06,
2167
+ "loss": 0.0456,
2168
+ "step": 8700
2169
+ },
2170
+ {
2171
+ "epoch": 4.04,
2172
+ "learning_rate": 7.918987341772152e-06,
2173
+ "loss": 0.0421,
2174
+ "step": 8725
2175
+ },
2176
+ {
2177
+ "epoch": 4.04,
2178
+ "learning_rate": 7.912658227848102e-06,
2179
+ "loss": 0.043,
2180
+ "step": 8750
2181
+ },
2182
+ {
2183
+ "epoch": 5.0,
2184
+ "learning_rate": 7.906329113924051e-06,
2185
+ "loss": 0.0428,
2186
+ "step": 8775
2187
+ },
2188
+ {
2189
+ "epoch": 5.0,
2190
+ "learning_rate": 7.9e-06,
2191
+ "loss": 0.0459,
2192
+ "step": 8800
2193
+ },
2194
+ {
2195
+ "epoch": 5.0,
2196
+ "learning_rate": 7.89367088607595e-06,
2197
+ "loss": 0.0467,
2198
+ "step": 8825
2199
+ },
2200
+ {
2201
+ "epoch": 5.0,
2202
+ "learning_rate": 7.8873417721519e-06,
2203
+ "loss": 0.0462,
2204
+ "step": 8850
2205
+ },
2206
+ {
2207
+ "epoch": 5.0,
2208
+ "learning_rate": 7.881012658227849e-06,
2209
+ "loss": 0.0578,
2210
+ "step": 8875
2211
+ },
2212
+ {
2213
+ "epoch": 5.0,
2214
+ "learning_rate": 7.874683544303798e-06,
2215
+ "loss": 0.0422,
2216
+ "step": 8900
2217
+ },
2218
+ {
2219
+ "epoch": 5.0,
2220
+ "learning_rate": 7.868354430379748e-06,
2221
+ "loss": 0.0546,
2222
+ "step": 8925
2223
+ },
2224
+ {
2225
+ "epoch": 5.0,
2226
+ "learning_rate": 7.862025316455697e-06,
2227
+ "loss": 0.0515,
2228
+ "step": 8950
2229
+ },
2230
+ {
2231
+ "epoch": 5.01,
2232
+ "learning_rate": 7.855696202531647e-06,
2233
+ "loss": 0.0541,
2234
+ "step": 8975
2235
+ },
2236
+ {
2237
+ "epoch": 5.01,
2238
+ "learning_rate": 7.849367088607596e-06,
2239
+ "loss": 0.0479,
2240
+ "step": 9000
2241
+ },
2242
+ {
2243
+ "epoch": 5.01,
2244
+ "eval_loss": 0.2375001162290573,
2245
+ "eval_runtime": 1456.1088,
2246
+ "eval_samples_per_second": 7.15,
2247
+ "eval_steps_per_second": 0.447,
2248
+ "eval_wer": 31.58262735979447,
2249
+ "step": 9000
2250
+ },
2251
+ {
2252
+ "epoch": 5.01,
2253
+ "learning_rate": 7.843037974683545e-06,
2254
+ "loss": 0.0434,
2255
+ "step": 9025
2256
+ },
2257
+ {
2258
+ "epoch": 5.01,
2259
+ "learning_rate": 7.836708860759495e-06,
2260
+ "loss": 0.033,
2261
+ "step": 9050
2262
+ },
2263
+ {
2264
+ "epoch": 5.01,
2265
+ "learning_rate": 7.830379746835444e-06,
2266
+ "loss": 0.0494,
2267
+ "step": 9075
2268
+ },
2269
+ {
2270
+ "epoch": 5.01,
2271
+ "learning_rate": 7.824050632911394e-06,
2272
+ "loss": 0.0522,
2273
+ "step": 9100
2274
+ },
2275
+ {
2276
+ "epoch": 5.01,
2277
+ "learning_rate": 7.817721518987343e-06,
2278
+ "loss": 0.0338,
2279
+ "step": 9125
2280
+ },
2281
+ {
2282
+ "epoch": 5.01,
2283
+ "learning_rate": 7.811392405063291e-06,
2284
+ "loss": 0.0395,
2285
+ "step": 9150
2286
+ },
2287
+ {
2288
+ "epoch": 5.01,
2289
+ "learning_rate": 7.80506329113924e-06,
2290
+ "loss": 0.0332,
2291
+ "step": 9175
2292
+ },
2293
+ {
2294
+ "epoch": 5.01,
2295
+ "learning_rate": 7.79873417721519e-06,
2296
+ "loss": 0.0381,
2297
+ "step": 9200
2298
+ },
2299
+ {
2300
+ "epoch": 5.01,
2301
+ "learning_rate": 7.79240506329114e-06,
2302
+ "loss": 0.0409,
2303
+ "step": 9225
2304
+ },
2305
+ {
2306
+ "epoch": 5.01,
2307
+ "learning_rate": 7.786075949367089e-06,
2308
+ "loss": 0.0529,
2309
+ "step": 9250
2310
+ },
2311
+ {
2312
+ "epoch": 5.01,
2313
+ "learning_rate": 7.779746835443038e-06,
2314
+ "loss": 0.0498,
2315
+ "step": 9275
2316
+ },
2317
+ {
2318
+ "epoch": 5.01,
2319
+ "learning_rate": 7.773417721518987e-06,
2320
+ "loss": 0.0379,
2321
+ "step": 9300
2322
+ },
2323
+ {
2324
+ "epoch": 5.01,
2325
+ "learning_rate": 7.767088607594937e-06,
2326
+ "loss": 0.0316,
2327
+ "step": 9325
2328
+ },
2329
+ {
2330
+ "epoch": 5.01,
2331
+ "learning_rate": 7.760759493670886e-06,
2332
+ "loss": 0.0398,
2333
+ "step": 9350
2334
+ },
2335
+ {
2336
+ "epoch": 5.02,
2337
+ "learning_rate": 7.754430379746836e-06,
2338
+ "loss": 0.0427,
2339
+ "step": 9375
2340
+ },
2341
+ {
2342
+ "epoch": 5.02,
2343
+ "learning_rate": 7.748101265822785e-06,
2344
+ "loss": 0.0363,
2345
+ "step": 9400
2346
+ },
2347
+ {
2348
+ "epoch": 5.02,
2349
+ "learning_rate": 7.741772151898735e-06,
2350
+ "loss": 0.0393,
2351
+ "step": 9425
2352
+ },
2353
+ {
2354
+ "epoch": 5.02,
2355
+ "learning_rate": 7.735443037974684e-06,
2356
+ "loss": 0.0377,
2357
+ "step": 9450
2358
+ },
2359
+ {
2360
+ "epoch": 5.02,
2361
+ "learning_rate": 7.729113924050633e-06,
2362
+ "loss": 0.0313,
2363
+ "step": 9475
2364
+ },
2365
+ {
2366
+ "epoch": 5.02,
2367
+ "learning_rate": 7.722784810126583e-06,
2368
+ "loss": 0.0359,
2369
+ "step": 9500
2370
+ },
2371
+ {
2372
+ "epoch": 5.02,
2373
+ "learning_rate": 7.716455696202532e-06,
2374
+ "loss": 0.0289,
2375
+ "step": 9525
2376
+ },
2377
+ {
2378
+ "epoch": 5.02,
2379
+ "learning_rate": 7.710126582278482e-06,
2380
+ "loss": 0.0354,
2381
+ "step": 9550
2382
+ },
2383
+ {
2384
+ "epoch": 5.02,
2385
+ "learning_rate": 7.703797468354431e-06,
2386
+ "loss": 0.0309,
2387
+ "step": 9575
2388
+ },
2389
+ {
2390
+ "epoch": 5.02,
2391
+ "learning_rate": 7.69746835443038e-06,
2392
+ "loss": 0.0321,
2393
+ "step": 9600
2394
+ },
2395
+ {
2396
+ "epoch": 5.02,
2397
+ "learning_rate": 7.69113924050633e-06,
2398
+ "loss": 0.0392,
2399
+ "step": 9625
2400
+ },
2401
+ {
2402
+ "epoch": 5.02,
2403
+ "learning_rate": 7.68481012658228e-06,
2404
+ "loss": 0.0362,
2405
+ "step": 9650
2406
+ },
2407
+ {
2408
+ "epoch": 5.02,
2409
+ "learning_rate": 7.678481012658229e-06,
2410
+ "loss": 0.0358,
2411
+ "step": 9675
2412
+ },
2413
+ {
2414
+ "epoch": 5.02,
2415
+ "learning_rate": 7.672151898734178e-06,
2416
+ "loss": 0.045,
2417
+ "step": 9700
2418
+ },
2419
+ {
2420
+ "epoch": 5.02,
2421
+ "learning_rate": 7.665822784810128e-06,
2422
+ "loss": 0.0401,
2423
+ "step": 9725
2424
+ },
2425
+ {
2426
+ "epoch": 5.02,
2427
+ "learning_rate": 7.659493670886077e-06,
2428
+ "loss": 0.0357,
2429
+ "step": 9750
2430
+ },
2431
+ {
2432
+ "epoch": 5.03,
2433
+ "learning_rate": 7.653164556962027e-06,
2434
+ "loss": 0.0256,
2435
+ "step": 9775
2436
+ },
2437
+ {
2438
+ "epoch": 5.03,
2439
+ "learning_rate": 7.646835443037976e-06,
2440
+ "loss": 0.0358,
2441
+ "step": 9800
2442
+ },
2443
+ {
2444
+ "epoch": 5.03,
2445
+ "learning_rate": 7.640506329113925e-06,
2446
+ "loss": 0.0361,
2447
+ "step": 9825
2448
+ },
2449
+ {
2450
+ "epoch": 5.03,
2451
+ "learning_rate": 7.634177215189875e-06,
2452
+ "loss": 0.0365,
2453
+ "step": 9850
2454
+ },
2455
+ {
2456
+ "epoch": 5.03,
2457
+ "learning_rate": 7.6278481012658234e-06,
2458
+ "loss": 0.0344,
2459
+ "step": 9875
2460
+ },
2461
+ {
2462
+ "epoch": 5.03,
2463
+ "learning_rate": 7.621518987341773e-06,
2464
+ "loss": 0.0349,
2465
+ "step": 9900
2466
+ },
2467
+ {
2468
+ "epoch": 5.03,
2469
+ "learning_rate": 7.615189873417722e-06,
2470
+ "loss": 0.0289,
2471
+ "step": 9925
2472
+ },
2473
+ {
2474
+ "epoch": 5.03,
2475
+ "learning_rate": 7.608860759493672e-06,
2476
+ "loss": 0.0259,
2477
+ "step": 9950
2478
+ },
2479
+ {
2480
+ "epoch": 5.03,
2481
+ "learning_rate": 7.602531645569621e-06,
2482
+ "loss": 0.0289,
2483
+ "step": 9975
2484
+ },
2485
+ {
2486
+ "epoch": 5.03,
2487
+ "learning_rate": 7.59620253164557e-06,
2488
+ "loss": 0.0269,
2489
+ "step": 10000
2490
+ },
2491
+ {
2492
+ "epoch": 5.03,
2493
+ "eval_loss": 0.2615405321121216,
2494
+ "eval_runtime": 1420.7197,
2495
+ "eval_samples_per_second": 7.328,
2496
+ "eval_steps_per_second": 0.458,
2497
+ "eval_wer": 36.23310948145498,
2498
+ "step": 10000
2499
+ }
2500
+ ],
2501
+ "max_steps": 40000,
2502
+ "num_train_epochs": 9223372036854775807,
2503
+ "total_flos": 4.6159234781184e+19,
2504
+ "trial_name": null,
2505
+ "trial_params": null
2506
+ }
checkpoint-10000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0289e1bf8fd1a33f7c3c9d3e15cc9bdd13d10e9774e0ada934c0b9d71501a993
3
+ size 4091
checkpoint-11000/config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-small",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": true,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 768,
17
+ "decoder_attention_heads": 12,
18
+ "decoder_ffn_dim": 3072,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 12,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 12,
24
+ "encoder_ffn_dim": 3072,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 12,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": null,
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "mask_feature_length": 10,
32
+ "mask_feature_min_masks": 0,
33
+ "mask_feature_prob": 0.05,
34
+ "mask_time_length": 10,
35
+ "mask_time_min_masks": 2,
36
+ "mask_time_prob": 0.05,
37
+ "max_length": 448,
38
+ "max_source_positions": 1500,
39
+ "max_target_positions": 448,
40
+ "model_type": "whisper",
41
+ "num_hidden_layers": 12,
42
+ "num_mel_bins": 80,
43
+ "pad_token_id": 50257,
44
+ "scale_embedding": false,
45
+ "suppress_tokens": [],
46
+ "torch_dtype": "float32",
47
+ "transformers_version": "4.30.0.dev0",
48
+ "use_cache": true,
49
+ "use_weighted_layer_sum": false,
50
+ "vocab_size": 51865
51
+ }
checkpoint-11000/generation_config.json ADDED
@@ -0,0 +1,221 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "begin_suppress_tokens": [
3
+ 220,
4
+ 50257
5
+ ],
6
+ "bos_token_id": 50257,
7
+ "decoder_start_token_id": 50258,
8
+ "eos_token_id": 50257,
9
+ "forced_decoder_ids": [
10
+ [
11
+ 1,
12
+ null
13
+ ],
14
+ [
15
+ 2,
16
+ 50359
17
+ ]
18
+ ],
19
+ "is_multilingual": true,
20
+ "lang_to_id": {
21
+ "<|af|>": 50327,
22
+ "<|am|>": 50334,
23
+ "<|ar|>": 50272,
24
+ "<|as|>": 50350,
25
+ "<|az|>": 50304,
26
+ "<|ba|>": 50355,
27
+ "<|be|>": 50330,
28
+ "<|bg|>": 50292,
29
+ "<|bn|>": 50302,
30
+ "<|bo|>": 50347,
31
+ "<|br|>": 50309,
32
+ "<|bs|>": 50315,
33
+ "<|ca|>": 50270,
34
+ "<|cs|>": 50283,
35
+ "<|cy|>": 50297,
36
+ "<|da|>": 50285,
37
+ "<|de|>": 50261,
38
+ "<|el|>": 50281,
39
+ "<|en|>": 50259,
40
+ "<|es|>": 50262,
41
+ "<|et|>": 50307,
42
+ "<|eu|>": 50310,
43
+ "<|fa|>": 50300,
44
+ "<|fi|>": 50277,
45
+ "<|fo|>": 50338,
46
+ "<|fr|>": 50265,
47
+ "<|gl|>": 50319,
48
+ "<|gu|>": 50333,
49
+ "<|haw|>": 50352,
50
+ "<|ha|>": 50354,
51
+ "<|he|>": 50279,
52
+ "<|hi|>": 50276,
53
+ "<|hr|>": 50291,
54
+ "<|ht|>": 50339,
55
+ "<|hu|>": 50286,
56
+ "<|hy|>": 50312,
57
+ "<|id|>": 50275,
58
+ "<|is|>": 50311,
59
+ "<|it|>": 50274,
60
+ "<|ja|>": 50266,
61
+ "<|jw|>": 50356,
62
+ "<|ka|>": 50329,
63
+ "<|kk|>": 50316,
64
+ "<|km|>": 50323,
65
+ "<|kn|>": 50306,
66
+ "<|ko|>": 50264,
67
+ "<|la|>": 50294,
68
+ "<|lb|>": 50345,
69
+ "<|ln|>": 50353,
70
+ "<|lo|>": 50336,
71
+ "<|lt|>": 50293,
72
+ "<|lv|>": 50301,
73
+ "<|mg|>": 50349,
74
+ "<|mi|>": 50295,
75
+ "<|mk|>": 50308,
76
+ "<|ml|>": 50296,
77
+ "<|mn|>": 50314,
78
+ "<|mr|>": 50320,
79
+ "<|ms|>": 50282,
80
+ "<|mt|>": 50343,
81
+ "<|my|>": 50346,
82
+ "<|ne|>": 50313,
83
+ "<|nl|>": 50271,
84
+ "<|nn|>": 50342,
85
+ "<|no|>": 50288,
86
+ "<|oc|>": 50328,
87
+ "<|pa|>": 50321,
88
+ "<|pl|>": 50269,
89
+ "<|ps|>": 50340,
90
+ "<|pt|>": 50267,
91
+ "<|ro|>": 50284,
92
+ "<|ru|>": 50263,
93
+ "<|sa|>": 50344,
94
+ "<|sd|>": 50332,
95
+ "<|si|>": 50322,
96
+ "<|sk|>": 50298,
97
+ "<|sl|>": 50305,
98
+ "<|sn|>": 50324,
99
+ "<|so|>": 50326,
100
+ "<|sq|>": 50317,
101
+ "<|sr|>": 50303,
102
+ "<|su|>": 50357,
103
+ "<|sv|>": 50273,
104
+ "<|sw|>": 50318,
105
+ "<|ta|>": 50287,
106
+ "<|te|>": 50299,
107
+ "<|tg|>": 50331,
108
+ "<|th|>": 50289,
109
+ "<|tk|>": 50341,
110
+ "<|tl|>": 50348,
111
+ "<|tr|>": 50268,
112
+ "<|tt|>": 50351,
113
+ "<|uk|>": 50280,
114
+ "<|ur|>": 50290,
115
+ "<|uz|>": 50337,
116
+ "<|vi|>": 50278,
117
+ "<|yi|>": 50335,
118
+ "<|yo|>": 50325,
119
+ "<|zh|>": 50260
120
+ },
121
+ "max_initial_timestamp_index": 1,
122
+ "max_length": 448,
123
+ "no_timestamps_token_id": 50363,
124
+ "pad_token_id": 50257,
125
+ "return_timestamps": false,
126
+ "suppress_tokens": [
127
+ 1,
128
+ 2,
129
+ 7,
130
+ 8,
131
+ 9,
132
+ 10,
133
+ 14,
134
+ 25,
135
+ 26,
136
+ 27,
137
+ 28,
138
+ 29,
139
+ 31,
140
+ 58,
141
+ 59,
142
+ 60,
143
+ 61,
144
+ 62,
145
+ 63,
146
+ 90,
147
+ 91,
148
+ 92,
149
+ 93,
150
+ 359,
151
+ 503,
152
+ 522,
153
+ 542,
154
+ 873,
155
+ 893,
156
+ 902,
157
+ 918,
158
+ 922,
159
+ 931,
160
+ 1350,
161
+ 1853,
162
+ 1982,
163
+ 2460,
164
+ 2627,
165
+ 3246,
166
+ 3253,
167
+ 3268,
168
+ 3536,
169
+ 3846,
170
+ 3961,
171
+ 4183,
172
+ 4667,
173
+ 6585,
174
+ 6647,
175
+ 7273,
176
+ 9061,
177
+ 9383,
178
+ 10428,
179
+ 10929,
180
+ 11938,
181
+ 12033,
182
+ 12331,
183
+ 12562,
184
+ 13793,
185
+ 14157,
186
+ 14635,
187
+ 15265,
188
+ 15618,
189
+ 16553,
190
+ 16604,
191
+ 18362,
192
+ 18956,
193
+ 20075,
194
+ 21675,
195
+ 22520,
196
+ 26130,
197
+ 26161,
198
+ 26435,
199
+ 28279,
200
+ 29464,
201
+ 31650,
202
+ 32302,
203
+ 32470,
204
+ 36865,
205
+ 42863,
206
+ 47425,
207
+ 49870,
208
+ 50254,
209
+ 50258,
210
+ 50358,
211
+ 50359,
212
+ 50360,
213
+ 50361,
214
+ 50362
215
+ ],
216
+ "task_to_id": {
217
+ "transcribe": 50359,
218
+ "translate": 50358
219
+ },
220
+ "transformers_version": "4.30.0.dev0"
221
+ }
checkpoint-11000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ebe9a7167ac6d8714c10503f224fa5b58a019c86e9388beb5e866d95654371d
3
+ size 1934161093
checkpoint-11000/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-11000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8185a13e4eb3a24a6da5cf406243ba0a61f2f3047f67b491635f12b941b8fbe8
3
+ size 967102601
checkpoint-11000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2fb11b8fe7afefbde3385e8e8f0bb0a92eca3f4813d20d1e16253d70edd930a0
3
+ size 14639
checkpoint-11000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:652e9e9b1f49df5c7d4e2eda2b4c0d4bbda8a9746e3a136d3022a7b8fb113822
3
+ size 557
checkpoint-11000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:724fd087a1e24275c921af58d6e25f4867cedc7578db7212e10c786cd348c477
3
+ size 627
checkpoint-11000/trainer_state.json ADDED
@@ -0,0 +1,2755 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 31.58262735979447,
3
+ "best_model_checkpoint": "./whisper-small-fa-aug/checkpoint-9000",
4
+ "epoch": 6.01235,
5
+ "global_step": 11000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 4.2000000000000006e-07,
13
+ "loss": 3.8,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 9.200000000000001e-07,
19
+ "loss": 3.0497,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.0,
24
+ "learning_rate": 1.42e-06,
25
+ "loss": 2.0102,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.0,
30
+ "learning_rate": 1.9200000000000003e-06,
31
+ "loss": 1.5948,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.0,
36
+ "learning_rate": 2.42e-06,
37
+ "loss": 1.4094,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.0,
42
+ "learning_rate": 2.92e-06,
43
+ "loss": 1.2611,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.0,
48
+ "learning_rate": 3.4200000000000007e-06,
49
+ "loss": 1.1366,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.01,
54
+ "learning_rate": 3.920000000000001e-06,
55
+ "loss": 1.0817,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.01,
60
+ "learning_rate": 4.42e-06,
61
+ "loss": 1.0324,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.01,
66
+ "learning_rate": 4.92e-06,
67
+ "loss": 0.9176,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.01,
72
+ "learning_rate": 5.420000000000001e-06,
73
+ "loss": 0.7975,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.01,
78
+ "learning_rate": 5.92e-06,
79
+ "loss": 0.7079,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.01,
84
+ "learning_rate": 6.42e-06,
85
+ "loss": 0.5832,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.01,
90
+ "learning_rate": 6.92e-06,
91
+ "loss": 0.4469,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.01,
96
+ "learning_rate": 7.420000000000001e-06,
97
+ "loss": 0.3913,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.01,
102
+ "learning_rate": 7.92e-06,
103
+ "loss": 0.4189,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.01,
108
+ "learning_rate": 8.42e-06,
109
+ "loss": 0.3756,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.01,
114
+ "learning_rate": 8.920000000000001e-06,
115
+ "loss": 0.3876,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.01,
120
+ "learning_rate": 9.42e-06,
121
+ "loss": 0.3802,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.01,
126
+ "learning_rate": 9.920000000000002e-06,
127
+ "loss": 0.4061,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.01,
132
+ "learning_rate": 9.994683544303798e-06,
133
+ "loss": 0.3885,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.01,
138
+ "learning_rate": 9.988354430379748e-06,
139
+ "loss": 0.3449,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.01,
144
+ "learning_rate": 9.982025316455697e-06,
145
+ "loss": 0.3292,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.01,
150
+ "learning_rate": 9.975696202531647e-06,
151
+ "loss": 0.3439,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.02,
156
+ "learning_rate": 9.969367088607596e-06,
157
+ "loss": 0.3031,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.02,
162
+ "learning_rate": 9.963037974683545e-06,
163
+ "loss": 0.3159,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.02,
168
+ "learning_rate": 9.956708860759495e-06,
169
+ "loss": 0.3304,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.02,
174
+ "learning_rate": 9.950379746835444e-06,
175
+ "loss": 0.2918,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.02,
180
+ "learning_rate": 9.944050632911392e-06,
181
+ "loss": 0.2785,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.02,
186
+ "learning_rate": 9.937721518987341e-06,
187
+ "loss": 0.2618,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.02,
192
+ "learning_rate": 9.931392405063291e-06,
193
+ "loss": 0.2862,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.02,
198
+ "learning_rate": 9.92506329113924e-06,
199
+ "loss": 0.2962,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.02,
204
+ "learning_rate": 9.91873417721519e-06,
205
+ "loss": 0.2935,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.02,
210
+ "learning_rate": 9.912405063291141e-06,
211
+ "loss": 0.2827,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.02,
216
+ "learning_rate": 9.90607594936709e-06,
217
+ "loss": 0.2702,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 0.02,
222
+ "learning_rate": 9.89974683544304e-06,
223
+ "loss": 0.2987,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 0.02,
228
+ "learning_rate": 9.89341772151899e-06,
229
+ "loss": 0.3067,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 0.02,
234
+ "learning_rate": 9.887088607594939e-06,
235
+ "loss": 0.2962,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 0.02,
240
+ "learning_rate": 9.880759493670888e-06,
241
+ "loss": 0.2807,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 0.03,
246
+ "learning_rate": 9.874430379746836e-06,
247
+ "loss": 0.2533,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 0.03,
252
+ "eval_loss": 0.3274388611316681,
253
+ "eval_runtime": 1481.9106,
254
+ "eval_samples_per_second": 7.025,
255
+ "eval_steps_per_second": 0.439,
256
+ "eval_wer": 40.657571667832,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 0.03,
261
+ "learning_rate": 9.868101265822785e-06,
262
+ "loss": 0.21,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 0.03,
267
+ "learning_rate": 9.861772151898735e-06,
268
+ "loss": 0.2463,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 0.03,
273
+ "learning_rate": 9.855443037974684e-06,
274
+ "loss": 0.2688,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 0.03,
279
+ "learning_rate": 9.849113924050633e-06,
280
+ "loss": 0.2293,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 0.03,
285
+ "learning_rate": 9.842784810126583e-06,
286
+ "loss": 0.2283,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 0.03,
291
+ "learning_rate": 9.836455696202532e-06,
292
+ "loss": 0.228,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 0.03,
297
+ "learning_rate": 9.830126582278482e-06,
298
+ "loss": 0.1919,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 0.03,
303
+ "learning_rate": 9.823797468354431e-06,
304
+ "loss": 0.2173,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 0.03,
309
+ "learning_rate": 9.81746835443038e-06,
310
+ "loss": 0.2066,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 0.03,
315
+ "learning_rate": 9.81113924050633e-06,
316
+ "loss": 0.1826,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 0.03,
321
+ "learning_rate": 9.80481012658228e-06,
322
+ "loss": 0.1855,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 0.03,
327
+ "learning_rate": 9.798481012658229e-06,
328
+ "loss": 0.1702,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 0.03,
333
+ "learning_rate": 9.792151898734178e-06,
334
+ "loss": 0.1714,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 0.03,
339
+ "learning_rate": 9.785822784810128e-06,
340
+ "loss": 0.1686,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 0.03,
345
+ "learning_rate": 9.779493670886077e-06,
346
+ "loss": 0.1762,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 0.04,
351
+ "learning_rate": 9.773164556962027e-06,
352
+ "loss": 0.1722,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 0.04,
357
+ "learning_rate": 9.766835443037976e-06,
358
+ "loss": 0.2051,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 0.04,
363
+ "learning_rate": 9.760506329113924e-06,
364
+ "loss": 0.2071,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 0.04,
369
+ "learning_rate": 9.754177215189873e-06,
370
+ "loss": 0.1664,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 0.04,
375
+ "learning_rate": 9.747848101265823e-06,
376
+ "loss": 0.1871,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 0.04,
381
+ "learning_rate": 9.741518987341772e-06,
382
+ "loss": 0.1893,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 0.04,
387
+ "learning_rate": 9.735189873417721e-06,
388
+ "loss": 0.1756,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 0.04,
393
+ "learning_rate": 9.728860759493671e-06,
394
+ "loss": 0.1687,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 0.04,
399
+ "learning_rate": 9.72253164556962e-06,
400
+ "loss": 0.1685,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 0.04,
405
+ "learning_rate": 9.71620253164557e-06,
406
+ "loss": 0.1805,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 0.04,
411
+ "learning_rate": 9.70987341772152e-06,
412
+ "loss": 0.1868,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 0.04,
417
+ "learning_rate": 9.70354430379747e-06,
418
+ "loss": 0.212,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 0.04,
423
+ "learning_rate": 9.69721518987342e-06,
424
+ "loss": 0.2019,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 0.04,
429
+ "learning_rate": 9.69088607594937e-06,
430
+ "loss": 0.2076,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 0.04,
435
+ "learning_rate": 9.684556962025317e-06,
436
+ "loss": 0.1971,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 1.0,
441
+ "learning_rate": 9.678227848101266e-06,
442
+ "loss": 0.2424,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 1.0,
447
+ "learning_rate": 9.671898734177216e-06,
448
+ "loss": 0.2346,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 1.0,
453
+ "learning_rate": 9.665569620253165e-06,
454
+ "loss": 0.2436,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 1.0,
459
+ "learning_rate": 9.659240506329115e-06,
460
+ "loss": 0.2613,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 1.0,
465
+ "learning_rate": 9.652911392405064e-06,
466
+ "loss": 0.2781,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 1.0,
471
+ "learning_rate": 9.646582278481013e-06,
472
+ "loss": 0.2465,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 1.0,
477
+ "learning_rate": 9.640253164556963e-06,
478
+ "loss": 0.2386,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 1.0,
483
+ "learning_rate": 9.633924050632912e-06,
484
+ "loss": 0.2298,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 1.01,
489
+ "learning_rate": 9.627594936708862e-06,
490
+ "loss": 0.264,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 1.01,
495
+ "learning_rate": 9.621265822784811e-06,
496
+ "loss": 0.2077,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 1.01,
501
+ "eval_loss": 0.25397753715515137,
502
+ "eval_runtime": 1477.9319,
503
+ "eval_samples_per_second": 7.044,
504
+ "eval_steps_per_second": 0.44,
505
+ "eval_wer": 35.12252231743606,
506
+ "step": 2000
507
+ },
508
+ {
509
+ "epoch": 1.01,
510
+ "learning_rate": 9.61493670886076e-06,
511
+ "loss": 0.192,
512
+ "step": 2025
513
+ },
514
+ {
515
+ "epoch": 1.01,
516
+ "learning_rate": 9.60860759493671e-06,
517
+ "loss": 0.1903,
518
+ "step": 2050
519
+ },
520
+ {
521
+ "epoch": 1.01,
522
+ "learning_rate": 9.60227848101266e-06,
523
+ "loss": 0.2135,
524
+ "step": 2075
525
+ },
526
+ {
527
+ "epoch": 1.01,
528
+ "learning_rate": 9.595949367088609e-06,
529
+ "loss": 0.1917,
530
+ "step": 2100
531
+ },
532
+ {
533
+ "epoch": 1.01,
534
+ "learning_rate": 9.589620253164558e-06,
535
+ "loss": 0.1579,
536
+ "step": 2125
537
+ },
538
+ {
539
+ "epoch": 1.01,
540
+ "learning_rate": 9.583291139240508e-06,
541
+ "loss": 0.1779,
542
+ "step": 2150
543
+ },
544
+ {
545
+ "epoch": 1.01,
546
+ "learning_rate": 9.576962025316457e-06,
547
+ "loss": 0.1504,
548
+ "step": 2175
549
+ },
550
+ {
551
+ "epoch": 1.01,
552
+ "learning_rate": 9.570632911392405e-06,
553
+ "loss": 0.1574,
554
+ "step": 2200
555
+ },
556
+ {
557
+ "epoch": 1.01,
558
+ "learning_rate": 9.564303797468354e-06,
559
+ "loss": 0.1757,
560
+ "step": 2225
561
+ },
562
+ {
563
+ "epoch": 1.01,
564
+ "learning_rate": 9.557974683544304e-06,
565
+ "loss": 0.1996,
566
+ "step": 2250
567
+ },
568
+ {
569
+ "epoch": 1.01,
570
+ "learning_rate": 9.551645569620253e-06,
571
+ "loss": 0.1716,
572
+ "step": 2275
573
+ },
574
+ {
575
+ "epoch": 1.01,
576
+ "learning_rate": 9.545316455696203e-06,
577
+ "loss": 0.1509,
578
+ "step": 2300
579
+ },
580
+ {
581
+ "epoch": 1.01,
582
+ "learning_rate": 9.538987341772152e-06,
583
+ "loss": 0.1402,
584
+ "step": 2325
585
+ },
586
+ {
587
+ "epoch": 1.01,
588
+ "learning_rate": 9.532658227848101e-06,
589
+ "loss": 0.1612,
590
+ "step": 2350
591
+ },
592
+ {
593
+ "epoch": 1.02,
594
+ "learning_rate": 9.526329113924051e-06,
595
+ "loss": 0.1357,
596
+ "step": 2375
597
+ },
598
+ {
599
+ "epoch": 1.02,
600
+ "learning_rate": 9.52e-06,
601
+ "loss": 0.1426,
602
+ "step": 2400
603
+ },
604
+ {
605
+ "epoch": 1.02,
606
+ "learning_rate": 9.51367088607595e-06,
607
+ "loss": 0.1644,
608
+ "step": 2425
609
+ },
610
+ {
611
+ "epoch": 1.02,
612
+ "learning_rate": 9.5073417721519e-06,
613
+ "loss": 0.1196,
614
+ "step": 2450
615
+ },
616
+ {
617
+ "epoch": 1.02,
618
+ "learning_rate": 9.501012658227849e-06,
619
+ "loss": 0.1332,
620
+ "step": 2475
621
+ },
622
+ {
623
+ "epoch": 1.02,
624
+ "learning_rate": 9.494683544303798e-06,
625
+ "loss": 0.1096,
626
+ "step": 2500
627
+ },
628
+ {
629
+ "epoch": 1.02,
630
+ "learning_rate": 9.488354430379747e-06,
631
+ "loss": 0.1263,
632
+ "step": 2525
633
+ },
634
+ {
635
+ "epoch": 1.02,
636
+ "learning_rate": 9.482025316455697e-06,
637
+ "loss": 0.1467,
638
+ "step": 2550
639
+ },
640
+ {
641
+ "epoch": 1.02,
642
+ "learning_rate": 9.475696202531646e-06,
643
+ "loss": 0.1372,
644
+ "step": 2575
645
+ },
646
+ {
647
+ "epoch": 1.02,
648
+ "learning_rate": 9.469367088607596e-06,
649
+ "loss": 0.1256,
650
+ "step": 2600
651
+ },
652
+ {
653
+ "epoch": 1.02,
654
+ "learning_rate": 9.463037974683545e-06,
655
+ "loss": 0.1319,
656
+ "step": 2625
657
+ },
658
+ {
659
+ "epoch": 1.02,
660
+ "learning_rate": 9.456708860759495e-06,
661
+ "loss": 0.1385,
662
+ "step": 2650
663
+ },
664
+ {
665
+ "epoch": 1.02,
666
+ "learning_rate": 9.450379746835444e-06,
667
+ "loss": 0.147,
668
+ "step": 2675
669
+ },
670
+ {
671
+ "epoch": 1.02,
672
+ "learning_rate": 9.444050632911393e-06,
673
+ "loss": 0.1527,
674
+ "step": 2700
675
+ },
676
+ {
677
+ "epoch": 1.02,
678
+ "learning_rate": 9.437721518987343e-06,
679
+ "loss": 0.1357,
680
+ "step": 2725
681
+ },
682
+ {
683
+ "epoch": 1.02,
684
+ "learning_rate": 9.431392405063292e-06,
685
+ "loss": 0.122,
686
+ "step": 2750
687
+ },
688
+ {
689
+ "epoch": 1.03,
690
+ "learning_rate": 9.425063291139242e-06,
691
+ "loss": 0.0918,
692
+ "step": 2775
693
+ },
694
+ {
695
+ "epoch": 1.03,
696
+ "learning_rate": 9.418734177215191e-06,
697
+ "loss": 0.1218,
698
+ "step": 2800
699
+ },
700
+ {
701
+ "epoch": 1.03,
702
+ "learning_rate": 9.41240506329114e-06,
703
+ "loss": 0.139,
704
+ "step": 2825
705
+ },
706
+ {
707
+ "epoch": 1.03,
708
+ "learning_rate": 9.40607594936709e-06,
709
+ "loss": 0.1205,
710
+ "step": 2850
711
+ },
712
+ {
713
+ "epoch": 1.03,
714
+ "learning_rate": 9.39974683544304e-06,
715
+ "loss": 0.1158,
716
+ "step": 2875
717
+ },
718
+ {
719
+ "epoch": 1.03,
720
+ "learning_rate": 9.393417721518989e-06,
721
+ "loss": 0.1156,
722
+ "step": 2900
723
+ },
724
+ {
725
+ "epoch": 1.03,
726
+ "learning_rate": 9.387088607594937e-06,
727
+ "loss": 0.1044,
728
+ "step": 2925
729
+ },
730
+ {
731
+ "epoch": 1.03,
732
+ "learning_rate": 9.380759493670886e-06,
733
+ "loss": 0.112,
734
+ "step": 2950
735
+ },
736
+ {
737
+ "epoch": 1.03,
738
+ "learning_rate": 9.374430379746835e-06,
739
+ "loss": 0.1023,
740
+ "step": 2975
741
+ },
742
+ {
743
+ "epoch": 1.03,
744
+ "learning_rate": 9.368101265822785e-06,
745
+ "loss": 0.0966,
746
+ "step": 3000
747
+ },
748
+ {
749
+ "epoch": 1.03,
750
+ "eval_loss": 0.2638459801673889,
751
+ "eval_runtime": 1500.936,
752
+ "eval_samples_per_second": 6.936,
753
+ "eval_steps_per_second": 0.434,
754
+ "eval_wer": 37.93069806013106,
755
+ "step": 3000
756
+ },
757
+ {
758
+ "epoch": 1.03,
759
+ "learning_rate": 9.361772151898734e-06,
760
+ "loss": 0.0942,
761
+ "step": 3025
762
+ },
763
+ {
764
+ "epoch": 1.03,
765
+ "learning_rate": 9.355443037974684e-06,
766
+ "loss": 0.0826,
767
+ "step": 3050
768
+ },
769
+ {
770
+ "epoch": 1.03,
771
+ "learning_rate": 9.349113924050633e-06,
772
+ "loss": 0.086,
773
+ "step": 3075
774
+ },
775
+ {
776
+ "epoch": 1.03,
777
+ "learning_rate": 9.342784810126583e-06,
778
+ "loss": 0.0956,
779
+ "step": 3100
780
+ },
781
+ {
782
+ "epoch": 1.03,
783
+ "learning_rate": 9.336455696202532e-06,
784
+ "loss": 0.0935,
785
+ "step": 3125
786
+ },
787
+ {
788
+ "epoch": 1.03,
789
+ "learning_rate": 9.330126582278481e-06,
790
+ "loss": 0.0923,
791
+ "step": 3150
792
+ },
793
+ {
794
+ "epoch": 1.04,
795
+ "learning_rate": 9.323797468354431e-06,
796
+ "loss": 0.1163,
797
+ "step": 3175
798
+ },
799
+ {
800
+ "epoch": 1.04,
801
+ "learning_rate": 9.31746835443038e-06,
802
+ "loss": 0.1099,
803
+ "step": 3200
804
+ },
805
+ {
806
+ "epoch": 1.04,
807
+ "learning_rate": 9.31113924050633e-06,
808
+ "loss": 0.0879,
809
+ "step": 3225
810
+ },
811
+ {
812
+ "epoch": 1.04,
813
+ "learning_rate": 9.30481012658228e-06,
814
+ "loss": 0.0994,
815
+ "step": 3250
816
+ },
817
+ {
818
+ "epoch": 1.04,
819
+ "learning_rate": 9.298481012658229e-06,
820
+ "loss": 0.1069,
821
+ "step": 3275
822
+ },
823
+ {
824
+ "epoch": 1.04,
825
+ "learning_rate": 9.292151898734178e-06,
826
+ "loss": 0.0957,
827
+ "step": 3300
828
+ },
829
+ {
830
+ "epoch": 1.04,
831
+ "learning_rate": 9.285822784810127e-06,
832
+ "loss": 0.0866,
833
+ "step": 3325
834
+ },
835
+ {
836
+ "epoch": 1.04,
837
+ "learning_rate": 9.279493670886077e-06,
838
+ "loss": 0.0965,
839
+ "step": 3350
840
+ },
841
+ {
842
+ "epoch": 1.04,
843
+ "learning_rate": 9.273164556962026e-06,
844
+ "loss": 0.0952,
845
+ "step": 3375
846
+ },
847
+ {
848
+ "epoch": 1.04,
849
+ "learning_rate": 9.266835443037976e-06,
850
+ "loss": 0.1178,
851
+ "step": 3400
852
+ },
853
+ {
854
+ "epoch": 1.04,
855
+ "learning_rate": 9.260506329113925e-06,
856
+ "loss": 0.1154,
857
+ "step": 3425
858
+ },
859
+ {
860
+ "epoch": 1.04,
861
+ "learning_rate": 9.254177215189875e-06,
862
+ "loss": 0.1075,
863
+ "step": 3450
864
+ },
865
+ {
866
+ "epoch": 1.04,
867
+ "learning_rate": 9.247848101265824e-06,
868
+ "loss": 0.1161,
869
+ "step": 3475
870
+ },
871
+ {
872
+ "epoch": 1.04,
873
+ "learning_rate": 9.241518987341773e-06,
874
+ "loss": 0.1153,
875
+ "step": 3500
876
+ },
877
+ {
878
+ "epoch": 2.0,
879
+ "learning_rate": 9.235189873417723e-06,
880
+ "loss": 0.1343,
881
+ "step": 3525
882
+ },
883
+ {
884
+ "epoch": 2.0,
885
+ "learning_rate": 9.228860759493672e-06,
886
+ "loss": 0.1374,
887
+ "step": 3550
888
+ },
889
+ {
890
+ "epoch": 2.0,
891
+ "learning_rate": 9.222531645569622e-06,
892
+ "loss": 0.1436,
893
+ "step": 3575
894
+ },
895
+ {
896
+ "epoch": 2.0,
897
+ "learning_rate": 9.216202531645571e-06,
898
+ "loss": 0.1561,
899
+ "step": 3600
900
+ },
901
+ {
902
+ "epoch": 2.0,
903
+ "learning_rate": 9.20987341772152e-06,
904
+ "loss": 0.1709,
905
+ "step": 3625
906
+ },
907
+ {
908
+ "epoch": 2.0,
909
+ "learning_rate": 9.20354430379747e-06,
910
+ "loss": 0.142,
911
+ "step": 3650
912
+ },
913
+ {
914
+ "epoch": 2.0,
915
+ "learning_rate": 9.197215189873418e-06,
916
+ "loss": 0.1389,
917
+ "step": 3675
918
+ },
919
+ {
920
+ "epoch": 2.0,
921
+ "learning_rate": 9.190886075949367e-06,
922
+ "loss": 0.1398,
923
+ "step": 3700
924
+ },
925
+ {
926
+ "epoch": 2.01,
927
+ "learning_rate": 9.184556962025317e-06,
928
+ "loss": 0.1526,
929
+ "step": 3725
930
+ },
931
+ {
932
+ "epoch": 2.01,
933
+ "learning_rate": 9.178227848101266e-06,
934
+ "loss": 0.1273,
935
+ "step": 3750
936
+ },
937
+ {
938
+ "epoch": 2.01,
939
+ "learning_rate": 9.171898734177215e-06,
940
+ "loss": 0.1119,
941
+ "step": 3775
942
+ },
943
+ {
944
+ "epoch": 2.01,
945
+ "learning_rate": 9.165569620253165e-06,
946
+ "loss": 0.1088,
947
+ "step": 3800
948
+ },
949
+ {
950
+ "epoch": 2.01,
951
+ "learning_rate": 9.159493670886076e-06,
952
+ "loss": 0.1362,
953
+ "step": 3825
954
+ },
955
+ {
956
+ "epoch": 2.01,
957
+ "learning_rate": 9.153164556962026e-06,
958
+ "loss": 0.1247,
959
+ "step": 3850
960
+ },
961
+ {
962
+ "epoch": 2.01,
963
+ "learning_rate": 9.146835443037975e-06,
964
+ "loss": 0.0974,
965
+ "step": 3875
966
+ },
967
+ {
968
+ "epoch": 2.01,
969
+ "learning_rate": 9.140506329113925e-06,
970
+ "loss": 0.1218,
971
+ "step": 3900
972
+ },
973
+ {
974
+ "epoch": 2.01,
975
+ "learning_rate": 9.134177215189874e-06,
976
+ "loss": 0.0846,
977
+ "step": 3925
978
+ },
979
+ {
980
+ "epoch": 2.01,
981
+ "learning_rate": 9.127848101265823e-06,
982
+ "loss": 0.0964,
983
+ "step": 3950
984
+ },
985
+ {
986
+ "epoch": 2.01,
987
+ "learning_rate": 9.121518987341773e-06,
988
+ "loss": 0.1166,
989
+ "step": 3975
990
+ },
991
+ {
992
+ "epoch": 2.01,
993
+ "learning_rate": 9.115189873417722e-06,
994
+ "loss": 0.1212,
995
+ "step": 4000
996
+ },
997
+ {
998
+ "epoch": 2.01,
999
+ "eval_loss": 0.22517943382263184,
1000
+ "eval_runtime": 1537.556,
1001
+ "eval_samples_per_second": 6.771,
1002
+ "eval_steps_per_second": 0.423,
1003
+ "eval_wer": 69.04990325045935,
1004
+ "step": 4000
1005
+ },
1006
+ {
1007
+ "epoch": 2.01,
1008
+ "learning_rate": 9.108860759493672e-06,
1009
+ "loss": 0.1076,
1010
+ "step": 4025
1011
+ },
1012
+ {
1013
+ "epoch": 2.01,
1014
+ "learning_rate": 9.102531645569621e-06,
1015
+ "loss": 0.086,
1016
+ "step": 4050
1017
+ },
1018
+ {
1019
+ "epoch": 2.01,
1020
+ "learning_rate": 9.09620253164557e-06,
1021
+ "loss": 0.0773,
1022
+ "step": 4075
1023
+ },
1024
+ {
1025
+ "epoch": 2.01,
1026
+ "learning_rate": 9.08987341772152e-06,
1027
+ "loss": 0.098,
1028
+ "step": 4100
1029
+ },
1030
+ {
1031
+ "epoch": 2.02,
1032
+ "learning_rate": 9.08354430379747e-06,
1033
+ "loss": 0.0747,
1034
+ "step": 4125
1035
+ },
1036
+ {
1037
+ "epoch": 2.02,
1038
+ "learning_rate": 9.077215189873419e-06,
1039
+ "loss": 0.0895,
1040
+ "step": 4150
1041
+ },
1042
+ {
1043
+ "epoch": 2.02,
1044
+ "learning_rate": 9.070886075949368e-06,
1045
+ "loss": 0.0957,
1046
+ "step": 4175
1047
+ },
1048
+ {
1049
+ "epoch": 2.02,
1050
+ "learning_rate": 9.064556962025318e-06,
1051
+ "loss": 0.0789,
1052
+ "step": 4200
1053
+ },
1054
+ {
1055
+ "epoch": 2.02,
1056
+ "learning_rate": 9.058227848101267e-06,
1057
+ "loss": 0.0718,
1058
+ "step": 4225
1059
+ },
1060
+ {
1061
+ "epoch": 2.02,
1062
+ "learning_rate": 9.051898734177215e-06,
1063
+ "loss": 0.0737,
1064
+ "step": 4250
1065
+ },
1066
+ {
1067
+ "epoch": 2.02,
1068
+ "learning_rate": 9.045569620253164e-06,
1069
+ "loss": 0.0731,
1070
+ "step": 4275
1071
+ },
1072
+ {
1073
+ "epoch": 2.02,
1074
+ "learning_rate": 9.039240506329114e-06,
1075
+ "loss": 0.087,
1076
+ "step": 4300
1077
+ },
1078
+ {
1079
+ "epoch": 2.02,
1080
+ "learning_rate": 9.032911392405063e-06,
1081
+ "loss": 0.0875,
1082
+ "step": 4325
1083
+ },
1084
+ {
1085
+ "epoch": 2.02,
1086
+ "learning_rate": 9.026582278481013e-06,
1087
+ "loss": 0.0768,
1088
+ "step": 4350
1089
+ },
1090
+ {
1091
+ "epoch": 2.02,
1092
+ "learning_rate": 9.020253164556962e-06,
1093
+ "loss": 0.0842,
1094
+ "step": 4375
1095
+ },
1096
+ {
1097
+ "epoch": 2.02,
1098
+ "learning_rate": 9.013924050632911e-06,
1099
+ "loss": 0.092,
1100
+ "step": 4400
1101
+ },
1102
+ {
1103
+ "epoch": 2.02,
1104
+ "learning_rate": 9.007594936708861e-06,
1105
+ "loss": 0.0949,
1106
+ "step": 4425
1107
+ },
1108
+ {
1109
+ "epoch": 2.02,
1110
+ "learning_rate": 9.00126582278481e-06,
1111
+ "loss": 0.0989,
1112
+ "step": 4450
1113
+ },
1114
+ {
1115
+ "epoch": 2.02,
1116
+ "learning_rate": 8.99493670886076e-06,
1117
+ "loss": 0.0908,
1118
+ "step": 4475
1119
+ },
1120
+ {
1121
+ "epoch": 2.02,
1122
+ "learning_rate": 8.98860759493671e-06,
1123
+ "loss": 0.0783,
1124
+ "step": 4500
1125
+ },
1126
+ {
1127
+ "epoch": 2.03,
1128
+ "learning_rate": 8.982278481012659e-06,
1129
+ "loss": 0.0651,
1130
+ "step": 4525
1131
+ },
1132
+ {
1133
+ "epoch": 2.03,
1134
+ "learning_rate": 8.975949367088608e-06,
1135
+ "loss": 0.0791,
1136
+ "step": 4550
1137
+ },
1138
+ {
1139
+ "epoch": 2.03,
1140
+ "learning_rate": 8.969620253164557e-06,
1141
+ "loss": 0.0829,
1142
+ "step": 4575
1143
+ },
1144
+ {
1145
+ "epoch": 2.03,
1146
+ "learning_rate": 8.963291139240507e-06,
1147
+ "loss": 0.0797,
1148
+ "step": 4600
1149
+ },
1150
+ {
1151
+ "epoch": 2.03,
1152
+ "learning_rate": 8.956962025316456e-06,
1153
+ "loss": 0.0743,
1154
+ "step": 4625
1155
+ },
1156
+ {
1157
+ "epoch": 2.03,
1158
+ "learning_rate": 8.950632911392406e-06,
1159
+ "loss": 0.0814,
1160
+ "step": 4650
1161
+ },
1162
+ {
1163
+ "epoch": 2.03,
1164
+ "learning_rate": 8.944303797468355e-06,
1165
+ "loss": 0.0594,
1166
+ "step": 4675
1167
+ },
1168
+ {
1169
+ "epoch": 2.03,
1170
+ "learning_rate": 8.937974683544305e-06,
1171
+ "loss": 0.0709,
1172
+ "step": 4700
1173
+ },
1174
+ {
1175
+ "epoch": 2.03,
1176
+ "learning_rate": 8.931645569620254e-06,
1177
+ "loss": 0.0621,
1178
+ "step": 4725
1179
+ },
1180
+ {
1181
+ "epoch": 2.03,
1182
+ "learning_rate": 8.925316455696203e-06,
1183
+ "loss": 0.0599,
1184
+ "step": 4750
1185
+ },
1186
+ {
1187
+ "epoch": 2.03,
1188
+ "learning_rate": 8.918987341772153e-06,
1189
+ "loss": 0.0625,
1190
+ "step": 4775
1191
+ },
1192
+ {
1193
+ "epoch": 2.03,
1194
+ "learning_rate": 8.912658227848102e-06,
1195
+ "loss": 0.0541,
1196
+ "step": 4800
1197
+ },
1198
+ {
1199
+ "epoch": 2.03,
1200
+ "learning_rate": 8.906329113924052e-06,
1201
+ "loss": 0.059,
1202
+ "step": 4825
1203
+ },
1204
+ {
1205
+ "epoch": 2.03,
1206
+ "learning_rate": 8.900000000000001e-06,
1207
+ "loss": 0.0576,
1208
+ "step": 4850
1209
+ },
1210
+ {
1211
+ "epoch": 2.03,
1212
+ "learning_rate": 8.89367088607595e-06,
1213
+ "loss": 0.0623,
1214
+ "step": 4875
1215
+ },
1216
+ {
1217
+ "epoch": 2.03,
1218
+ "learning_rate": 8.8873417721519e-06,
1219
+ "loss": 0.0592,
1220
+ "step": 4900
1221
+ },
1222
+ {
1223
+ "epoch": 2.04,
1224
+ "learning_rate": 8.88101265822785e-06,
1225
+ "loss": 0.0692,
1226
+ "step": 4925
1227
+ },
1228
+ {
1229
+ "epoch": 2.04,
1230
+ "learning_rate": 8.874683544303799e-06,
1231
+ "loss": 0.077,
1232
+ "step": 4950
1233
+ },
1234
+ {
1235
+ "epoch": 2.04,
1236
+ "learning_rate": 8.868354430379748e-06,
1237
+ "loss": 0.0662,
1238
+ "step": 4975
1239
+ },
1240
+ {
1241
+ "epoch": 2.04,
1242
+ "learning_rate": 8.862025316455696e-06,
1243
+ "loss": 0.0605,
1244
+ "step": 5000
1245
+ },
1246
+ {
1247
+ "epoch": 2.04,
1248
+ "eval_loss": 0.2468421906232834,
1249
+ "eval_runtime": 1442.9116,
1250
+ "eval_samples_per_second": 7.215,
1251
+ "eval_steps_per_second": 0.451,
1252
+ "eval_wer": 33.46883689165677,
1253
+ "step": 5000
1254
+ },
1255
+ {
1256
+ "epoch": 2.04,
1257
+ "learning_rate": 8.855696202531645e-06,
1258
+ "loss": 0.0769,
1259
+ "step": 5025
1260
+ },
1261
+ {
1262
+ "epoch": 2.04,
1263
+ "learning_rate": 8.849367088607595e-06,
1264
+ "loss": 0.0601,
1265
+ "step": 5050
1266
+ },
1267
+ {
1268
+ "epoch": 2.04,
1269
+ "learning_rate": 8.843037974683544e-06,
1270
+ "loss": 0.0698,
1271
+ "step": 5075
1272
+ },
1273
+ {
1274
+ "epoch": 2.04,
1275
+ "learning_rate": 8.836708860759494e-06,
1276
+ "loss": 0.0593,
1277
+ "step": 5100
1278
+ },
1279
+ {
1280
+ "epoch": 2.04,
1281
+ "learning_rate": 8.830379746835443e-06,
1282
+ "loss": 0.0521,
1283
+ "step": 5125
1284
+ },
1285
+ {
1286
+ "epoch": 2.04,
1287
+ "learning_rate": 8.824050632911393e-06,
1288
+ "loss": 0.0687,
1289
+ "step": 5150
1290
+ },
1291
+ {
1292
+ "epoch": 2.04,
1293
+ "learning_rate": 8.817721518987342e-06,
1294
+ "loss": 0.0818,
1295
+ "step": 5175
1296
+ },
1297
+ {
1298
+ "epoch": 2.04,
1299
+ "learning_rate": 8.811392405063291e-06,
1300
+ "loss": 0.0706,
1301
+ "step": 5200
1302
+ },
1303
+ {
1304
+ "epoch": 2.04,
1305
+ "learning_rate": 8.805063291139241e-06,
1306
+ "loss": 0.0734,
1307
+ "step": 5225
1308
+ },
1309
+ {
1310
+ "epoch": 2.04,
1311
+ "learning_rate": 8.79873417721519e-06,
1312
+ "loss": 0.0795,
1313
+ "step": 5250
1314
+ },
1315
+ {
1316
+ "epoch": 3.0,
1317
+ "learning_rate": 8.79240506329114e-06,
1318
+ "loss": 0.0905,
1319
+ "step": 5275
1320
+ },
1321
+ {
1322
+ "epoch": 3.0,
1323
+ "learning_rate": 8.78607594936709e-06,
1324
+ "loss": 0.086,
1325
+ "step": 5300
1326
+ },
1327
+ {
1328
+ "epoch": 3.0,
1329
+ "learning_rate": 8.779746835443039e-06,
1330
+ "loss": 0.105,
1331
+ "step": 5325
1332
+ },
1333
+ {
1334
+ "epoch": 3.0,
1335
+ "learning_rate": 8.773417721518988e-06,
1336
+ "loss": 0.0981,
1337
+ "step": 5350
1338
+ },
1339
+ {
1340
+ "epoch": 3.0,
1341
+ "learning_rate": 8.767088607594937e-06,
1342
+ "loss": 0.1131,
1343
+ "step": 5375
1344
+ },
1345
+ {
1346
+ "epoch": 3.0,
1347
+ "learning_rate": 8.760759493670887e-06,
1348
+ "loss": 0.0916,
1349
+ "step": 5400
1350
+ },
1351
+ {
1352
+ "epoch": 3.0,
1353
+ "learning_rate": 8.754430379746836e-06,
1354
+ "loss": 0.0974,
1355
+ "step": 5425
1356
+ },
1357
+ {
1358
+ "epoch": 3.0,
1359
+ "learning_rate": 8.748101265822786e-06,
1360
+ "loss": 0.0899,
1361
+ "step": 5450
1362
+ },
1363
+ {
1364
+ "epoch": 3.01,
1365
+ "learning_rate": 8.741772151898735e-06,
1366
+ "loss": 0.1104,
1367
+ "step": 5475
1368
+ },
1369
+ {
1370
+ "epoch": 3.01,
1371
+ "learning_rate": 8.735443037974685e-06,
1372
+ "loss": 0.0818,
1373
+ "step": 5500
1374
+ },
1375
+ {
1376
+ "epoch": 3.01,
1377
+ "learning_rate": 8.729113924050634e-06,
1378
+ "loss": 0.0764,
1379
+ "step": 5525
1380
+ },
1381
+ {
1382
+ "epoch": 3.01,
1383
+ "learning_rate": 8.722784810126583e-06,
1384
+ "loss": 0.0709,
1385
+ "step": 5550
1386
+ },
1387
+ {
1388
+ "epoch": 3.01,
1389
+ "learning_rate": 8.716455696202533e-06,
1390
+ "loss": 0.0891,
1391
+ "step": 5575
1392
+ },
1393
+ {
1394
+ "epoch": 3.01,
1395
+ "learning_rate": 8.710126582278482e-06,
1396
+ "loss": 0.0869,
1397
+ "step": 5600
1398
+ },
1399
+ {
1400
+ "epoch": 3.01,
1401
+ "learning_rate": 8.703797468354432e-06,
1402
+ "loss": 0.0631,
1403
+ "step": 5625
1404
+ },
1405
+ {
1406
+ "epoch": 3.01,
1407
+ "learning_rate": 8.697468354430381e-06,
1408
+ "loss": 0.0795,
1409
+ "step": 5650
1410
+ },
1411
+ {
1412
+ "epoch": 3.01,
1413
+ "learning_rate": 8.69113924050633e-06,
1414
+ "loss": 0.0603,
1415
+ "step": 5675
1416
+ },
1417
+ {
1418
+ "epoch": 3.01,
1419
+ "learning_rate": 8.68481012658228e-06,
1420
+ "loss": 0.064,
1421
+ "step": 5700
1422
+ },
1423
+ {
1424
+ "epoch": 3.01,
1425
+ "learning_rate": 8.67848101265823e-06,
1426
+ "loss": 0.0738,
1427
+ "step": 5725
1428
+ },
1429
+ {
1430
+ "epoch": 3.01,
1431
+ "learning_rate": 8.672151898734177e-06,
1432
+ "loss": 0.0796,
1433
+ "step": 5750
1434
+ },
1435
+ {
1436
+ "epoch": 3.01,
1437
+ "learning_rate": 8.665822784810127e-06,
1438
+ "loss": 0.0723,
1439
+ "step": 5775
1440
+ },
1441
+ {
1442
+ "epoch": 3.01,
1443
+ "learning_rate": 8.659493670886076e-06,
1444
+ "loss": 0.0592,
1445
+ "step": 5800
1446
+ },
1447
+ {
1448
+ "epoch": 3.01,
1449
+ "learning_rate": 8.653164556962025e-06,
1450
+ "loss": 0.051,
1451
+ "step": 5825
1452
+ },
1453
+ {
1454
+ "epoch": 3.01,
1455
+ "learning_rate": 8.646835443037975e-06,
1456
+ "loss": 0.0716,
1457
+ "step": 5850
1458
+ },
1459
+ {
1460
+ "epoch": 3.02,
1461
+ "learning_rate": 8.640506329113924e-06,
1462
+ "loss": 0.0598,
1463
+ "step": 5875
1464
+ },
1465
+ {
1466
+ "epoch": 3.02,
1467
+ "learning_rate": 8.634177215189874e-06,
1468
+ "loss": 0.0524,
1469
+ "step": 5900
1470
+ },
1471
+ {
1472
+ "epoch": 3.02,
1473
+ "learning_rate": 8.627848101265823e-06,
1474
+ "loss": 0.0623,
1475
+ "step": 5925
1476
+ },
1477
+ {
1478
+ "epoch": 3.02,
1479
+ "learning_rate": 8.621518987341773e-06,
1480
+ "loss": 0.054,
1481
+ "step": 5950
1482
+ },
1483
+ {
1484
+ "epoch": 3.02,
1485
+ "learning_rate": 8.615189873417722e-06,
1486
+ "loss": 0.057,
1487
+ "step": 5975
1488
+ },
1489
+ {
1490
+ "epoch": 3.02,
1491
+ "learning_rate": 8.608860759493671e-06,
1492
+ "loss": 0.0479,
1493
+ "step": 6000
1494
+ },
1495
+ {
1496
+ "epoch": 3.02,
1497
+ "eval_loss": 0.24428367614746094,
1498
+ "eval_runtime": 1457.5568,
1499
+ "eval_samples_per_second": 7.143,
1500
+ "eval_steps_per_second": 0.447,
1501
+ "eval_wer": 36.51929299663409,
1502
+ "step": 6000
1503
+ },
1504
+ {
1505
+ "epoch": 3.02,
1506
+ "learning_rate": 8.602531645569621e-06,
1507
+ "loss": 0.0501,
1508
+ "step": 6025
1509
+ },
1510
+ {
1511
+ "epoch": 3.02,
1512
+ "learning_rate": 8.59620253164557e-06,
1513
+ "loss": 0.0527,
1514
+ "step": 6050
1515
+ },
1516
+ {
1517
+ "epoch": 3.02,
1518
+ "learning_rate": 8.58987341772152e-06,
1519
+ "loss": 0.0558,
1520
+ "step": 6075
1521
+ },
1522
+ {
1523
+ "epoch": 3.02,
1524
+ "learning_rate": 8.58354430379747e-06,
1525
+ "loss": 0.0548,
1526
+ "step": 6100
1527
+ },
1528
+ {
1529
+ "epoch": 3.02,
1530
+ "learning_rate": 8.577215189873419e-06,
1531
+ "loss": 0.0503,
1532
+ "step": 6125
1533
+ },
1534
+ {
1535
+ "epoch": 3.02,
1536
+ "learning_rate": 8.570886075949368e-06,
1537
+ "loss": 0.0641,
1538
+ "step": 6150
1539
+ },
1540
+ {
1541
+ "epoch": 3.02,
1542
+ "learning_rate": 8.564556962025316e-06,
1543
+ "loss": 0.0748,
1544
+ "step": 6175
1545
+ },
1546
+ {
1547
+ "epoch": 3.02,
1548
+ "learning_rate": 8.558227848101267e-06,
1549
+ "loss": 0.0775,
1550
+ "step": 6200
1551
+ },
1552
+ {
1553
+ "epoch": 3.02,
1554
+ "learning_rate": 8.551898734177216e-06,
1555
+ "loss": 0.0662,
1556
+ "step": 6225
1557
+ },
1558
+ {
1559
+ "epoch": 3.02,
1560
+ "learning_rate": 8.545569620253166e-06,
1561
+ "loss": 0.0591,
1562
+ "step": 6250
1563
+ },
1564
+ {
1565
+ "epoch": 3.03,
1566
+ "learning_rate": 8.539240506329115e-06,
1567
+ "loss": 0.0421,
1568
+ "step": 6275
1569
+ },
1570
+ {
1571
+ "epoch": 3.03,
1572
+ "learning_rate": 8.532911392405065e-06,
1573
+ "loss": 0.0565,
1574
+ "step": 6300
1575
+ },
1576
+ {
1577
+ "epoch": 3.03,
1578
+ "learning_rate": 8.526582278481014e-06,
1579
+ "loss": 0.0574,
1580
+ "step": 6325
1581
+ },
1582
+ {
1583
+ "epoch": 3.03,
1584
+ "learning_rate": 8.520253164556963e-06,
1585
+ "loss": 0.0466,
1586
+ "step": 6350
1587
+ },
1588
+ {
1589
+ "epoch": 3.03,
1590
+ "learning_rate": 8.513924050632913e-06,
1591
+ "loss": 0.0486,
1592
+ "step": 6375
1593
+ },
1594
+ {
1595
+ "epoch": 3.03,
1596
+ "learning_rate": 8.507594936708862e-06,
1597
+ "loss": 0.0502,
1598
+ "step": 6400
1599
+ },
1600
+ {
1601
+ "epoch": 3.03,
1602
+ "learning_rate": 8.501265822784812e-06,
1603
+ "loss": 0.0415,
1604
+ "step": 6425
1605
+ },
1606
+ {
1607
+ "epoch": 3.03,
1608
+ "learning_rate": 8.494936708860761e-06,
1609
+ "loss": 0.0449,
1610
+ "step": 6450
1611
+ },
1612
+ {
1613
+ "epoch": 3.03,
1614
+ "learning_rate": 8.488607594936709e-06,
1615
+ "loss": 0.0525,
1616
+ "step": 6475
1617
+ },
1618
+ {
1619
+ "epoch": 3.03,
1620
+ "learning_rate": 8.482278481012658e-06,
1621
+ "loss": 0.0401,
1622
+ "step": 6500
1623
+ },
1624
+ {
1625
+ "epoch": 3.03,
1626
+ "learning_rate": 8.475949367088608e-06,
1627
+ "loss": 0.0502,
1628
+ "step": 6525
1629
+ },
1630
+ {
1631
+ "epoch": 3.03,
1632
+ "learning_rate": 8.469620253164557e-06,
1633
+ "loss": 0.0453,
1634
+ "step": 6550
1635
+ },
1636
+ {
1637
+ "epoch": 3.03,
1638
+ "learning_rate": 8.463291139240507e-06,
1639
+ "loss": 0.0442,
1640
+ "step": 6575
1641
+ },
1642
+ {
1643
+ "epoch": 3.03,
1644
+ "learning_rate": 8.456962025316456e-06,
1645
+ "loss": 0.0423,
1646
+ "step": 6600
1647
+ },
1648
+ {
1649
+ "epoch": 3.03,
1650
+ "learning_rate": 8.450632911392405e-06,
1651
+ "loss": 0.0518,
1652
+ "step": 6625
1653
+ },
1654
+ {
1655
+ "epoch": 3.03,
1656
+ "learning_rate": 8.444303797468355e-06,
1657
+ "loss": 0.0479,
1658
+ "step": 6650
1659
+ },
1660
+ {
1661
+ "epoch": 3.04,
1662
+ "learning_rate": 8.437974683544304e-06,
1663
+ "loss": 0.0521,
1664
+ "step": 6675
1665
+ },
1666
+ {
1667
+ "epoch": 3.04,
1668
+ "learning_rate": 8.431645569620254e-06,
1669
+ "loss": 0.0436,
1670
+ "step": 6700
1671
+ },
1672
+ {
1673
+ "epoch": 3.04,
1674
+ "learning_rate": 8.425316455696203e-06,
1675
+ "loss": 0.0539,
1676
+ "step": 6725
1677
+ },
1678
+ {
1679
+ "epoch": 3.04,
1680
+ "learning_rate": 8.418987341772153e-06,
1681
+ "loss": 0.0446,
1682
+ "step": 6750
1683
+ },
1684
+ {
1685
+ "epoch": 3.04,
1686
+ "learning_rate": 8.412658227848102e-06,
1687
+ "loss": 0.0521,
1688
+ "step": 6775
1689
+ },
1690
+ {
1691
+ "epoch": 3.04,
1692
+ "learning_rate": 8.406329113924051e-06,
1693
+ "loss": 0.0391,
1694
+ "step": 6800
1695
+ },
1696
+ {
1697
+ "epoch": 3.04,
1698
+ "learning_rate": 8.400000000000001e-06,
1699
+ "loss": 0.0513,
1700
+ "step": 6825
1701
+ },
1702
+ {
1703
+ "epoch": 3.04,
1704
+ "learning_rate": 8.39367088607595e-06,
1705
+ "loss": 0.0449,
1706
+ "step": 6850
1707
+ },
1708
+ {
1709
+ "epoch": 3.04,
1710
+ "learning_rate": 8.3873417721519e-06,
1711
+ "loss": 0.0463,
1712
+ "step": 6875
1713
+ },
1714
+ {
1715
+ "epoch": 3.04,
1716
+ "learning_rate": 8.38101265822785e-06,
1717
+ "loss": 0.0495,
1718
+ "step": 6900
1719
+ },
1720
+ {
1721
+ "epoch": 3.04,
1722
+ "learning_rate": 8.374683544303797e-06,
1723
+ "loss": 0.0628,
1724
+ "step": 6925
1725
+ },
1726
+ {
1727
+ "epoch": 3.04,
1728
+ "learning_rate": 8.368354430379746e-06,
1729
+ "loss": 0.0594,
1730
+ "step": 6950
1731
+ },
1732
+ {
1733
+ "epoch": 3.04,
1734
+ "learning_rate": 8.362025316455696e-06,
1735
+ "loss": 0.0573,
1736
+ "step": 6975
1737
+ },
1738
+ {
1739
+ "epoch": 3.04,
1740
+ "learning_rate": 8.355696202531645e-06,
1741
+ "loss": 0.057,
1742
+ "step": 7000
1743
+ },
1744
+ {
1745
+ "epoch": 3.04,
1746
+ "eval_loss": 0.25129273533821106,
1747
+ "eval_runtime": 1479.2043,
1748
+ "eval_samples_per_second": 7.038,
1749
+ "eval_steps_per_second": 0.44,
1750
+ "eval_wer": 57.010683100538216,
1751
+ "step": 7000
1752
+ },
1753
+ {
1754
+ "epoch": 4.0,
1755
+ "learning_rate": 8.349367088607596e-06,
1756
+ "loss": 0.0534,
1757
+ "step": 7025
1758
+ },
1759
+ {
1760
+ "epoch": 4.0,
1761
+ "learning_rate": 8.343037974683546e-06,
1762
+ "loss": 0.0663,
1763
+ "step": 7050
1764
+ },
1765
+ {
1766
+ "epoch": 4.0,
1767
+ "learning_rate": 8.336708860759495e-06,
1768
+ "loss": 0.0698,
1769
+ "step": 7075
1770
+ },
1771
+ {
1772
+ "epoch": 4.0,
1773
+ "learning_rate": 8.330379746835445e-06,
1774
+ "loss": 0.0633,
1775
+ "step": 7100
1776
+ },
1777
+ {
1778
+ "epoch": 4.0,
1779
+ "learning_rate": 8.324050632911394e-06,
1780
+ "loss": 0.0808,
1781
+ "step": 7125
1782
+ },
1783
+ {
1784
+ "epoch": 4.0,
1785
+ "learning_rate": 8.317721518987343e-06,
1786
+ "loss": 0.0699,
1787
+ "step": 7150
1788
+ },
1789
+ {
1790
+ "epoch": 4.0,
1791
+ "learning_rate": 8.311392405063293e-06,
1792
+ "loss": 0.0736,
1793
+ "step": 7175
1794
+ },
1795
+ {
1796
+ "epoch": 4.0,
1797
+ "learning_rate": 8.305063291139242e-06,
1798
+ "loss": 0.069,
1799
+ "step": 7200
1800
+ },
1801
+ {
1802
+ "epoch": 4.01,
1803
+ "learning_rate": 8.29873417721519e-06,
1804
+ "loss": 0.0718,
1805
+ "step": 7225
1806
+ },
1807
+ {
1808
+ "epoch": 4.01,
1809
+ "learning_rate": 8.29240506329114e-06,
1810
+ "loss": 0.0674,
1811
+ "step": 7250
1812
+ },
1813
+ {
1814
+ "epoch": 4.01,
1815
+ "learning_rate": 8.286075949367089e-06,
1816
+ "loss": 0.0568,
1817
+ "step": 7275
1818
+ },
1819
+ {
1820
+ "epoch": 4.01,
1821
+ "learning_rate": 8.279746835443038e-06,
1822
+ "loss": 0.0497,
1823
+ "step": 7300
1824
+ },
1825
+ {
1826
+ "epoch": 4.01,
1827
+ "learning_rate": 8.273417721518988e-06,
1828
+ "loss": 0.0563,
1829
+ "step": 7325
1830
+ },
1831
+ {
1832
+ "epoch": 4.01,
1833
+ "learning_rate": 8.267088607594937e-06,
1834
+ "loss": 0.0649,
1835
+ "step": 7350
1836
+ },
1837
+ {
1838
+ "epoch": 4.01,
1839
+ "learning_rate": 8.260759493670887e-06,
1840
+ "loss": 0.042,
1841
+ "step": 7375
1842
+ },
1843
+ {
1844
+ "epoch": 4.01,
1845
+ "learning_rate": 8.254430379746836e-06,
1846
+ "loss": 0.0559,
1847
+ "step": 7400
1848
+ },
1849
+ {
1850
+ "epoch": 4.01,
1851
+ "learning_rate": 8.248101265822785e-06,
1852
+ "loss": 0.0476,
1853
+ "step": 7425
1854
+ },
1855
+ {
1856
+ "epoch": 4.01,
1857
+ "learning_rate": 8.241772151898735e-06,
1858
+ "loss": 0.0427,
1859
+ "step": 7450
1860
+ },
1861
+ {
1862
+ "epoch": 4.01,
1863
+ "learning_rate": 8.235443037974684e-06,
1864
+ "loss": 0.0533,
1865
+ "step": 7475
1866
+ },
1867
+ {
1868
+ "epoch": 4.01,
1869
+ "learning_rate": 8.229113924050634e-06,
1870
+ "loss": 0.0633,
1871
+ "step": 7500
1872
+ },
1873
+ {
1874
+ "epoch": 4.01,
1875
+ "learning_rate": 8.222784810126583e-06,
1876
+ "loss": 0.055,
1877
+ "step": 7525
1878
+ },
1879
+ {
1880
+ "epoch": 4.01,
1881
+ "learning_rate": 8.216455696202533e-06,
1882
+ "loss": 0.043,
1883
+ "step": 7550
1884
+ },
1885
+ {
1886
+ "epoch": 4.01,
1887
+ "learning_rate": 8.210126582278482e-06,
1888
+ "loss": 0.0421,
1889
+ "step": 7575
1890
+ },
1891
+ {
1892
+ "epoch": 4.01,
1893
+ "learning_rate": 8.203797468354431e-06,
1894
+ "loss": 0.0519,
1895
+ "step": 7600
1896
+ },
1897
+ {
1898
+ "epoch": 4.02,
1899
+ "learning_rate": 8.197468354430381e-06,
1900
+ "loss": 0.0398,
1901
+ "step": 7625
1902
+ },
1903
+ {
1904
+ "epoch": 4.02,
1905
+ "learning_rate": 8.19113924050633e-06,
1906
+ "loss": 0.041,
1907
+ "step": 7650
1908
+ },
1909
+ {
1910
+ "epoch": 4.02,
1911
+ "learning_rate": 8.184810126582278e-06,
1912
+ "loss": 0.0477,
1913
+ "step": 7675
1914
+ },
1915
+ {
1916
+ "epoch": 4.02,
1917
+ "learning_rate": 8.178481012658227e-06,
1918
+ "loss": 0.046,
1919
+ "step": 7700
1920
+ },
1921
+ {
1922
+ "epoch": 4.02,
1923
+ "learning_rate": 8.172151898734177e-06,
1924
+ "loss": 0.0465,
1925
+ "step": 7725
1926
+ },
1927
+ {
1928
+ "epoch": 4.02,
1929
+ "learning_rate": 8.165822784810126e-06,
1930
+ "loss": 0.0359,
1931
+ "step": 7750
1932
+ },
1933
+ {
1934
+ "epoch": 4.02,
1935
+ "learning_rate": 8.159493670886076e-06,
1936
+ "loss": 0.0362,
1937
+ "step": 7775
1938
+ },
1939
+ {
1940
+ "epoch": 4.02,
1941
+ "learning_rate": 8.153164556962025e-06,
1942
+ "loss": 0.04,
1943
+ "step": 7800
1944
+ },
1945
+ {
1946
+ "epoch": 4.02,
1947
+ "learning_rate": 8.146835443037975e-06,
1948
+ "loss": 0.0433,
1949
+ "step": 7825
1950
+ },
1951
+ {
1952
+ "epoch": 4.02,
1953
+ "learning_rate": 8.140506329113926e-06,
1954
+ "loss": 0.0399,
1955
+ "step": 7850
1956
+ },
1957
+ {
1958
+ "epoch": 4.02,
1959
+ "learning_rate": 8.134177215189875e-06,
1960
+ "loss": 0.0396,
1961
+ "step": 7875
1962
+ },
1963
+ {
1964
+ "epoch": 4.02,
1965
+ "learning_rate": 8.127848101265825e-06,
1966
+ "loss": 0.0472,
1967
+ "step": 7900
1968
+ },
1969
+ {
1970
+ "epoch": 4.02,
1971
+ "learning_rate": 8.121518987341774e-06,
1972
+ "loss": 0.0533,
1973
+ "step": 7925
1974
+ },
1975
+ {
1976
+ "epoch": 4.02,
1977
+ "learning_rate": 8.115189873417722e-06,
1978
+ "loss": 0.0564,
1979
+ "step": 7950
1980
+ },
1981
+ {
1982
+ "epoch": 4.02,
1983
+ "learning_rate": 8.108860759493671e-06,
1984
+ "loss": 0.0508,
1985
+ "step": 7975
1986
+ },
1987
+ {
1988
+ "epoch": 4.02,
1989
+ "learning_rate": 8.10253164556962e-06,
1990
+ "loss": 0.0375,
1991
+ "step": 8000
1992
+ },
1993
+ {
1994
+ "epoch": 4.02,
1995
+ "eval_loss": 0.252152681350708,
1996
+ "eval_runtime": 1456.2607,
1997
+ "eval_samples_per_second": 7.149,
1998
+ "eval_steps_per_second": 0.447,
1999
+ "eval_wer": 44.38120945056017,
2000
+ "step": 8000
2001
+ },
2002
+ {
2003
+ "epoch": 4.03,
2004
+ "learning_rate": 8.09620253164557e-06,
2005
+ "loss": 0.0296,
2006
+ "step": 8025
2007
+ },
2008
+ {
2009
+ "epoch": 4.03,
2010
+ "learning_rate": 8.08987341772152e-06,
2011
+ "loss": 0.0395,
2012
+ "step": 8050
2013
+ },
2014
+ {
2015
+ "epoch": 4.03,
2016
+ "learning_rate": 8.083544303797469e-06,
2017
+ "loss": 0.0409,
2018
+ "step": 8075
2019
+ },
2020
+ {
2021
+ "epoch": 4.03,
2022
+ "learning_rate": 8.077215189873418e-06,
2023
+ "loss": 0.0343,
2024
+ "step": 8100
2025
+ },
2026
+ {
2027
+ "epoch": 4.03,
2028
+ "learning_rate": 8.070886075949368e-06,
2029
+ "loss": 0.0321,
2030
+ "step": 8125
2031
+ },
2032
+ {
2033
+ "epoch": 4.03,
2034
+ "learning_rate": 8.064556962025317e-06,
2035
+ "loss": 0.0333,
2036
+ "step": 8150
2037
+ },
2038
+ {
2039
+ "epoch": 4.03,
2040
+ "learning_rate": 8.058227848101267e-06,
2041
+ "loss": 0.0327,
2042
+ "step": 8175
2043
+ },
2044
+ {
2045
+ "epoch": 4.03,
2046
+ "learning_rate": 8.051898734177216e-06,
2047
+ "loss": 0.0318,
2048
+ "step": 8200
2049
+ },
2050
+ {
2051
+ "epoch": 4.03,
2052
+ "learning_rate": 8.045569620253165e-06,
2053
+ "loss": 0.0405,
2054
+ "step": 8225
2055
+ },
2056
+ {
2057
+ "epoch": 4.03,
2058
+ "learning_rate": 8.039240506329115e-06,
2059
+ "loss": 0.0382,
2060
+ "step": 8250
2061
+ },
2062
+ {
2063
+ "epoch": 4.03,
2064
+ "learning_rate": 8.032911392405064e-06,
2065
+ "loss": 0.0317,
2066
+ "step": 8275
2067
+ },
2068
+ {
2069
+ "epoch": 4.03,
2070
+ "learning_rate": 8.026582278481014e-06,
2071
+ "loss": 0.0285,
2072
+ "step": 8300
2073
+ },
2074
+ {
2075
+ "epoch": 4.03,
2076
+ "learning_rate": 8.020253164556963e-06,
2077
+ "loss": 0.0285,
2078
+ "step": 8325
2079
+ },
2080
+ {
2081
+ "epoch": 4.03,
2082
+ "learning_rate": 8.013924050632913e-06,
2083
+ "loss": 0.0269,
2084
+ "step": 8350
2085
+ },
2086
+ {
2087
+ "epoch": 4.03,
2088
+ "learning_rate": 8.007594936708862e-06,
2089
+ "loss": 0.0334,
2090
+ "step": 8375
2091
+ },
2092
+ {
2093
+ "epoch": 4.03,
2094
+ "learning_rate": 8.00126582278481e-06,
2095
+ "loss": 0.0406,
2096
+ "step": 8400
2097
+ },
2098
+ {
2099
+ "epoch": 4.04,
2100
+ "learning_rate": 7.99493670886076e-06,
2101
+ "loss": 0.038,
2102
+ "step": 8425
2103
+ },
2104
+ {
2105
+ "epoch": 4.04,
2106
+ "learning_rate": 7.988607594936709e-06,
2107
+ "loss": 0.0315,
2108
+ "step": 8450
2109
+ },
2110
+ {
2111
+ "epoch": 4.04,
2112
+ "learning_rate": 7.982278481012658e-06,
2113
+ "loss": 0.0329,
2114
+ "step": 8475
2115
+ },
2116
+ {
2117
+ "epoch": 4.04,
2118
+ "learning_rate": 7.975949367088607e-06,
2119
+ "loss": 0.0373,
2120
+ "step": 8500
2121
+ },
2122
+ {
2123
+ "epoch": 4.04,
2124
+ "learning_rate": 7.969620253164557e-06,
2125
+ "loss": 0.0514,
2126
+ "step": 8525
2127
+ },
2128
+ {
2129
+ "epoch": 4.04,
2130
+ "learning_rate": 7.963291139240506e-06,
2131
+ "loss": 0.0349,
2132
+ "step": 8550
2133
+ },
2134
+ {
2135
+ "epoch": 4.04,
2136
+ "learning_rate": 7.956962025316456e-06,
2137
+ "loss": 0.0385,
2138
+ "step": 8575
2139
+ },
2140
+ {
2141
+ "epoch": 4.04,
2142
+ "learning_rate": 7.950632911392405e-06,
2143
+ "loss": 0.0367,
2144
+ "step": 8600
2145
+ },
2146
+ {
2147
+ "epoch": 4.04,
2148
+ "learning_rate": 7.944303797468355e-06,
2149
+ "loss": 0.0317,
2150
+ "step": 8625
2151
+ },
2152
+ {
2153
+ "epoch": 4.04,
2154
+ "learning_rate": 7.937974683544304e-06,
2155
+ "loss": 0.0368,
2156
+ "step": 8650
2157
+ },
2158
+ {
2159
+ "epoch": 4.04,
2160
+ "learning_rate": 7.931645569620255e-06,
2161
+ "loss": 0.045,
2162
+ "step": 8675
2163
+ },
2164
+ {
2165
+ "epoch": 4.04,
2166
+ "learning_rate": 7.925316455696203e-06,
2167
+ "loss": 0.0456,
2168
+ "step": 8700
2169
+ },
2170
+ {
2171
+ "epoch": 4.04,
2172
+ "learning_rate": 7.918987341772152e-06,
2173
+ "loss": 0.0421,
2174
+ "step": 8725
2175
+ },
2176
+ {
2177
+ "epoch": 4.04,
2178
+ "learning_rate": 7.912658227848102e-06,
2179
+ "loss": 0.043,
2180
+ "step": 8750
2181
+ },
2182
+ {
2183
+ "epoch": 5.0,
2184
+ "learning_rate": 7.906329113924051e-06,
2185
+ "loss": 0.0428,
2186
+ "step": 8775
2187
+ },
2188
+ {
2189
+ "epoch": 5.0,
2190
+ "learning_rate": 7.9e-06,
2191
+ "loss": 0.0459,
2192
+ "step": 8800
2193
+ },
2194
+ {
2195
+ "epoch": 5.0,
2196
+ "learning_rate": 7.89367088607595e-06,
2197
+ "loss": 0.0467,
2198
+ "step": 8825
2199
+ },
2200
+ {
2201
+ "epoch": 5.0,
2202
+ "learning_rate": 7.8873417721519e-06,
2203
+ "loss": 0.0462,
2204
+ "step": 8850
2205
+ },
2206
+ {
2207
+ "epoch": 5.0,
2208
+ "learning_rate": 7.881012658227849e-06,
2209
+ "loss": 0.0578,
2210
+ "step": 8875
2211
+ },
2212
+ {
2213
+ "epoch": 5.0,
2214
+ "learning_rate": 7.874683544303798e-06,
2215
+ "loss": 0.0422,
2216
+ "step": 8900
2217
+ },
2218
+ {
2219
+ "epoch": 5.0,
2220
+ "learning_rate": 7.868354430379748e-06,
2221
+ "loss": 0.0546,
2222
+ "step": 8925
2223
+ },
2224
+ {
2225
+ "epoch": 5.0,
2226
+ "learning_rate": 7.862025316455697e-06,
2227
+ "loss": 0.0515,
2228
+ "step": 8950
2229
+ },
2230
+ {
2231
+ "epoch": 5.01,
2232
+ "learning_rate": 7.855696202531647e-06,
2233
+ "loss": 0.0541,
2234
+ "step": 8975
2235
+ },
2236
+ {
2237
+ "epoch": 5.01,
2238
+ "learning_rate": 7.849367088607596e-06,
2239
+ "loss": 0.0479,
2240
+ "step": 9000
2241
+ },
2242
+ {
2243
+ "epoch": 5.01,
2244
+ "eval_loss": 0.2375001162290573,
2245
+ "eval_runtime": 1456.1088,
2246
+ "eval_samples_per_second": 7.15,
2247
+ "eval_steps_per_second": 0.447,
2248
+ "eval_wer": 31.58262735979447,
2249
+ "step": 9000
2250
+ },
2251
+ {
2252
+ "epoch": 5.01,
2253
+ "learning_rate": 7.843037974683545e-06,
2254
+ "loss": 0.0434,
2255
+ "step": 9025
2256
+ },
2257
+ {
2258
+ "epoch": 5.01,
2259
+ "learning_rate": 7.836708860759495e-06,
2260
+ "loss": 0.033,
2261
+ "step": 9050
2262
+ },
2263
+ {
2264
+ "epoch": 5.01,
2265
+ "learning_rate": 7.830379746835444e-06,
2266
+ "loss": 0.0494,
2267
+ "step": 9075
2268
+ },
2269
+ {
2270
+ "epoch": 5.01,
2271
+ "learning_rate": 7.824050632911394e-06,
2272
+ "loss": 0.0522,
2273
+ "step": 9100
2274
+ },
2275
+ {
2276
+ "epoch": 5.01,
2277
+ "learning_rate": 7.817721518987343e-06,
2278
+ "loss": 0.0338,
2279
+ "step": 9125
2280
+ },
2281
+ {
2282
+ "epoch": 5.01,
2283
+ "learning_rate": 7.811392405063291e-06,
2284
+ "loss": 0.0395,
2285
+ "step": 9150
2286
+ },
2287
+ {
2288
+ "epoch": 5.01,
2289
+ "learning_rate": 7.80506329113924e-06,
2290
+ "loss": 0.0332,
2291
+ "step": 9175
2292
+ },
2293
+ {
2294
+ "epoch": 5.01,
2295
+ "learning_rate": 7.79873417721519e-06,
2296
+ "loss": 0.0381,
2297
+ "step": 9200
2298
+ },
2299
+ {
2300
+ "epoch": 5.01,
2301
+ "learning_rate": 7.79240506329114e-06,
2302
+ "loss": 0.0409,
2303
+ "step": 9225
2304
+ },
2305
+ {
2306
+ "epoch": 5.01,
2307
+ "learning_rate": 7.786075949367089e-06,
2308
+ "loss": 0.0529,
2309
+ "step": 9250
2310
+ },
2311
+ {
2312
+ "epoch": 5.01,
2313
+ "learning_rate": 7.779746835443038e-06,
2314
+ "loss": 0.0498,
2315
+ "step": 9275
2316
+ },
2317
+ {
2318
+ "epoch": 5.01,
2319
+ "learning_rate": 7.773417721518987e-06,
2320
+ "loss": 0.0379,
2321
+ "step": 9300
2322
+ },
2323
+ {
2324
+ "epoch": 5.01,
2325
+ "learning_rate": 7.767088607594937e-06,
2326
+ "loss": 0.0316,
2327
+ "step": 9325
2328
+ },
2329
+ {
2330
+ "epoch": 5.01,
2331
+ "learning_rate": 7.760759493670886e-06,
2332
+ "loss": 0.0398,
2333
+ "step": 9350
2334
+ },
2335
+ {
2336
+ "epoch": 5.02,
2337
+ "learning_rate": 7.754430379746836e-06,
2338
+ "loss": 0.0427,
2339
+ "step": 9375
2340
+ },
2341
+ {
2342
+ "epoch": 5.02,
2343
+ "learning_rate": 7.748101265822785e-06,
2344
+ "loss": 0.0363,
2345
+ "step": 9400
2346
+ },
2347
+ {
2348
+ "epoch": 5.02,
2349
+ "learning_rate": 7.741772151898735e-06,
2350
+ "loss": 0.0393,
2351
+ "step": 9425
2352
+ },
2353
+ {
2354
+ "epoch": 5.02,
2355
+ "learning_rate": 7.735443037974684e-06,
2356
+ "loss": 0.0377,
2357
+ "step": 9450
2358
+ },
2359
+ {
2360
+ "epoch": 5.02,
2361
+ "learning_rate": 7.729113924050633e-06,
2362
+ "loss": 0.0313,
2363
+ "step": 9475
2364
+ },
2365
+ {
2366
+ "epoch": 5.02,
2367
+ "learning_rate": 7.722784810126583e-06,
2368
+ "loss": 0.0359,
2369
+ "step": 9500
2370
+ },
2371
+ {
2372
+ "epoch": 5.02,
2373
+ "learning_rate": 7.716455696202532e-06,
2374
+ "loss": 0.0289,
2375
+ "step": 9525
2376
+ },
2377
+ {
2378
+ "epoch": 5.02,
2379
+ "learning_rate": 7.710126582278482e-06,
2380
+ "loss": 0.0354,
2381
+ "step": 9550
2382
+ },
2383
+ {
2384
+ "epoch": 5.02,
2385
+ "learning_rate": 7.703797468354431e-06,
2386
+ "loss": 0.0309,
2387
+ "step": 9575
2388
+ },
2389
+ {
2390
+ "epoch": 5.02,
2391
+ "learning_rate": 7.69746835443038e-06,
2392
+ "loss": 0.0321,
2393
+ "step": 9600
2394
+ },
2395
+ {
2396
+ "epoch": 5.02,
2397
+ "learning_rate": 7.69113924050633e-06,
2398
+ "loss": 0.0392,
2399
+ "step": 9625
2400
+ },
2401
+ {
2402
+ "epoch": 5.02,
2403
+ "learning_rate": 7.68481012658228e-06,
2404
+ "loss": 0.0362,
2405
+ "step": 9650
2406
+ },
2407
+ {
2408
+ "epoch": 5.02,
2409
+ "learning_rate": 7.678481012658229e-06,
2410
+ "loss": 0.0358,
2411
+ "step": 9675
2412
+ },
2413
+ {
2414
+ "epoch": 5.02,
2415
+ "learning_rate": 7.672151898734178e-06,
2416
+ "loss": 0.045,
2417
+ "step": 9700
2418
+ },
2419
+ {
2420
+ "epoch": 5.02,
2421
+ "learning_rate": 7.665822784810128e-06,
2422
+ "loss": 0.0401,
2423
+ "step": 9725
2424
+ },
2425
+ {
2426
+ "epoch": 5.02,
2427
+ "learning_rate": 7.659493670886077e-06,
2428
+ "loss": 0.0357,
2429
+ "step": 9750
2430
+ },
2431
+ {
2432
+ "epoch": 5.03,
2433
+ "learning_rate": 7.653164556962027e-06,
2434
+ "loss": 0.0256,
2435
+ "step": 9775
2436
+ },
2437
+ {
2438
+ "epoch": 5.03,
2439
+ "learning_rate": 7.646835443037976e-06,
2440
+ "loss": 0.0358,
2441
+ "step": 9800
2442
+ },
2443
+ {
2444
+ "epoch": 5.03,
2445
+ "learning_rate": 7.640506329113925e-06,
2446
+ "loss": 0.0361,
2447
+ "step": 9825
2448
+ },
2449
+ {
2450
+ "epoch": 5.03,
2451
+ "learning_rate": 7.634177215189875e-06,
2452
+ "loss": 0.0365,
2453
+ "step": 9850
2454
+ },
2455
+ {
2456
+ "epoch": 5.03,
2457
+ "learning_rate": 7.6278481012658234e-06,
2458
+ "loss": 0.0344,
2459
+ "step": 9875
2460
+ },
2461
+ {
2462
+ "epoch": 5.03,
2463
+ "learning_rate": 7.621518987341773e-06,
2464
+ "loss": 0.0349,
2465
+ "step": 9900
2466
+ },
2467
+ {
2468
+ "epoch": 5.03,
2469
+ "learning_rate": 7.615189873417722e-06,
2470
+ "loss": 0.0289,
2471
+ "step": 9925
2472
+ },
2473
+ {
2474
+ "epoch": 5.03,
2475
+ "learning_rate": 7.608860759493672e-06,
2476
+ "loss": 0.0259,
2477
+ "step": 9950
2478
+ },
2479
+ {
2480
+ "epoch": 5.03,
2481
+ "learning_rate": 7.602531645569621e-06,
2482
+ "loss": 0.0289,
2483
+ "step": 9975
2484
+ },
2485
+ {
2486
+ "epoch": 5.03,
2487
+ "learning_rate": 7.59620253164557e-06,
2488
+ "loss": 0.0269,
2489
+ "step": 10000
2490
+ },
2491
+ {
2492
+ "epoch": 5.03,
2493
+ "eval_loss": 0.2615405321121216,
2494
+ "eval_runtime": 1420.7197,
2495
+ "eval_samples_per_second": 7.328,
2496
+ "eval_steps_per_second": 0.458,
2497
+ "eval_wer": 36.23310948145498,
2498
+ "step": 10000
2499
+ },
2500
+ {
2501
+ "epoch": 5.03,
2502
+ "learning_rate": 7.589873417721519e-06,
2503
+ "loss": 0.0261,
2504
+ "step": 10025
2505
+ },
2506
+ {
2507
+ "epoch": 5.03,
2508
+ "learning_rate": 7.583544303797469e-06,
2509
+ "loss": 0.0278,
2510
+ "step": 10050
2511
+ },
2512
+ {
2513
+ "epoch": 5.03,
2514
+ "learning_rate": 7.577215189873418e-06,
2515
+ "loss": 0.0248,
2516
+ "step": 10075
2517
+ },
2518
+ {
2519
+ "epoch": 5.03,
2520
+ "learning_rate": 7.5708860759493674e-06,
2521
+ "loss": 0.0263,
2522
+ "step": 10100
2523
+ },
2524
+ {
2525
+ "epoch": 5.03,
2526
+ "learning_rate": 7.564556962025317e-06,
2527
+ "loss": 0.0253,
2528
+ "step": 10125
2529
+ },
2530
+ {
2531
+ "epoch": 5.03,
2532
+ "learning_rate": 7.558227848101266e-06,
2533
+ "loss": 0.0248,
2534
+ "step": 10150
2535
+ },
2536
+ {
2537
+ "epoch": 5.04,
2538
+ "learning_rate": 7.551898734177216e-06,
2539
+ "loss": 0.0281,
2540
+ "step": 10175
2541
+ },
2542
+ {
2543
+ "epoch": 5.04,
2544
+ "learning_rate": 7.545569620253165e-06,
2545
+ "loss": 0.0276,
2546
+ "step": 10200
2547
+ },
2548
+ {
2549
+ "epoch": 5.04,
2550
+ "learning_rate": 7.539240506329114e-06,
2551
+ "loss": 0.0285,
2552
+ "step": 10225
2553
+ },
2554
+ {
2555
+ "epoch": 5.04,
2556
+ "learning_rate": 7.532911392405063e-06,
2557
+ "loss": 0.0265,
2558
+ "step": 10250
2559
+ },
2560
+ {
2561
+ "epoch": 5.04,
2562
+ "learning_rate": 7.526582278481013e-06,
2563
+ "loss": 0.0288,
2564
+ "step": 10275
2565
+ },
2566
+ {
2567
+ "epoch": 5.04,
2568
+ "learning_rate": 7.520253164556963e-06,
2569
+ "loss": 0.0247,
2570
+ "step": 10300
2571
+ },
2572
+ {
2573
+ "epoch": 5.04,
2574
+ "learning_rate": 7.513924050632912e-06,
2575
+ "loss": 0.0281,
2576
+ "step": 10325
2577
+ },
2578
+ {
2579
+ "epoch": 5.04,
2580
+ "learning_rate": 7.507594936708862e-06,
2581
+ "loss": 0.0262,
2582
+ "step": 10350
2583
+ },
2584
+ {
2585
+ "epoch": 5.04,
2586
+ "learning_rate": 7.501265822784811e-06,
2587
+ "loss": 0.0243,
2588
+ "step": 10375
2589
+ },
2590
+ {
2591
+ "epoch": 5.04,
2592
+ "learning_rate": 7.494936708860761e-06,
2593
+ "loss": 0.0331,
2594
+ "step": 10400
2595
+ },
2596
+ {
2597
+ "epoch": 5.04,
2598
+ "learning_rate": 7.48860759493671e-06,
2599
+ "loss": 0.0346,
2600
+ "step": 10425
2601
+ },
2602
+ {
2603
+ "epoch": 5.04,
2604
+ "learning_rate": 7.4822784810126594e-06,
2605
+ "loss": 0.0348,
2606
+ "step": 10450
2607
+ },
2608
+ {
2609
+ "epoch": 5.04,
2610
+ "learning_rate": 7.475949367088609e-06,
2611
+ "loss": 0.0272,
2612
+ "step": 10475
2613
+ },
2614
+ {
2615
+ "epoch": 5.04,
2616
+ "learning_rate": 7.4696202531645574e-06,
2617
+ "loss": 0.0326,
2618
+ "step": 10500
2619
+ },
2620
+ {
2621
+ "epoch": 6.0,
2622
+ "learning_rate": 7.463291139240507e-06,
2623
+ "loss": 0.033,
2624
+ "step": 10525
2625
+ },
2626
+ {
2627
+ "epoch": 6.0,
2628
+ "learning_rate": 7.456962025316456e-06,
2629
+ "loss": 0.0333,
2630
+ "step": 10550
2631
+ },
2632
+ {
2633
+ "epoch": 6.0,
2634
+ "learning_rate": 7.450632911392406e-06,
2635
+ "loss": 0.0371,
2636
+ "step": 10575
2637
+ },
2638
+ {
2639
+ "epoch": 6.0,
2640
+ "learning_rate": 7.444303797468355e-06,
2641
+ "loss": 0.0359,
2642
+ "step": 10600
2643
+ },
2644
+ {
2645
+ "epoch": 6.0,
2646
+ "learning_rate": 7.437974683544305e-06,
2647
+ "loss": 0.0438,
2648
+ "step": 10625
2649
+ },
2650
+ {
2651
+ "epoch": 6.0,
2652
+ "learning_rate": 7.431645569620254e-06,
2653
+ "loss": 0.0382,
2654
+ "step": 10650
2655
+ },
2656
+ {
2657
+ "epoch": 6.0,
2658
+ "learning_rate": 7.4253164556962034e-06,
2659
+ "loss": 0.0337,
2660
+ "step": 10675
2661
+ },
2662
+ {
2663
+ "epoch": 6.0,
2664
+ "learning_rate": 7.418987341772153e-06,
2665
+ "loss": 0.0422,
2666
+ "step": 10700
2667
+ },
2668
+ {
2669
+ "epoch": 6.01,
2670
+ "learning_rate": 7.4126582278481014e-06,
2671
+ "loss": 0.0467,
2672
+ "step": 10725
2673
+ },
2674
+ {
2675
+ "epoch": 6.01,
2676
+ "learning_rate": 7.406329113924051e-06,
2677
+ "loss": 0.0437,
2678
+ "step": 10750
2679
+ },
2680
+ {
2681
+ "epoch": 6.01,
2682
+ "learning_rate": 7.4e-06,
2683
+ "loss": 0.0328,
2684
+ "step": 10775
2685
+ },
2686
+ {
2687
+ "epoch": 6.01,
2688
+ "learning_rate": 7.39367088607595e-06,
2689
+ "loss": 0.0295,
2690
+ "step": 10800
2691
+ },
2692
+ {
2693
+ "epoch": 6.01,
2694
+ "learning_rate": 7.387341772151899e-06,
2695
+ "loss": 0.0431,
2696
+ "step": 10825
2697
+ },
2698
+ {
2699
+ "epoch": 6.01,
2700
+ "learning_rate": 7.381012658227849e-06,
2701
+ "loss": 0.0456,
2702
+ "step": 10850
2703
+ },
2704
+ {
2705
+ "epoch": 6.01,
2706
+ "learning_rate": 7.374683544303798e-06,
2707
+ "loss": 0.0267,
2708
+ "step": 10875
2709
+ },
2710
+ {
2711
+ "epoch": 6.01,
2712
+ "learning_rate": 7.3683544303797474e-06,
2713
+ "loss": 0.0334,
2714
+ "step": 10900
2715
+ },
2716
+ {
2717
+ "epoch": 6.01,
2718
+ "learning_rate": 7.362025316455697e-06,
2719
+ "loss": 0.0296,
2720
+ "step": 10925
2721
+ },
2722
+ {
2723
+ "epoch": 6.01,
2724
+ "learning_rate": 7.3556962025316454e-06,
2725
+ "loss": 0.0261,
2726
+ "step": 10950
2727
+ },
2728
+ {
2729
+ "epoch": 6.01,
2730
+ "learning_rate": 7.349367088607595e-06,
2731
+ "loss": 0.0299,
2732
+ "step": 10975
2733
+ },
2734
+ {
2735
+ "epoch": 6.01,
2736
+ "learning_rate": 7.343037974683544e-06,
2737
+ "loss": 0.0379,
2738
+ "step": 11000
2739
+ },
2740
+ {
2741
+ "epoch": 6.01,
2742
+ "eval_loss": 0.23971551656723022,
2743
+ "eval_runtime": 1435.6772,
2744
+ "eval_samples_per_second": 7.252,
2745
+ "eval_steps_per_second": 0.453,
2746
+ "eval_wer": 31.7712483129807,
2747
+ "step": 11000
2748
+ }
2749
+ ],
2750
+ "max_steps": 40000,
2751
+ "num_train_epochs": 9223372036854775807,
2752
+ "total_flos": 5.07737153323008e+19,
2753
+ "trial_name": null,
2754
+ "trial_params": null
2755
+ }
checkpoint-11000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0289e1bf8fd1a33f7c3c9d3e15cc9bdd13d10e9774e0ada934c0b9d71501a993
3
+ size 4091
checkpoint-12000/config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-small",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": true,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 768,
17
+ "decoder_attention_heads": 12,
18
+ "decoder_ffn_dim": 3072,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 12,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 12,
24
+ "encoder_ffn_dim": 3072,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 12,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": null,
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "mask_feature_length": 10,
32
+ "mask_feature_min_masks": 0,
33
+ "mask_feature_prob": 0.05,
34
+ "mask_time_length": 10,
35
+ "mask_time_min_masks": 2,
36
+ "mask_time_prob": 0.05,
37
+ "max_length": 448,
38
+ "max_source_positions": 1500,
39
+ "max_target_positions": 448,
40
+ "model_type": "whisper",
41
+ "num_hidden_layers": 12,
42
+ "num_mel_bins": 80,
43
+ "pad_token_id": 50257,
44
+ "scale_embedding": false,
45
+ "suppress_tokens": [],
46
+ "torch_dtype": "float32",
47
+ "transformers_version": "4.30.0.dev0",
48
+ "use_cache": true,
49
+ "use_weighted_layer_sum": false,
50
+ "vocab_size": 51865
51
+ }
checkpoint-12000/generation_config.json ADDED
@@ -0,0 +1,221 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "begin_suppress_tokens": [
3
+ 220,
4
+ 50257
5
+ ],
6
+ "bos_token_id": 50257,
7
+ "decoder_start_token_id": 50258,
8
+ "eos_token_id": 50257,
9
+ "forced_decoder_ids": [
10
+ [
11
+ 1,
12
+ null
13
+ ],
14
+ [
15
+ 2,
16
+ 50359
17
+ ]
18
+ ],
19
+ "is_multilingual": true,
20
+ "lang_to_id": {
21
+ "<|af|>": 50327,
22
+ "<|am|>": 50334,
23
+ "<|ar|>": 50272,
24
+ "<|as|>": 50350,
25
+ "<|az|>": 50304,
26
+ "<|ba|>": 50355,
27
+ "<|be|>": 50330,
28
+ "<|bg|>": 50292,
29
+ "<|bn|>": 50302,
30
+ "<|bo|>": 50347,
31
+ "<|br|>": 50309,
32
+ "<|bs|>": 50315,
33
+ "<|ca|>": 50270,
34
+ "<|cs|>": 50283,
35
+ "<|cy|>": 50297,
36
+ "<|da|>": 50285,
37
+ "<|de|>": 50261,
38
+ "<|el|>": 50281,
39
+ "<|en|>": 50259,
40
+ "<|es|>": 50262,
41
+ "<|et|>": 50307,
42
+ "<|eu|>": 50310,
43
+ "<|fa|>": 50300,
44
+ "<|fi|>": 50277,
45
+ "<|fo|>": 50338,
46
+ "<|fr|>": 50265,
47
+ "<|gl|>": 50319,
48
+ "<|gu|>": 50333,
49
+ "<|haw|>": 50352,
50
+ "<|ha|>": 50354,
51
+ "<|he|>": 50279,
52
+ "<|hi|>": 50276,
53
+ "<|hr|>": 50291,
54
+ "<|ht|>": 50339,
55
+ "<|hu|>": 50286,
56
+ "<|hy|>": 50312,
57
+ "<|id|>": 50275,
58
+ "<|is|>": 50311,
59
+ "<|it|>": 50274,
60
+ "<|ja|>": 50266,
61
+ "<|jw|>": 50356,
62
+ "<|ka|>": 50329,
63
+ "<|kk|>": 50316,
64
+ "<|km|>": 50323,
65
+ "<|kn|>": 50306,
66
+ "<|ko|>": 50264,
67
+ "<|la|>": 50294,
68
+ "<|lb|>": 50345,
69
+ "<|ln|>": 50353,
70
+ "<|lo|>": 50336,
71
+ "<|lt|>": 50293,
72
+ "<|lv|>": 50301,
73
+ "<|mg|>": 50349,
74
+ "<|mi|>": 50295,
75
+ "<|mk|>": 50308,
76
+ "<|ml|>": 50296,
77
+ "<|mn|>": 50314,
78
+ "<|mr|>": 50320,
79
+ "<|ms|>": 50282,
80
+ "<|mt|>": 50343,
81
+ "<|my|>": 50346,
82
+ "<|ne|>": 50313,
83
+ "<|nl|>": 50271,
84
+ "<|nn|>": 50342,
85
+ "<|no|>": 50288,
86
+ "<|oc|>": 50328,
87
+ "<|pa|>": 50321,
88
+ "<|pl|>": 50269,
89
+ "<|ps|>": 50340,
90
+ "<|pt|>": 50267,
91
+ "<|ro|>": 50284,
92
+ "<|ru|>": 50263,
93
+ "<|sa|>": 50344,
94
+ "<|sd|>": 50332,
95
+ "<|si|>": 50322,
96
+ "<|sk|>": 50298,
97
+ "<|sl|>": 50305,
98
+ "<|sn|>": 50324,
99
+ "<|so|>": 50326,
100
+ "<|sq|>": 50317,
101
+ "<|sr|>": 50303,
102
+ "<|su|>": 50357,
103
+ "<|sv|>": 50273,
104
+ "<|sw|>": 50318,
105
+ "<|ta|>": 50287,
106
+ "<|te|>": 50299,
107
+ "<|tg|>": 50331,
108
+ "<|th|>": 50289,
109
+ "<|tk|>": 50341,
110
+ "<|tl|>": 50348,
111
+ "<|tr|>": 50268,
112
+ "<|tt|>": 50351,
113
+ "<|uk|>": 50280,
114
+ "<|ur|>": 50290,
115
+ "<|uz|>": 50337,
116
+ "<|vi|>": 50278,
117
+ "<|yi|>": 50335,
118
+ "<|yo|>": 50325,
119
+ "<|zh|>": 50260
120
+ },
121
+ "max_initial_timestamp_index": 1,
122
+ "max_length": 448,
123
+ "no_timestamps_token_id": 50363,
124
+ "pad_token_id": 50257,
125
+ "return_timestamps": false,
126
+ "suppress_tokens": [
127
+ 1,
128
+ 2,
129
+ 7,
130
+ 8,
131
+ 9,
132
+ 10,
133
+ 14,
134
+ 25,
135
+ 26,
136
+ 27,
137
+ 28,
138
+ 29,
139
+ 31,
140
+ 58,
141
+ 59,
142
+ 60,
143
+ 61,
144
+ 62,
145
+ 63,
146
+ 90,
147
+ 91,
148
+ 92,
149
+ 93,
150
+ 359,
151
+ 503,
152
+ 522,
153
+ 542,
154
+ 873,
155
+ 893,
156
+ 902,
157
+ 918,
158
+ 922,
159
+ 931,
160
+ 1350,
161
+ 1853,
162
+ 1982,
163
+ 2460,
164
+ 2627,
165
+ 3246,
166
+ 3253,
167
+ 3268,
168
+ 3536,
169
+ 3846,
170
+ 3961,
171
+ 4183,
172
+ 4667,
173
+ 6585,
174
+ 6647,
175
+ 7273,
176
+ 9061,
177
+ 9383,
178
+ 10428,
179
+ 10929,
180
+ 11938,
181
+ 12033,
182
+ 12331,
183
+ 12562,
184
+ 13793,
185
+ 14157,
186
+ 14635,
187
+ 15265,
188
+ 15618,
189
+ 16553,
190
+ 16604,
191
+ 18362,
192
+ 18956,
193
+ 20075,
194
+ 21675,
195
+ 22520,
196
+ 26130,
197
+ 26161,
198
+ 26435,
199
+ 28279,
200
+ 29464,
201
+ 31650,
202
+ 32302,
203
+ 32470,
204
+ 36865,
205
+ 42863,
206
+ 47425,
207
+ 49870,
208
+ 50254,
209
+ 50258,
210
+ 50358,
211
+ 50359,
212
+ 50360,
213
+ 50361,
214
+ 50362
215
+ ],
216
+ "task_to_id": {
217
+ "transcribe": 50359,
218
+ "translate": 50358
219
+ },
220
+ "transformers_version": "4.30.0.dev0"
221
+ }
checkpoint-12000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:92e664ec5154e726a35e6d42bf3352ab51341193a6689efdaea9604923bda9cc
3
+ size 1934161093
checkpoint-12000/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-12000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:163238221ef8cc3002eb5d1ae169577ad908ddae8b499b8370880e5e726f9ef6
3
+ size 967102601
checkpoint-12000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8c552ce94d91a466160e0526f23f77fc2c5a8568c307954a912646bc8e340ae
3
+ size 14639
checkpoint-12000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b5a815a4758f309647094935e93cd183cdf97272d965f1f809d37a65b8075906
3
+ size 557
checkpoint-12000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e340fa9311f70ee0dc384eeb43490c95a739982fb89fc03044dd1c37ffa4ee2
3
+ size 627
checkpoint-12000/trainer_state.json ADDED
@@ -0,0 +1,3004 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 31.58262735979447,
3
+ "best_model_checkpoint": "./whisper-small-fa-aug/checkpoint-9000",
4
+ "epoch": 6.03735,
5
+ "global_step": 12000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 4.2000000000000006e-07,
13
+ "loss": 3.8,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 9.200000000000001e-07,
19
+ "loss": 3.0497,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.0,
24
+ "learning_rate": 1.42e-06,
25
+ "loss": 2.0102,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.0,
30
+ "learning_rate": 1.9200000000000003e-06,
31
+ "loss": 1.5948,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.0,
36
+ "learning_rate": 2.42e-06,
37
+ "loss": 1.4094,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.0,
42
+ "learning_rate": 2.92e-06,
43
+ "loss": 1.2611,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.0,
48
+ "learning_rate": 3.4200000000000007e-06,
49
+ "loss": 1.1366,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.01,
54
+ "learning_rate": 3.920000000000001e-06,
55
+ "loss": 1.0817,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.01,
60
+ "learning_rate": 4.42e-06,
61
+ "loss": 1.0324,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.01,
66
+ "learning_rate": 4.92e-06,
67
+ "loss": 0.9176,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.01,
72
+ "learning_rate": 5.420000000000001e-06,
73
+ "loss": 0.7975,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.01,
78
+ "learning_rate": 5.92e-06,
79
+ "loss": 0.7079,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.01,
84
+ "learning_rate": 6.42e-06,
85
+ "loss": 0.5832,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.01,
90
+ "learning_rate": 6.92e-06,
91
+ "loss": 0.4469,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.01,
96
+ "learning_rate": 7.420000000000001e-06,
97
+ "loss": 0.3913,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.01,
102
+ "learning_rate": 7.92e-06,
103
+ "loss": 0.4189,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.01,
108
+ "learning_rate": 8.42e-06,
109
+ "loss": 0.3756,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.01,
114
+ "learning_rate": 8.920000000000001e-06,
115
+ "loss": 0.3876,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.01,
120
+ "learning_rate": 9.42e-06,
121
+ "loss": 0.3802,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.01,
126
+ "learning_rate": 9.920000000000002e-06,
127
+ "loss": 0.4061,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.01,
132
+ "learning_rate": 9.994683544303798e-06,
133
+ "loss": 0.3885,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.01,
138
+ "learning_rate": 9.988354430379748e-06,
139
+ "loss": 0.3449,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.01,
144
+ "learning_rate": 9.982025316455697e-06,
145
+ "loss": 0.3292,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.01,
150
+ "learning_rate": 9.975696202531647e-06,
151
+ "loss": 0.3439,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.02,
156
+ "learning_rate": 9.969367088607596e-06,
157
+ "loss": 0.3031,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.02,
162
+ "learning_rate": 9.963037974683545e-06,
163
+ "loss": 0.3159,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.02,
168
+ "learning_rate": 9.956708860759495e-06,
169
+ "loss": 0.3304,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.02,
174
+ "learning_rate": 9.950379746835444e-06,
175
+ "loss": 0.2918,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.02,
180
+ "learning_rate": 9.944050632911392e-06,
181
+ "loss": 0.2785,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.02,
186
+ "learning_rate": 9.937721518987341e-06,
187
+ "loss": 0.2618,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.02,
192
+ "learning_rate": 9.931392405063291e-06,
193
+ "loss": 0.2862,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.02,
198
+ "learning_rate": 9.92506329113924e-06,
199
+ "loss": 0.2962,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.02,
204
+ "learning_rate": 9.91873417721519e-06,
205
+ "loss": 0.2935,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.02,
210
+ "learning_rate": 9.912405063291141e-06,
211
+ "loss": 0.2827,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.02,
216
+ "learning_rate": 9.90607594936709e-06,
217
+ "loss": 0.2702,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 0.02,
222
+ "learning_rate": 9.89974683544304e-06,
223
+ "loss": 0.2987,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 0.02,
228
+ "learning_rate": 9.89341772151899e-06,
229
+ "loss": 0.3067,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 0.02,
234
+ "learning_rate": 9.887088607594939e-06,
235
+ "loss": 0.2962,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 0.02,
240
+ "learning_rate": 9.880759493670888e-06,
241
+ "loss": 0.2807,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 0.03,
246
+ "learning_rate": 9.874430379746836e-06,
247
+ "loss": 0.2533,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 0.03,
252
+ "eval_loss": 0.3274388611316681,
253
+ "eval_runtime": 1481.9106,
254
+ "eval_samples_per_second": 7.025,
255
+ "eval_steps_per_second": 0.439,
256
+ "eval_wer": 40.657571667832,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 0.03,
261
+ "learning_rate": 9.868101265822785e-06,
262
+ "loss": 0.21,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 0.03,
267
+ "learning_rate": 9.861772151898735e-06,
268
+ "loss": 0.2463,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 0.03,
273
+ "learning_rate": 9.855443037974684e-06,
274
+ "loss": 0.2688,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 0.03,
279
+ "learning_rate": 9.849113924050633e-06,
280
+ "loss": 0.2293,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 0.03,
285
+ "learning_rate": 9.842784810126583e-06,
286
+ "loss": 0.2283,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 0.03,
291
+ "learning_rate": 9.836455696202532e-06,
292
+ "loss": 0.228,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 0.03,
297
+ "learning_rate": 9.830126582278482e-06,
298
+ "loss": 0.1919,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 0.03,
303
+ "learning_rate": 9.823797468354431e-06,
304
+ "loss": 0.2173,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 0.03,
309
+ "learning_rate": 9.81746835443038e-06,
310
+ "loss": 0.2066,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 0.03,
315
+ "learning_rate": 9.81113924050633e-06,
316
+ "loss": 0.1826,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 0.03,
321
+ "learning_rate": 9.80481012658228e-06,
322
+ "loss": 0.1855,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 0.03,
327
+ "learning_rate": 9.798481012658229e-06,
328
+ "loss": 0.1702,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 0.03,
333
+ "learning_rate": 9.792151898734178e-06,
334
+ "loss": 0.1714,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 0.03,
339
+ "learning_rate": 9.785822784810128e-06,
340
+ "loss": 0.1686,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 0.03,
345
+ "learning_rate": 9.779493670886077e-06,
346
+ "loss": 0.1762,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 0.04,
351
+ "learning_rate": 9.773164556962027e-06,
352
+ "loss": 0.1722,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 0.04,
357
+ "learning_rate": 9.766835443037976e-06,
358
+ "loss": 0.2051,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 0.04,
363
+ "learning_rate": 9.760506329113924e-06,
364
+ "loss": 0.2071,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 0.04,
369
+ "learning_rate": 9.754177215189873e-06,
370
+ "loss": 0.1664,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 0.04,
375
+ "learning_rate": 9.747848101265823e-06,
376
+ "loss": 0.1871,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 0.04,
381
+ "learning_rate": 9.741518987341772e-06,
382
+ "loss": 0.1893,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 0.04,
387
+ "learning_rate": 9.735189873417721e-06,
388
+ "loss": 0.1756,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 0.04,
393
+ "learning_rate": 9.728860759493671e-06,
394
+ "loss": 0.1687,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 0.04,
399
+ "learning_rate": 9.72253164556962e-06,
400
+ "loss": 0.1685,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 0.04,
405
+ "learning_rate": 9.71620253164557e-06,
406
+ "loss": 0.1805,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 0.04,
411
+ "learning_rate": 9.70987341772152e-06,
412
+ "loss": 0.1868,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 0.04,
417
+ "learning_rate": 9.70354430379747e-06,
418
+ "loss": 0.212,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 0.04,
423
+ "learning_rate": 9.69721518987342e-06,
424
+ "loss": 0.2019,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 0.04,
429
+ "learning_rate": 9.69088607594937e-06,
430
+ "loss": 0.2076,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 0.04,
435
+ "learning_rate": 9.684556962025317e-06,
436
+ "loss": 0.1971,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 1.0,
441
+ "learning_rate": 9.678227848101266e-06,
442
+ "loss": 0.2424,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 1.0,
447
+ "learning_rate": 9.671898734177216e-06,
448
+ "loss": 0.2346,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 1.0,
453
+ "learning_rate": 9.665569620253165e-06,
454
+ "loss": 0.2436,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 1.0,
459
+ "learning_rate": 9.659240506329115e-06,
460
+ "loss": 0.2613,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 1.0,
465
+ "learning_rate": 9.652911392405064e-06,
466
+ "loss": 0.2781,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 1.0,
471
+ "learning_rate": 9.646582278481013e-06,
472
+ "loss": 0.2465,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 1.0,
477
+ "learning_rate": 9.640253164556963e-06,
478
+ "loss": 0.2386,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 1.0,
483
+ "learning_rate": 9.633924050632912e-06,
484
+ "loss": 0.2298,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 1.01,
489
+ "learning_rate": 9.627594936708862e-06,
490
+ "loss": 0.264,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 1.01,
495
+ "learning_rate": 9.621265822784811e-06,
496
+ "loss": 0.2077,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 1.01,
501
+ "eval_loss": 0.25397753715515137,
502
+ "eval_runtime": 1477.9319,
503
+ "eval_samples_per_second": 7.044,
504
+ "eval_steps_per_second": 0.44,
505
+ "eval_wer": 35.12252231743606,
506
+ "step": 2000
507
+ },
508
+ {
509
+ "epoch": 1.01,
510
+ "learning_rate": 9.61493670886076e-06,
511
+ "loss": 0.192,
512
+ "step": 2025
513
+ },
514
+ {
515
+ "epoch": 1.01,
516
+ "learning_rate": 9.60860759493671e-06,
517
+ "loss": 0.1903,
518
+ "step": 2050
519
+ },
520
+ {
521
+ "epoch": 1.01,
522
+ "learning_rate": 9.60227848101266e-06,
523
+ "loss": 0.2135,
524
+ "step": 2075
525
+ },
526
+ {
527
+ "epoch": 1.01,
528
+ "learning_rate": 9.595949367088609e-06,
529
+ "loss": 0.1917,
530
+ "step": 2100
531
+ },
532
+ {
533
+ "epoch": 1.01,
534
+ "learning_rate": 9.589620253164558e-06,
535
+ "loss": 0.1579,
536
+ "step": 2125
537
+ },
538
+ {
539
+ "epoch": 1.01,
540
+ "learning_rate": 9.583291139240508e-06,
541
+ "loss": 0.1779,
542
+ "step": 2150
543
+ },
544
+ {
545
+ "epoch": 1.01,
546
+ "learning_rate": 9.576962025316457e-06,
547
+ "loss": 0.1504,
548
+ "step": 2175
549
+ },
550
+ {
551
+ "epoch": 1.01,
552
+ "learning_rate": 9.570632911392405e-06,
553
+ "loss": 0.1574,
554
+ "step": 2200
555
+ },
556
+ {
557
+ "epoch": 1.01,
558
+ "learning_rate": 9.564303797468354e-06,
559
+ "loss": 0.1757,
560
+ "step": 2225
561
+ },
562
+ {
563
+ "epoch": 1.01,
564
+ "learning_rate": 9.557974683544304e-06,
565
+ "loss": 0.1996,
566
+ "step": 2250
567
+ },
568
+ {
569
+ "epoch": 1.01,
570
+ "learning_rate": 9.551645569620253e-06,
571
+ "loss": 0.1716,
572
+ "step": 2275
573
+ },
574
+ {
575
+ "epoch": 1.01,
576
+ "learning_rate": 9.545316455696203e-06,
577
+ "loss": 0.1509,
578
+ "step": 2300
579
+ },
580
+ {
581
+ "epoch": 1.01,
582
+ "learning_rate": 9.538987341772152e-06,
583
+ "loss": 0.1402,
584
+ "step": 2325
585
+ },
586
+ {
587
+ "epoch": 1.01,
588
+ "learning_rate": 9.532658227848101e-06,
589
+ "loss": 0.1612,
590
+ "step": 2350
591
+ },
592
+ {
593
+ "epoch": 1.02,
594
+ "learning_rate": 9.526329113924051e-06,
595
+ "loss": 0.1357,
596
+ "step": 2375
597
+ },
598
+ {
599
+ "epoch": 1.02,
600
+ "learning_rate": 9.52e-06,
601
+ "loss": 0.1426,
602
+ "step": 2400
603
+ },
604
+ {
605
+ "epoch": 1.02,
606
+ "learning_rate": 9.51367088607595e-06,
607
+ "loss": 0.1644,
608
+ "step": 2425
609
+ },
610
+ {
611
+ "epoch": 1.02,
612
+ "learning_rate": 9.5073417721519e-06,
613
+ "loss": 0.1196,
614
+ "step": 2450
615
+ },
616
+ {
617
+ "epoch": 1.02,
618
+ "learning_rate": 9.501012658227849e-06,
619
+ "loss": 0.1332,
620
+ "step": 2475
621
+ },
622
+ {
623
+ "epoch": 1.02,
624
+ "learning_rate": 9.494683544303798e-06,
625
+ "loss": 0.1096,
626
+ "step": 2500
627
+ },
628
+ {
629
+ "epoch": 1.02,
630
+ "learning_rate": 9.488354430379747e-06,
631
+ "loss": 0.1263,
632
+ "step": 2525
633
+ },
634
+ {
635
+ "epoch": 1.02,
636
+ "learning_rate": 9.482025316455697e-06,
637
+ "loss": 0.1467,
638
+ "step": 2550
639
+ },
640
+ {
641
+ "epoch": 1.02,
642
+ "learning_rate": 9.475696202531646e-06,
643
+ "loss": 0.1372,
644
+ "step": 2575
645
+ },
646
+ {
647
+ "epoch": 1.02,
648
+ "learning_rate": 9.469367088607596e-06,
649
+ "loss": 0.1256,
650
+ "step": 2600
651
+ },
652
+ {
653
+ "epoch": 1.02,
654
+ "learning_rate": 9.463037974683545e-06,
655
+ "loss": 0.1319,
656
+ "step": 2625
657
+ },
658
+ {
659
+ "epoch": 1.02,
660
+ "learning_rate": 9.456708860759495e-06,
661
+ "loss": 0.1385,
662
+ "step": 2650
663
+ },
664
+ {
665
+ "epoch": 1.02,
666
+ "learning_rate": 9.450379746835444e-06,
667
+ "loss": 0.147,
668
+ "step": 2675
669
+ },
670
+ {
671
+ "epoch": 1.02,
672
+ "learning_rate": 9.444050632911393e-06,
673
+ "loss": 0.1527,
674
+ "step": 2700
675
+ },
676
+ {
677
+ "epoch": 1.02,
678
+ "learning_rate": 9.437721518987343e-06,
679
+ "loss": 0.1357,
680
+ "step": 2725
681
+ },
682
+ {
683
+ "epoch": 1.02,
684
+ "learning_rate": 9.431392405063292e-06,
685
+ "loss": 0.122,
686
+ "step": 2750
687
+ },
688
+ {
689
+ "epoch": 1.03,
690
+ "learning_rate": 9.425063291139242e-06,
691
+ "loss": 0.0918,
692
+ "step": 2775
693
+ },
694
+ {
695
+ "epoch": 1.03,
696
+ "learning_rate": 9.418734177215191e-06,
697
+ "loss": 0.1218,
698
+ "step": 2800
699
+ },
700
+ {
701
+ "epoch": 1.03,
702
+ "learning_rate": 9.41240506329114e-06,
703
+ "loss": 0.139,
704
+ "step": 2825
705
+ },
706
+ {
707
+ "epoch": 1.03,
708
+ "learning_rate": 9.40607594936709e-06,
709
+ "loss": 0.1205,
710
+ "step": 2850
711
+ },
712
+ {
713
+ "epoch": 1.03,
714
+ "learning_rate": 9.39974683544304e-06,
715
+ "loss": 0.1158,
716
+ "step": 2875
717
+ },
718
+ {
719
+ "epoch": 1.03,
720
+ "learning_rate": 9.393417721518989e-06,
721
+ "loss": 0.1156,
722
+ "step": 2900
723
+ },
724
+ {
725
+ "epoch": 1.03,
726
+ "learning_rate": 9.387088607594937e-06,
727
+ "loss": 0.1044,
728
+ "step": 2925
729
+ },
730
+ {
731
+ "epoch": 1.03,
732
+ "learning_rate": 9.380759493670886e-06,
733
+ "loss": 0.112,
734
+ "step": 2950
735
+ },
736
+ {
737
+ "epoch": 1.03,
738
+ "learning_rate": 9.374430379746835e-06,
739
+ "loss": 0.1023,
740
+ "step": 2975
741
+ },
742
+ {
743
+ "epoch": 1.03,
744
+ "learning_rate": 9.368101265822785e-06,
745
+ "loss": 0.0966,
746
+ "step": 3000
747
+ },
748
+ {
749
+ "epoch": 1.03,
750
+ "eval_loss": 0.2638459801673889,
751
+ "eval_runtime": 1500.936,
752
+ "eval_samples_per_second": 6.936,
753
+ "eval_steps_per_second": 0.434,
754
+ "eval_wer": 37.93069806013106,
755
+ "step": 3000
756
+ },
757
+ {
758
+ "epoch": 1.03,
759
+ "learning_rate": 9.361772151898734e-06,
760
+ "loss": 0.0942,
761
+ "step": 3025
762
+ },
763
+ {
764
+ "epoch": 1.03,
765
+ "learning_rate": 9.355443037974684e-06,
766
+ "loss": 0.0826,
767
+ "step": 3050
768
+ },
769
+ {
770
+ "epoch": 1.03,
771
+ "learning_rate": 9.349113924050633e-06,
772
+ "loss": 0.086,
773
+ "step": 3075
774
+ },
775
+ {
776
+ "epoch": 1.03,
777
+ "learning_rate": 9.342784810126583e-06,
778
+ "loss": 0.0956,
779
+ "step": 3100
780
+ },
781
+ {
782
+ "epoch": 1.03,
783
+ "learning_rate": 9.336455696202532e-06,
784
+ "loss": 0.0935,
785
+ "step": 3125
786
+ },
787
+ {
788
+ "epoch": 1.03,
789
+ "learning_rate": 9.330126582278481e-06,
790
+ "loss": 0.0923,
791
+ "step": 3150
792
+ },
793
+ {
794
+ "epoch": 1.04,
795
+ "learning_rate": 9.323797468354431e-06,
796
+ "loss": 0.1163,
797
+ "step": 3175
798
+ },
799
+ {
800
+ "epoch": 1.04,
801
+ "learning_rate": 9.31746835443038e-06,
802
+ "loss": 0.1099,
803
+ "step": 3200
804
+ },
805
+ {
806
+ "epoch": 1.04,
807
+ "learning_rate": 9.31113924050633e-06,
808
+ "loss": 0.0879,
809
+ "step": 3225
810
+ },
811
+ {
812
+ "epoch": 1.04,
813
+ "learning_rate": 9.30481012658228e-06,
814
+ "loss": 0.0994,
815
+ "step": 3250
816
+ },
817
+ {
818
+ "epoch": 1.04,
819
+ "learning_rate": 9.298481012658229e-06,
820
+ "loss": 0.1069,
821
+ "step": 3275
822
+ },
823
+ {
824
+ "epoch": 1.04,
825
+ "learning_rate": 9.292151898734178e-06,
826
+ "loss": 0.0957,
827
+ "step": 3300
828
+ },
829
+ {
830
+ "epoch": 1.04,
831
+ "learning_rate": 9.285822784810127e-06,
832
+ "loss": 0.0866,
833
+ "step": 3325
834
+ },
835
+ {
836
+ "epoch": 1.04,
837
+ "learning_rate": 9.279493670886077e-06,
838
+ "loss": 0.0965,
839
+ "step": 3350
840
+ },
841
+ {
842
+ "epoch": 1.04,
843
+ "learning_rate": 9.273164556962026e-06,
844
+ "loss": 0.0952,
845
+ "step": 3375
846
+ },
847
+ {
848
+ "epoch": 1.04,
849
+ "learning_rate": 9.266835443037976e-06,
850
+ "loss": 0.1178,
851
+ "step": 3400
852
+ },
853
+ {
854
+ "epoch": 1.04,
855
+ "learning_rate": 9.260506329113925e-06,
856
+ "loss": 0.1154,
857
+ "step": 3425
858
+ },
859
+ {
860
+ "epoch": 1.04,
861
+ "learning_rate": 9.254177215189875e-06,
862
+ "loss": 0.1075,
863
+ "step": 3450
864
+ },
865
+ {
866
+ "epoch": 1.04,
867
+ "learning_rate": 9.247848101265824e-06,
868
+ "loss": 0.1161,
869
+ "step": 3475
870
+ },
871
+ {
872
+ "epoch": 1.04,
873
+ "learning_rate": 9.241518987341773e-06,
874
+ "loss": 0.1153,
875
+ "step": 3500
876
+ },
877
+ {
878
+ "epoch": 2.0,
879
+ "learning_rate": 9.235189873417723e-06,
880
+ "loss": 0.1343,
881
+ "step": 3525
882
+ },
883
+ {
884
+ "epoch": 2.0,
885
+ "learning_rate": 9.228860759493672e-06,
886
+ "loss": 0.1374,
887
+ "step": 3550
888
+ },
889
+ {
890
+ "epoch": 2.0,
891
+ "learning_rate": 9.222531645569622e-06,
892
+ "loss": 0.1436,
893
+ "step": 3575
894
+ },
895
+ {
896
+ "epoch": 2.0,
897
+ "learning_rate": 9.216202531645571e-06,
898
+ "loss": 0.1561,
899
+ "step": 3600
900
+ },
901
+ {
902
+ "epoch": 2.0,
903
+ "learning_rate": 9.20987341772152e-06,
904
+ "loss": 0.1709,
905
+ "step": 3625
906
+ },
907
+ {
908
+ "epoch": 2.0,
909
+ "learning_rate": 9.20354430379747e-06,
910
+ "loss": 0.142,
911
+ "step": 3650
912
+ },
913
+ {
914
+ "epoch": 2.0,
915
+ "learning_rate": 9.197215189873418e-06,
916
+ "loss": 0.1389,
917
+ "step": 3675
918
+ },
919
+ {
920
+ "epoch": 2.0,
921
+ "learning_rate": 9.190886075949367e-06,
922
+ "loss": 0.1398,
923
+ "step": 3700
924
+ },
925
+ {
926
+ "epoch": 2.01,
927
+ "learning_rate": 9.184556962025317e-06,
928
+ "loss": 0.1526,
929
+ "step": 3725
930
+ },
931
+ {
932
+ "epoch": 2.01,
933
+ "learning_rate": 9.178227848101266e-06,
934
+ "loss": 0.1273,
935
+ "step": 3750
936
+ },
937
+ {
938
+ "epoch": 2.01,
939
+ "learning_rate": 9.171898734177215e-06,
940
+ "loss": 0.1119,
941
+ "step": 3775
942
+ },
943
+ {
944
+ "epoch": 2.01,
945
+ "learning_rate": 9.165569620253165e-06,
946
+ "loss": 0.1088,
947
+ "step": 3800
948
+ },
949
+ {
950
+ "epoch": 2.01,
951
+ "learning_rate": 9.159493670886076e-06,
952
+ "loss": 0.1362,
953
+ "step": 3825
954
+ },
955
+ {
956
+ "epoch": 2.01,
957
+ "learning_rate": 9.153164556962026e-06,
958
+ "loss": 0.1247,
959
+ "step": 3850
960
+ },
961
+ {
962
+ "epoch": 2.01,
963
+ "learning_rate": 9.146835443037975e-06,
964
+ "loss": 0.0974,
965
+ "step": 3875
966
+ },
967
+ {
968
+ "epoch": 2.01,
969
+ "learning_rate": 9.140506329113925e-06,
970
+ "loss": 0.1218,
971
+ "step": 3900
972
+ },
973
+ {
974
+ "epoch": 2.01,
975
+ "learning_rate": 9.134177215189874e-06,
976
+ "loss": 0.0846,
977
+ "step": 3925
978
+ },
979
+ {
980
+ "epoch": 2.01,
981
+ "learning_rate": 9.127848101265823e-06,
982
+ "loss": 0.0964,
983
+ "step": 3950
984
+ },
985
+ {
986
+ "epoch": 2.01,
987
+ "learning_rate": 9.121518987341773e-06,
988
+ "loss": 0.1166,
989
+ "step": 3975
990
+ },
991
+ {
992
+ "epoch": 2.01,
993
+ "learning_rate": 9.115189873417722e-06,
994
+ "loss": 0.1212,
995
+ "step": 4000
996
+ },
997
+ {
998
+ "epoch": 2.01,
999
+ "eval_loss": 0.22517943382263184,
1000
+ "eval_runtime": 1537.556,
1001
+ "eval_samples_per_second": 6.771,
1002
+ "eval_steps_per_second": 0.423,
1003
+ "eval_wer": 69.04990325045935,
1004
+ "step": 4000
1005
+ },
1006
+ {
1007
+ "epoch": 2.01,
1008
+ "learning_rate": 9.108860759493672e-06,
1009
+ "loss": 0.1076,
1010
+ "step": 4025
1011
+ },
1012
+ {
1013
+ "epoch": 2.01,
1014
+ "learning_rate": 9.102531645569621e-06,
1015
+ "loss": 0.086,
1016
+ "step": 4050
1017
+ },
1018
+ {
1019
+ "epoch": 2.01,
1020
+ "learning_rate": 9.09620253164557e-06,
1021
+ "loss": 0.0773,
1022
+ "step": 4075
1023
+ },
1024
+ {
1025
+ "epoch": 2.01,
1026
+ "learning_rate": 9.08987341772152e-06,
1027
+ "loss": 0.098,
1028
+ "step": 4100
1029
+ },
1030
+ {
1031
+ "epoch": 2.02,
1032
+ "learning_rate": 9.08354430379747e-06,
1033
+ "loss": 0.0747,
1034
+ "step": 4125
1035
+ },
1036
+ {
1037
+ "epoch": 2.02,
1038
+ "learning_rate": 9.077215189873419e-06,
1039
+ "loss": 0.0895,
1040
+ "step": 4150
1041
+ },
1042
+ {
1043
+ "epoch": 2.02,
1044
+ "learning_rate": 9.070886075949368e-06,
1045
+ "loss": 0.0957,
1046
+ "step": 4175
1047
+ },
1048
+ {
1049
+ "epoch": 2.02,
1050
+ "learning_rate": 9.064556962025318e-06,
1051
+ "loss": 0.0789,
1052
+ "step": 4200
1053
+ },
1054
+ {
1055
+ "epoch": 2.02,
1056
+ "learning_rate": 9.058227848101267e-06,
1057
+ "loss": 0.0718,
1058
+ "step": 4225
1059
+ },
1060
+ {
1061
+ "epoch": 2.02,
1062
+ "learning_rate": 9.051898734177215e-06,
1063
+ "loss": 0.0737,
1064
+ "step": 4250
1065
+ },
1066
+ {
1067
+ "epoch": 2.02,
1068
+ "learning_rate": 9.045569620253164e-06,
1069
+ "loss": 0.0731,
1070
+ "step": 4275
1071
+ },
1072
+ {
1073
+ "epoch": 2.02,
1074
+ "learning_rate": 9.039240506329114e-06,
1075
+ "loss": 0.087,
1076
+ "step": 4300
1077
+ },
1078
+ {
1079
+ "epoch": 2.02,
1080
+ "learning_rate": 9.032911392405063e-06,
1081
+ "loss": 0.0875,
1082
+ "step": 4325
1083
+ },
1084
+ {
1085
+ "epoch": 2.02,
1086
+ "learning_rate": 9.026582278481013e-06,
1087
+ "loss": 0.0768,
1088
+ "step": 4350
1089
+ },
1090
+ {
1091
+ "epoch": 2.02,
1092
+ "learning_rate": 9.020253164556962e-06,
1093
+ "loss": 0.0842,
1094
+ "step": 4375
1095
+ },
1096
+ {
1097
+ "epoch": 2.02,
1098
+ "learning_rate": 9.013924050632911e-06,
1099
+ "loss": 0.092,
1100
+ "step": 4400
1101
+ },
1102
+ {
1103
+ "epoch": 2.02,
1104
+ "learning_rate": 9.007594936708861e-06,
1105
+ "loss": 0.0949,
1106
+ "step": 4425
1107
+ },
1108
+ {
1109
+ "epoch": 2.02,
1110
+ "learning_rate": 9.00126582278481e-06,
1111
+ "loss": 0.0989,
1112
+ "step": 4450
1113
+ },
1114
+ {
1115
+ "epoch": 2.02,
1116
+ "learning_rate": 8.99493670886076e-06,
1117
+ "loss": 0.0908,
1118
+ "step": 4475
1119
+ },
1120
+ {
1121
+ "epoch": 2.02,
1122
+ "learning_rate": 8.98860759493671e-06,
1123
+ "loss": 0.0783,
1124
+ "step": 4500
1125
+ },
1126
+ {
1127
+ "epoch": 2.03,
1128
+ "learning_rate": 8.982278481012659e-06,
1129
+ "loss": 0.0651,
1130
+ "step": 4525
1131
+ },
1132
+ {
1133
+ "epoch": 2.03,
1134
+ "learning_rate": 8.975949367088608e-06,
1135
+ "loss": 0.0791,
1136
+ "step": 4550
1137
+ },
1138
+ {
1139
+ "epoch": 2.03,
1140
+ "learning_rate": 8.969620253164557e-06,
1141
+ "loss": 0.0829,
1142
+ "step": 4575
1143
+ },
1144
+ {
1145
+ "epoch": 2.03,
1146
+ "learning_rate": 8.963291139240507e-06,
1147
+ "loss": 0.0797,
1148
+ "step": 4600
1149
+ },
1150
+ {
1151
+ "epoch": 2.03,
1152
+ "learning_rate": 8.956962025316456e-06,
1153
+ "loss": 0.0743,
1154
+ "step": 4625
1155
+ },
1156
+ {
1157
+ "epoch": 2.03,
1158
+ "learning_rate": 8.950632911392406e-06,
1159
+ "loss": 0.0814,
1160
+ "step": 4650
1161
+ },
1162
+ {
1163
+ "epoch": 2.03,
1164
+ "learning_rate": 8.944303797468355e-06,
1165
+ "loss": 0.0594,
1166
+ "step": 4675
1167
+ },
1168
+ {
1169
+ "epoch": 2.03,
1170
+ "learning_rate": 8.937974683544305e-06,
1171
+ "loss": 0.0709,
1172
+ "step": 4700
1173
+ },
1174
+ {
1175
+ "epoch": 2.03,
1176
+ "learning_rate": 8.931645569620254e-06,
1177
+ "loss": 0.0621,
1178
+ "step": 4725
1179
+ },
1180
+ {
1181
+ "epoch": 2.03,
1182
+ "learning_rate": 8.925316455696203e-06,
1183
+ "loss": 0.0599,
1184
+ "step": 4750
1185
+ },
1186
+ {
1187
+ "epoch": 2.03,
1188
+ "learning_rate": 8.918987341772153e-06,
1189
+ "loss": 0.0625,
1190
+ "step": 4775
1191
+ },
1192
+ {
1193
+ "epoch": 2.03,
1194
+ "learning_rate": 8.912658227848102e-06,
1195
+ "loss": 0.0541,
1196
+ "step": 4800
1197
+ },
1198
+ {
1199
+ "epoch": 2.03,
1200
+ "learning_rate": 8.906329113924052e-06,
1201
+ "loss": 0.059,
1202
+ "step": 4825
1203
+ },
1204
+ {
1205
+ "epoch": 2.03,
1206
+ "learning_rate": 8.900000000000001e-06,
1207
+ "loss": 0.0576,
1208
+ "step": 4850
1209
+ },
1210
+ {
1211
+ "epoch": 2.03,
1212
+ "learning_rate": 8.89367088607595e-06,
1213
+ "loss": 0.0623,
1214
+ "step": 4875
1215
+ },
1216
+ {
1217
+ "epoch": 2.03,
1218
+ "learning_rate": 8.8873417721519e-06,
1219
+ "loss": 0.0592,
1220
+ "step": 4900
1221
+ },
1222
+ {
1223
+ "epoch": 2.04,
1224
+ "learning_rate": 8.88101265822785e-06,
1225
+ "loss": 0.0692,
1226
+ "step": 4925
1227
+ },
1228
+ {
1229
+ "epoch": 2.04,
1230
+ "learning_rate": 8.874683544303799e-06,
1231
+ "loss": 0.077,
1232
+ "step": 4950
1233
+ },
1234
+ {
1235
+ "epoch": 2.04,
1236
+ "learning_rate": 8.868354430379748e-06,
1237
+ "loss": 0.0662,
1238
+ "step": 4975
1239
+ },
1240
+ {
1241
+ "epoch": 2.04,
1242
+ "learning_rate": 8.862025316455696e-06,
1243
+ "loss": 0.0605,
1244
+ "step": 5000
1245
+ },
1246
+ {
1247
+ "epoch": 2.04,
1248
+ "eval_loss": 0.2468421906232834,
1249
+ "eval_runtime": 1442.9116,
1250
+ "eval_samples_per_second": 7.215,
1251
+ "eval_steps_per_second": 0.451,
1252
+ "eval_wer": 33.46883689165677,
1253
+ "step": 5000
1254
+ },
1255
+ {
1256
+ "epoch": 2.04,
1257
+ "learning_rate": 8.855696202531645e-06,
1258
+ "loss": 0.0769,
1259
+ "step": 5025
1260
+ },
1261
+ {
1262
+ "epoch": 2.04,
1263
+ "learning_rate": 8.849367088607595e-06,
1264
+ "loss": 0.0601,
1265
+ "step": 5050
1266
+ },
1267
+ {
1268
+ "epoch": 2.04,
1269
+ "learning_rate": 8.843037974683544e-06,
1270
+ "loss": 0.0698,
1271
+ "step": 5075
1272
+ },
1273
+ {
1274
+ "epoch": 2.04,
1275
+ "learning_rate": 8.836708860759494e-06,
1276
+ "loss": 0.0593,
1277
+ "step": 5100
1278
+ },
1279
+ {
1280
+ "epoch": 2.04,
1281
+ "learning_rate": 8.830379746835443e-06,
1282
+ "loss": 0.0521,
1283
+ "step": 5125
1284
+ },
1285
+ {
1286
+ "epoch": 2.04,
1287
+ "learning_rate": 8.824050632911393e-06,
1288
+ "loss": 0.0687,
1289
+ "step": 5150
1290
+ },
1291
+ {
1292
+ "epoch": 2.04,
1293
+ "learning_rate": 8.817721518987342e-06,
1294
+ "loss": 0.0818,
1295
+ "step": 5175
1296
+ },
1297
+ {
1298
+ "epoch": 2.04,
1299
+ "learning_rate": 8.811392405063291e-06,
1300
+ "loss": 0.0706,
1301
+ "step": 5200
1302
+ },
1303
+ {
1304
+ "epoch": 2.04,
1305
+ "learning_rate": 8.805063291139241e-06,
1306
+ "loss": 0.0734,
1307
+ "step": 5225
1308
+ },
1309
+ {
1310
+ "epoch": 2.04,
1311
+ "learning_rate": 8.79873417721519e-06,
1312
+ "loss": 0.0795,
1313
+ "step": 5250
1314
+ },
1315
+ {
1316
+ "epoch": 3.0,
1317
+ "learning_rate": 8.79240506329114e-06,
1318
+ "loss": 0.0905,
1319
+ "step": 5275
1320
+ },
1321
+ {
1322
+ "epoch": 3.0,
1323
+ "learning_rate": 8.78607594936709e-06,
1324
+ "loss": 0.086,
1325
+ "step": 5300
1326
+ },
1327
+ {
1328
+ "epoch": 3.0,
1329
+ "learning_rate": 8.779746835443039e-06,
1330
+ "loss": 0.105,
1331
+ "step": 5325
1332
+ },
1333
+ {
1334
+ "epoch": 3.0,
1335
+ "learning_rate": 8.773417721518988e-06,
1336
+ "loss": 0.0981,
1337
+ "step": 5350
1338
+ },
1339
+ {
1340
+ "epoch": 3.0,
1341
+ "learning_rate": 8.767088607594937e-06,
1342
+ "loss": 0.1131,
1343
+ "step": 5375
1344
+ },
1345
+ {
1346
+ "epoch": 3.0,
1347
+ "learning_rate": 8.760759493670887e-06,
1348
+ "loss": 0.0916,
1349
+ "step": 5400
1350
+ },
1351
+ {
1352
+ "epoch": 3.0,
1353
+ "learning_rate": 8.754430379746836e-06,
1354
+ "loss": 0.0974,
1355
+ "step": 5425
1356
+ },
1357
+ {
1358
+ "epoch": 3.0,
1359
+ "learning_rate": 8.748101265822786e-06,
1360
+ "loss": 0.0899,
1361
+ "step": 5450
1362
+ },
1363
+ {
1364
+ "epoch": 3.01,
1365
+ "learning_rate": 8.741772151898735e-06,
1366
+ "loss": 0.1104,
1367
+ "step": 5475
1368
+ },
1369
+ {
1370
+ "epoch": 3.01,
1371
+ "learning_rate": 8.735443037974685e-06,
1372
+ "loss": 0.0818,
1373
+ "step": 5500
1374
+ },
1375
+ {
1376
+ "epoch": 3.01,
1377
+ "learning_rate": 8.729113924050634e-06,
1378
+ "loss": 0.0764,
1379
+ "step": 5525
1380
+ },
1381
+ {
1382
+ "epoch": 3.01,
1383
+ "learning_rate": 8.722784810126583e-06,
1384
+ "loss": 0.0709,
1385
+ "step": 5550
1386
+ },
1387
+ {
1388
+ "epoch": 3.01,
1389
+ "learning_rate": 8.716455696202533e-06,
1390
+ "loss": 0.0891,
1391
+ "step": 5575
1392
+ },
1393
+ {
1394
+ "epoch": 3.01,
1395
+ "learning_rate": 8.710126582278482e-06,
1396
+ "loss": 0.0869,
1397
+ "step": 5600
1398
+ },
1399
+ {
1400
+ "epoch": 3.01,
1401
+ "learning_rate": 8.703797468354432e-06,
1402
+ "loss": 0.0631,
1403
+ "step": 5625
1404
+ },
1405
+ {
1406
+ "epoch": 3.01,
1407
+ "learning_rate": 8.697468354430381e-06,
1408
+ "loss": 0.0795,
1409
+ "step": 5650
1410
+ },
1411
+ {
1412
+ "epoch": 3.01,
1413
+ "learning_rate": 8.69113924050633e-06,
1414
+ "loss": 0.0603,
1415
+ "step": 5675
1416
+ },
1417
+ {
1418
+ "epoch": 3.01,
1419
+ "learning_rate": 8.68481012658228e-06,
1420
+ "loss": 0.064,
1421
+ "step": 5700
1422
+ },
1423
+ {
1424
+ "epoch": 3.01,
1425
+ "learning_rate": 8.67848101265823e-06,
1426
+ "loss": 0.0738,
1427
+ "step": 5725
1428
+ },
1429
+ {
1430
+ "epoch": 3.01,
1431
+ "learning_rate": 8.672151898734177e-06,
1432
+ "loss": 0.0796,
1433
+ "step": 5750
1434
+ },
1435
+ {
1436
+ "epoch": 3.01,
1437
+ "learning_rate": 8.665822784810127e-06,
1438
+ "loss": 0.0723,
1439
+ "step": 5775
1440
+ },
1441
+ {
1442
+ "epoch": 3.01,
1443
+ "learning_rate": 8.659493670886076e-06,
1444
+ "loss": 0.0592,
1445
+ "step": 5800
1446
+ },
1447
+ {
1448
+ "epoch": 3.01,
1449
+ "learning_rate": 8.653164556962025e-06,
1450
+ "loss": 0.051,
1451
+ "step": 5825
1452
+ },
1453
+ {
1454
+ "epoch": 3.01,
1455
+ "learning_rate": 8.646835443037975e-06,
1456
+ "loss": 0.0716,
1457
+ "step": 5850
1458
+ },
1459
+ {
1460
+ "epoch": 3.02,
1461
+ "learning_rate": 8.640506329113924e-06,
1462
+ "loss": 0.0598,
1463
+ "step": 5875
1464
+ },
1465
+ {
1466
+ "epoch": 3.02,
1467
+ "learning_rate": 8.634177215189874e-06,
1468
+ "loss": 0.0524,
1469
+ "step": 5900
1470
+ },
1471
+ {
1472
+ "epoch": 3.02,
1473
+ "learning_rate": 8.627848101265823e-06,
1474
+ "loss": 0.0623,
1475
+ "step": 5925
1476
+ },
1477
+ {
1478
+ "epoch": 3.02,
1479
+ "learning_rate": 8.621518987341773e-06,
1480
+ "loss": 0.054,
1481
+ "step": 5950
1482
+ },
1483
+ {
1484
+ "epoch": 3.02,
1485
+ "learning_rate": 8.615189873417722e-06,
1486
+ "loss": 0.057,
1487
+ "step": 5975
1488
+ },
1489
+ {
1490
+ "epoch": 3.02,
1491
+ "learning_rate": 8.608860759493671e-06,
1492
+ "loss": 0.0479,
1493
+ "step": 6000
1494
+ },
1495
+ {
1496
+ "epoch": 3.02,
1497
+ "eval_loss": 0.24428367614746094,
1498
+ "eval_runtime": 1457.5568,
1499
+ "eval_samples_per_second": 7.143,
1500
+ "eval_steps_per_second": 0.447,
1501
+ "eval_wer": 36.51929299663409,
1502
+ "step": 6000
1503
+ },
1504
+ {
1505
+ "epoch": 3.02,
1506
+ "learning_rate": 8.602531645569621e-06,
1507
+ "loss": 0.0501,
1508
+ "step": 6025
1509
+ },
1510
+ {
1511
+ "epoch": 3.02,
1512
+ "learning_rate": 8.59620253164557e-06,
1513
+ "loss": 0.0527,
1514
+ "step": 6050
1515
+ },
1516
+ {
1517
+ "epoch": 3.02,
1518
+ "learning_rate": 8.58987341772152e-06,
1519
+ "loss": 0.0558,
1520
+ "step": 6075
1521
+ },
1522
+ {
1523
+ "epoch": 3.02,
1524
+ "learning_rate": 8.58354430379747e-06,
1525
+ "loss": 0.0548,
1526
+ "step": 6100
1527
+ },
1528
+ {
1529
+ "epoch": 3.02,
1530
+ "learning_rate": 8.577215189873419e-06,
1531
+ "loss": 0.0503,
1532
+ "step": 6125
1533
+ },
1534
+ {
1535
+ "epoch": 3.02,
1536
+ "learning_rate": 8.570886075949368e-06,
1537
+ "loss": 0.0641,
1538
+ "step": 6150
1539
+ },
1540
+ {
1541
+ "epoch": 3.02,
1542
+ "learning_rate": 8.564556962025316e-06,
1543
+ "loss": 0.0748,
1544
+ "step": 6175
1545
+ },
1546
+ {
1547
+ "epoch": 3.02,
1548
+ "learning_rate": 8.558227848101267e-06,
1549
+ "loss": 0.0775,
1550
+ "step": 6200
1551
+ },
1552
+ {
1553
+ "epoch": 3.02,
1554
+ "learning_rate": 8.551898734177216e-06,
1555
+ "loss": 0.0662,
1556
+ "step": 6225
1557
+ },
1558
+ {
1559
+ "epoch": 3.02,
1560
+ "learning_rate": 8.545569620253166e-06,
1561
+ "loss": 0.0591,
1562
+ "step": 6250
1563
+ },
1564
+ {
1565
+ "epoch": 3.03,
1566
+ "learning_rate": 8.539240506329115e-06,
1567
+ "loss": 0.0421,
1568
+ "step": 6275
1569
+ },
1570
+ {
1571
+ "epoch": 3.03,
1572
+ "learning_rate": 8.532911392405065e-06,
1573
+ "loss": 0.0565,
1574
+ "step": 6300
1575
+ },
1576
+ {
1577
+ "epoch": 3.03,
1578
+ "learning_rate": 8.526582278481014e-06,
1579
+ "loss": 0.0574,
1580
+ "step": 6325
1581
+ },
1582
+ {
1583
+ "epoch": 3.03,
1584
+ "learning_rate": 8.520253164556963e-06,
1585
+ "loss": 0.0466,
1586
+ "step": 6350
1587
+ },
1588
+ {
1589
+ "epoch": 3.03,
1590
+ "learning_rate": 8.513924050632913e-06,
1591
+ "loss": 0.0486,
1592
+ "step": 6375
1593
+ },
1594
+ {
1595
+ "epoch": 3.03,
1596
+ "learning_rate": 8.507594936708862e-06,
1597
+ "loss": 0.0502,
1598
+ "step": 6400
1599
+ },
1600
+ {
1601
+ "epoch": 3.03,
1602
+ "learning_rate": 8.501265822784812e-06,
1603
+ "loss": 0.0415,
1604
+ "step": 6425
1605
+ },
1606
+ {
1607
+ "epoch": 3.03,
1608
+ "learning_rate": 8.494936708860761e-06,
1609
+ "loss": 0.0449,
1610
+ "step": 6450
1611
+ },
1612
+ {
1613
+ "epoch": 3.03,
1614
+ "learning_rate": 8.488607594936709e-06,
1615
+ "loss": 0.0525,
1616
+ "step": 6475
1617
+ },
1618
+ {
1619
+ "epoch": 3.03,
1620
+ "learning_rate": 8.482278481012658e-06,
1621
+ "loss": 0.0401,
1622
+ "step": 6500
1623
+ },
1624
+ {
1625
+ "epoch": 3.03,
1626
+ "learning_rate": 8.475949367088608e-06,
1627
+ "loss": 0.0502,
1628
+ "step": 6525
1629
+ },
1630
+ {
1631
+ "epoch": 3.03,
1632
+ "learning_rate": 8.469620253164557e-06,
1633
+ "loss": 0.0453,
1634
+ "step": 6550
1635
+ },
1636
+ {
1637
+ "epoch": 3.03,
1638
+ "learning_rate": 8.463291139240507e-06,
1639
+ "loss": 0.0442,
1640
+ "step": 6575
1641
+ },
1642
+ {
1643
+ "epoch": 3.03,
1644
+ "learning_rate": 8.456962025316456e-06,
1645
+ "loss": 0.0423,
1646
+ "step": 6600
1647
+ },
1648
+ {
1649
+ "epoch": 3.03,
1650
+ "learning_rate": 8.450632911392405e-06,
1651
+ "loss": 0.0518,
1652
+ "step": 6625
1653
+ },
1654
+ {
1655
+ "epoch": 3.03,
1656
+ "learning_rate": 8.444303797468355e-06,
1657
+ "loss": 0.0479,
1658
+ "step": 6650
1659
+ },
1660
+ {
1661
+ "epoch": 3.04,
1662
+ "learning_rate": 8.437974683544304e-06,
1663
+ "loss": 0.0521,
1664
+ "step": 6675
1665
+ },
1666
+ {
1667
+ "epoch": 3.04,
1668
+ "learning_rate": 8.431645569620254e-06,
1669
+ "loss": 0.0436,
1670
+ "step": 6700
1671
+ },
1672
+ {
1673
+ "epoch": 3.04,
1674
+ "learning_rate": 8.425316455696203e-06,
1675
+ "loss": 0.0539,
1676
+ "step": 6725
1677
+ },
1678
+ {
1679
+ "epoch": 3.04,
1680
+ "learning_rate": 8.418987341772153e-06,
1681
+ "loss": 0.0446,
1682
+ "step": 6750
1683
+ },
1684
+ {
1685
+ "epoch": 3.04,
1686
+ "learning_rate": 8.412658227848102e-06,
1687
+ "loss": 0.0521,
1688
+ "step": 6775
1689
+ },
1690
+ {
1691
+ "epoch": 3.04,
1692
+ "learning_rate": 8.406329113924051e-06,
1693
+ "loss": 0.0391,
1694
+ "step": 6800
1695
+ },
1696
+ {
1697
+ "epoch": 3.04,
1698
+ "learning_rate": 8.400000000000001e-06,
1699
+ "loss": 0.0513,
1700
+ "step": 6825
1701
+ },
1702
+ {
1703
+ "epoch": 3.04,
1704
+ "learning_rate": 8.39367088607595e-06,
1705
+ "loss": 0.0449,
1706
+ "step": 6850
1707
+ },
1708
+ {
1709
+ "epoch": 3.04,
1710
+ "learning_rate": 8.3873417721519e-06,
1711
+ "loss": 0.0463,
1712
+ "step": 6875
1713
+ },
1714
+ {
1715
+ "epoch": 3.04,
1716
+ "learning_rate": 8.38101265822785e-06,
1717
+ "loss": 0.0495,
1718
+ "step": 6900
1719
+ },
1720
+ {
1721
+ "epoch": 3.04,
1722
+ "learning_rate": 8.374683544303797e-06,
1723
+ "loss": 0.0628,
1724
+ "step": 6925
1725
+ },
1726
+ {
1727
+ "epoch": 3.04,
1728
+ "learning_rate": 8.368354430379746e-06,
1729
+ "loss": 0.0594,
1730
+ "step": 6950
1731
+ },
1732
+ {
1733
+ "epoch": 3.04,
1734
+ "learning_rate": 8.362025316455696e-06,
1735
+ "loss": 0.0573,
1736
+ "step": 6975
1737
+ },
1738
+ {
1739
+ "epoch": 3.04,
1740
+ "learning_rate": 8.355696202531645e-06,
1741
+ "loss": 0.057,
1742
+ "step": 7000
1743
+ },
1744
+ {
1745
+ "epoch": 3.04,
1746
+ "eval_loss": 0.25129273533821106,
1747
+ "eval_runtime": 1479.2043,
1748
+ "eval_samples_per_second": 7.038,
1749
+ "eval_steps_per_second": 0.44,
1750
+ "eval_wer": 57.010683100538216,
1751
+ "step": 7000
1752
+ },
1753
+ {
1754
+ "epoch": 4.0,
1755
+ "learning_rate": 8.349367088607596e-06,
1756
+ "loss": 0.0534,
1757
+ "step": 7025
1758
+ },
1759
+ {
1760
+ "epoch": 4.0,
1761
+ "learning_rate": 8.343037974683546e-06,
1762
+ "loss": 0.0663,
1763
+ "step": 7050
1764
+ },
1765
+ {
1766
+ "epoch": 4.0,
1767
+ "learning_rate": 8.336708860759495e-06,
1768
+ "loss": 0.0698,
1769
+ "step": 7075
1770
+ },
1771
+ {
1772
+ "epoch": 4.0,
1773
+ "learning_rate": 8.330379746835445e-06,
1774
+ "loss": 0.0633,
1775
+ "step": 7100
1776
+ },
1777
+ {
1778
+ "epoch": 4.0,
1779
+ "learning_rate": 8.324050632911394e-06,
1780
+ "loss": 0.0808,
1781
+ "step": 7125
1782
+ },
1783
+ {
1784
+ "epoch": 4.0,
1785
+ "learning_rate": 8.317721518987343e-06,
1786
+ "loss": 0.0699,
1787
+ "step": 7150
1788
+ },
1789
+ {
1790
+ "epoch": 4.0,
1791
+ "learning_rate": 8.311392405063293e-06,
1792
+ "loss": 0.0736,
1793
+ "step": 7175
1794
+ },
1795
+ {
1796
+ "epoch": 4.0,
1797
+ "learning_rate": 8.305063291139242e-06,
1798
+ "loss": 0.069,
1799
+ "step": 7200
1800
+ },
1801
+ {
1802
+ "epoch": 4.01,
1803
+ "learning_rate": 8.29873417721519e-06,
1804
+ "loss": 0.0718,
1805
+ "step": 7225
1806
+ },
1807
+ {
1808
+ "epoch": 4.01,
1809
+ "learning_rate": 8.29240506329114e-06,
1810
+ "loss": 0.0674,
1811
+ "step": 7250
1812
+ },
1813
+ {
1814
+ "epoch": 4.01,
1815
+ "learning_rate": 8.286075949367089e-06,
1816
+ "loss": 0.0568,
1817
+ "step": 7275
1818
+ },
1819
+ {
1820
+ "epoch": 4.01,
1821
+ "learning_rate": 8.279746835443038e-06,
1822
+ "loss": 0.0497,
1823
+ "step": 7300
1824
+ },
1825
+ {
1826
+ "epoch": 4.01,
1827
+ "learning_rate": 8.273417721518988e-06,
1828
+ "loss": 0.0563,
1829
+ "step": 7325
1830
+ },
1831
+ {
1832
+ "epoch": 4.01,
1833
+ "learning_rate": 8.267088607594937e-06,
1834
+ "loss": 0.0649,
1835
+ "step": 7350
1836
+ },
1837
+ {
1838
+ "epoch": 4.01,
1839
+ "learning_rate": 8.260759493670887e-06,
1840
+ "loss": 0.042,
1841
+ "step": 7375
1842
+ },
1843
+ {
1844
+ "epoch": 4.01,
1845
+ "learning_rate": 8.254430379746836e-06,
1846
+ "loss": 0.0559,
1847
+ "step": 7400
1848
+ },
1849
+ {
1850
+ "epoch": 4.01,
1851
+ "learning_rate": 8.248101265822785e-06,
1852
+ "loss": 0.0476,
1853
+ "step": 7425
1854
+ },
1855
+ {
1856
+ "epoch": 4.01,
1857
+ "learning_rate": 8.241772151898735e-06,
1858
+ "loss": 0.0427,
1859
+ "step": 7450
1860
+ },
1861
+ {
1862
+ "epoch": 4.01,
1863
+ "learning_rate": 8.235443037974684e-06,
1864
+ "loss": 0.0533,
1865
+ "step": 7475
1866
+ },
1867
+ {
1868
+ "epoch": 4.01,
1869
+ "learning_rate": 8.229113924050634e-06,
1870
+ "loss": 0.0633,
1871
+ "step": 7500
1872
+ },
1873
+ {
1874
+ "epoch": 4.01,
1875
+ "learning_rate": 8.222784810126583e-06,
1876
+ "loss": 0.055,
1877
+ "step": 7525
1878
+ },
1879
+ {
1880
+ "epoch": 4.01,
1881
+ "learning_rate": 8.216455696202533e-06,
1882
+ "loss": 0.043,
1883
+ "step": 7550
1884
+ },
1885
+ {
1886
+ "epoch": 4.01,
1887
+ "learning_rate": 8.210126582278482e-06,
1888
+ "loss": 0.0421,
1889
+ "step": 7575
1890
+ },
1891
+ {
1892
+ "epoch": 4.01,
1893
+ "learning_rate": 8.203797468354431e-06,
1894
+ "loss": 0.0519,
1895
+ "step": 7600
1896
+ },
1897
+ {
1898
+ "epoch": 4.02,
1899
+ "learning_rate": 8.197468354430381e-06,
1900
+ "loss": 0.0398,
1901
+ "step": 7625
1902
+ },
1903
+ {
1904
+ "epoch": 4.02,
1905
+ "learning_rate": 8.19113924050633e-06,
1906
+ "loss": 0.041,
1907
+ "step": 7650
1908
+ },
1909
+ {
1910
+ "epoch": 4.02,
1911
+ "learning_rate": 8.184810126582278e-06,
1912
+ "loss": 0.0477,
1913
+ "step": 7675
1914
+ },
1915
+ {
1916
+ "epoch": 4.02,
1917
+ "learning_rate": 8.178481012658227e-06,
1918
+ "loss": 0.046,
1919
+ "step": 7700
1920
+ },
1921
+ {
1922
+ "epoch": 4.02,
1923
+ "learning_rate": 8.172151898734177e-06,
1924
+ "loss": 0.0465,
1925
+ "step": 7725
1926
+ },
1927
+ {
1928
+ "epoch": 4.02,
1929
+ "learning_rate": 8.165822784810126e-06,
1930
+ "loss": 0.0359,
1931
+ "step": 7750
1932
+ },
1933
+ {
1934
+ "epoch": 4.02,
1935
+ "learning_rate": 8.159493670886076e-06,
1936
+ "loss": 0.0362,
1937
+ "step": 7775
1938
+ },
1939
+ {
1940
+ "epoch": 4.02,
1941
+ "learning_rate": 8.153164556962025e-06,
1942
+ "loss": 0.04,
1943
+ "step": 7800
1944
+ },
1945
+ {
1946
+ "epoch": 4.02,
1947
+ "learning_rate": 8.146835443037975e-06,
1948
+ "loss": 0.0433,
1949
+ "step": 7825
1950
+ },
1951
+ {
1952
+ "epoch": 4.02,
1953
+ "learning_rate": 8.140506329113926e-06,
1954
+ "loss": 0.0399,
1955
+ "step": 7850
1956
+ },
1957
+ {
1958
+ "epoch": 4.02,
1959
+ "learning_rate": 8.134177215189875e-06,
1960
+ "loss": 0.0396,
1961
+ "step": 7875
1962
+ },
1963
+ {
1964
+ "epoch": 4.02,
1965
+ "learning_rate": 8.127848101265825e-06,
1966
+ "loss": 0.0472,
1967
+ "step": 7900
1968
+ },
1969
+ {
1970
+ "epoch": 4.02,
1971
+ "learning_rate": 8.121518987341774e-06,
1972
+ "loss": 0.0533,
1973
+ "step": 7925
1974
+ },
1975
+ {
1976
+ "epoch": 4.02,
1977
+ "learning_rate": 8.115189873417722e-06,
1978
+ "loss": 0.0564,
1979
+ "step": 7950
1980
+ },
1981
+ {
1982
+ "epoch": 4.02,
1983
+ "learning_rate": 8.108860759493671e-06,
1984
+ "loss": 0.0508,
1985
+ "step": 7975
1986
+ },
1987
+ {
1988
+ "epoch": 4.02,
1989
+ "learning_rate": 8.10253164556962e-06,
1990
+ "loss": 0.0375,
1991
+ "step": 8000
1992
+ },
1993
+ {
1994
+ "epoch": 4.02,
1995
+ "eval_loss": 0.252152681350708,
1996
+ "eval_runtime": 1456.2607,
1997
+ "eval_samples_per_second": 7.149,
1998
+ "eval_steps_per_second": 0.447,
1999
+ "eval_wer": 44.38120945056017,
2000
+ "step": 8000
2001
+ },
2002
+ {
2003
+ "epoch": 4.03,
2004
+ "learning_rate": 8.09620253164557e-06,
2005
+ "loss": 0.0296,
2006
+ "step": 8025
2007
+ },
2008
+ {
2009
+ "epoch": 4.03,
2010
+ "learning_rate": 8.08987341772152e-06,
2011
+ "loss": 0.0395,
2012
+ "step": 8050
2013
+ },
2014
+ {
2015
+ "epoch": 4.03,
2016
+ "learning_rate": 8.083544303797469e-06,
2017
+ "loss": 0.0409,
2018
+ "step": 8075
2019
+ },
2020
+ {
2021
+ "epoch": 4.03,
2022
+ "learning_rate": 8.077215189873418e-06,
2023
+ "loss": 0.0343,
2024
+ "step": 8100
2025
+ },
2026
+ {
2027
+ "epoch": 4.03,
2028
+ "learning_rate": 8.070886075949368e-06,
2029
+ "loss": 0.0321,
2030
+ "step": 8125
2031
+ },
2032
+ {
2033
+ "epoch": 4.03,
2034
+ "learning_rate": 8.064556962025317e-06,
2035
+ "loss": 0.0333,
2036
+ "step": 8150
2037
+ },
2038
+ {
2039
+ "epoch": 4.03,
2040
+ "learning_rate": 8.058227848101267e-06,
2041
+ "loss": 0.0327,
2042
+ "step": 8175
2043
+ },
2044
+ {
2045
+ "epoch": 4.03,
2046
+ "learning_rate": 8.051898734177216e-06,
2047
+ "loss": 0.0318,
2048
+ "step": 8200
2049
+ },
2050
+ {
2051
+ "epoch": 4.03,
2052
+ "learning_rate": 8.045569620253165e-06,
2053
+ "loss": 0.0405,
2054
+ "step": 8225
2055
+ },
2056
+ {
2057
+ "epoch": 4.03,
2058
+ "learning_rate": 8.039240506329115e-06,
2059
+ "loss": 0.0382,
2060
+ "step": 8250
2061
+ },
2062
+ {
2063
+ "epoch": 4.03,
2064
+ "learning_rate": 8.032911392405064e-06,
2065
+ "loss": 0.0317,
2066
+ "step": 8275
2067
+ },
2068
+ {
2069
+ "epoch": 4.03,
2070
+ "learning_rate": 8.026582278481014e-06,
2071
+ "loss": 0.0285,
2072
+ "step": 8300
2073
+ },
2074
+ {
2075
+ "epoch": 4.03,
2076
+ "learning_rate": 8.020253164556963e-06,
2077
+ "loss": 0.0285,
2078
+ "step": 8325
2079
+ },
2080
+ {
2081
+ "epoch": 4.03,
2082
+ "learning_rate": 8.013924050632913e-06,
2083
+ "loss": 0.0269,
2084
+ "step": 8350
2085
+ },
2086
+ {
2087
+ "epoch": 4.03,
2088
+ "learning_rate": 8.007594936708862e-06,
2089
+ "loss": 0.0334,
2090
+ "step": 8375
2091
+ },
2092
+ {
2093
+ "epoch": 4.03,
2094
+ "learning_rate": 8.00126582278481e-06,
2095
+ "loss": 0.0406,
2096
+ "step": 8400
2097
+ },
2098
+ {
2099
+ "epoch": 4.04,
2100
+ "learning_rate": 7.99493670886076e-06,
2101
+ "loss": 0.038,
2102
+ "step": 8425
2103
+ },
2104
+ {
2105
+ "epoch": 4.04,
2106
+ "learning_rate": 7.988607594936709e-06,
2107
+ "loss": 0.0315,
2108
+ "step": 8450
2109
+ },
2110
+ {
2111
+ "epoch": 4.04,
2112
+ "learning_rate": 7.982278481012658e-06,
2113
+ "loss": 0.0329,
2114
+ "step": 8475
2115
+ },
2116
+ {
2117
+ "epoch": 4.04,
2118
+ "learning_rate": 7.975949367088607e-06,
2119
+ "loss": 0.0373,
2120
+ "step": 8500
2121
+ },
2122
+ {
2123
+ "epoch": 4.04,
2124
+ "learning_rate": 7.969620253164557e-06,
2125
+ "loss": 0.0514,
2126
+ "step": 8525
2127
+ },
2128
+ {
2129
+ "epoch": 4.04,
2130
+ "learning_rate": 7.963291139240506e-06,
2131
+ "loss": 0.0349,
2132
+ "step": 8550
2133
+ },
2134
+ {
2135
+ "epoch": 4.04,
2136
+ "learning_rate": 7.956962025316456e-06,
2137
+ "loss": 0.0385,
2138
+ "step": 8575
2139
+ },
2140
+ {
2141
+ "epoch": 4.04,
2142
+ "learning_rate": 7.950632911392405e-06,
2143
+ "loss": 0.0367,
2144
+ "step": 8600
2145
+ },
2146
+ {
2147
+ "epoch": 4.04,
2148
+ "learning_rate": 7.944303797468355e-06,
2149
+ "loss": 0.0317,
2150
+ "step": 8625
2151
+ },
2152
+ {
2153
+ "epoch": 4.04,
2154
+ "learning_rate": 7.937974683544304e-06,
2155
+ "loss": 0.0368,
2156
+ "step": 8650
2157
+ },
2158
+ {
2159
+ "epoch": 4.04,
2160
+ "learning_rate": 7.931645569620255e-06,
2161
+ "loss": 0.045,
2162
+ "step": 8675
2163
+ },
2164
+ {
2165
+ "epoch": 4.04,
2166
+ "learning_rate": 7.925316455696203e-06,
2167
+ "loss": 0.0456,
2168
+ "step": 8700
2169
+ },
2170
+ {
2171
+ "epoch": 4.04,
2172
+ "learning_rate": 7.918987341772152e-06,
2173
+ "loss": 0.0421,
2174
+ "step": 8725
2175
+ },
2176
+ {
2177
+ "epoch": 4.04,
2178
+ "learning_rate": 7.912658227848102e-06,
2179
+ "loss": 0.043,
2180
+ "step": 8750
2181
+ },
2182
+ {
2183
+ "epoch": 5.0,
2184
+ "learning_rate": 7.906329113924051e-06,
2185
+ "loss": 0.0428,
2186
+ "step": 8775
2187
+ },
2188
+ {
2189
+ "epoch": 5.0,
2190
+ "learning_rate": 7.9e-06,
2191
+ "loss": 0.0459,
2192
+ "step": 8800
2193
+ },
2194
+ {
2195
+ "epoch": 5.0,
2196
+ "learning_rate": 7.89367088607595e-06,
2197
+ "loss": 0.0467,
2198
+ "step": 8825
2199
+ },
2200
+ {
2201
+ "epoch": 5.0,
2202
+ "learning_rate": 7.8873417721519e-06,
2203
+ "loss": 0.0462,
2204
+ "step": 8850
2205
+ },
2206
+ {
2207
+ "epoch": 5.0,
2208
+ "learning_rate": 7.881012658227849e-06,
2209
+ "loss": 0.0578,
2210
+ "step": 8875
2211
+ },
2212
+ {
2213
+ "epoch": 5.0,
2214
+ "learning_rate": 7.874683544303798e-06,
2215
+ "loss": 0.0422,
2216
+ "step": 8900
2217
+ },
2218
+ {
2219
+ "epoch": 5.0,
2220
+ "learning_rate": 7.868354430379748e-06,
2221
+ "loss": 0.0546,
2222
+ "step": 8925
2223
+ },
2224
+ {
2225
+ "epoch": 5.0,
2226
+ "learning_rate": 7.862025316455697e-06,
2227
+ "loss": 0.0515,
2228
+ "step": 8950
2229
+ },
2230
+ {
2231
+ "epoch": 5.01,
2232
+ "learning_rate": 7.855696202531647e-06,
2233
+ "loss": 0.0541,
2234
+ "step": 8975
2235
+ },
2236
+ {
2237
+ "epoch": 5.01,
2238
+ "learning_rate": 7.849367088607596e-06,
2239
+ "loss": 0.0479,
2240
+ "step": 9000
2241
+ },
2242
+ {
2243
+ "epoch": 5.01,
2244
+ "eval_loss": 0.2375001162290573,
2245
+ "eval_runtime": 1456.1088,
2246
+ "eval_samples_per_second": 7.15,
2247
+ "eval_steps_per_second": 0.447,
2248
+ "eval_wer": 31.58262735979447,
2249
+ "step": 9000
2250
+ },
2251
+ {
2252
+ "epoch": 5.01,
2253
+ "learning_rate": 7.843037974683545e-06,
2254
+ "loss": 0.0434,
2255
+ "step": 9025
2256
+ },
2257
+ {
2258
+ "epoch": 5.01,
2259
+ "learning_rate": 7.836708860759495e-06,
2260
+ "loss": 0.033,
2261
+ "step": 9050
2262
+ },
2263
+ {
2264
+ "epoch": 5.01,
2265
+ "learning_rate": 7.830379746835444e-06,
2266
+ "loss": 0.0494,
2267
+ "step": 9075
2268
+ },
2269
+ {
2270
+ "epoch": 5.01,
2271
+ "learning_rate": 7.824050632911394e-06,
2272
+ "loss": 0.0522,
2273
+ "step": 9100
2274
+ },
2275
+ {
2276
+ "epoch": 5.01,
2277
+ "learning_rate": 7.817721518987343e-06,
2278
+ "loss": 0.0338,
2279
+ "step": 9125
2280
+ },
2281
+ {
2282
+ "epoch": 5.01,
2283
+ "learning_rate": 7.811392405063291e-06,
2284
+ "loss": 0.0395,
2285
+ "step": 9150
2286
+ },
2287
+ {
2288
+ "epoch": 5.01,
2289
+ "learning_rate": 7.80506329113924e-06,
2290
+ "loss": 0.0332,
2291
+ "step": 9175
2292
+ },
2293
+ {
2294
+ "epoch": 5.01,
2295
+ "learning_rate": 7.79873417721519e-06,
2296
+ "loss": 0.0381,
2297
+ "step": 9200
2298
+ },
2299
+ {
2300
+ "epoch": 5.01,
2301
+ "learning_rate": 7.79240506329114e-06,
2302
+ "loss": 0.0409,
2303
+ "step": 9225
2304
+ },
2305
+ {
2306
+ "epoch": 5.01,
2307
+ "learning_rate": 7.786075949367089e-06,
2308
+ "loss": 0.0529,
2309
+ "step": 9250
2310
+ },
2311
+ {
2312
+ "epoch": 5.01,
2313
+ "learning_rate": 7.779746835443038e-06,
2314
+ "loss": 0.0498,
2315
+ "step": 9275
2316
+ },
2317
+ {
2318
+ "epoch": 5.01,
2319
+ "learning_rate": 7.773417721518987e-06,
2320
+ "loss": 0.0379,
2321
+ "step": 9300
2322
+ },
2323
+ {
2324
+ "epoch": 5.01,
2325
+ "learning_rate": 7.767088607594937e-06,
2326
+ "loss": 0.0316,
2327
+ "step": 9325
2328
+ },
2329
+ {
2330
+ "epoch": 5.01,
2331
+ "learning_rate": 7.760759493670886e-06,
2332
+ "loss": 0.0398,
2333
+ "step": 9350
2334
+ },
2335
+ {
2336
+ "epoch": 5.02,
2337
+ "learning_rate": 7.754430379746836e-06,
2338
+ "loss": 0.0427,
2339
+ "step": 9375
2340
+ },
2341
+ {
2342
+ "epoch": 5.02,
2343
+ "learning_rate": 7.748101265822785e-06,
2344
+ "loss": 0.0363,
2345
+ "step": 9400
2346
+ },
2347
+ {
2348
+ "epoch": 5.02,
2349
+ "learning_rate": 7.741772151898735e-06,
2350
+ "loss": 0.0393,
2351
+ "step": 9425
2352
+ },
2353
+ {
2354
+ "epoch": 5.02,
2355
+ "learning_rate": 7.735443037974684e-06,
2356
+ "loss": 0.0377,
2357
+ "step": 9450
2358
+ },
2359
+ {
2360
+ "epoch": 5.02,
2361
+ "learning_rate": 7.729113924050633e-06,
2362
+ "loss": 0.0313,
2363
+ "step": 9475
2364
+ },
2365
+ {
2366
+ "epoch": 5.02,
2367
+ "learning_rate": 7.722784810126583e-06,
2368
+ "loss": 0.0359,
2369
+ "step": 9500
2370
+ },
2371
+ {
2372
+ "epoch": 5.02,
2373
+ "learning_rate": 7.716455696202532e-06,
2374
+ "loss": 0.0289,
2375
+ "step": 9525
2376
+ },
2377
+ {
2378
+ "epoch": 5.02,
2379
+ "learning_rate": 7.710126582278482e-06,
2380
+ "loss": 0.0354,
2381
+ "step": 9550
2382
+ },
2383
+ {
2384
+ "epoch": 5.02,
2385
+ "learning_rate": 7.703797468354431e-06,
2386
+ "loss": 0.0309,
2387
+ "step": 9575
2388
+ },
2389
+ {
2390
+ "epoch": 5.02,
2391
+ "learning_rate": 7.69746835443038e-06,
2392
+ "loss": 0.0321,
2393
+ "step": 9600
2394
+ },
2395
+ {
2396
+ "epoch": 5.02,
2397
+ "learning_rate": 7.69113924050633e-06,
2398
+ "loss": 0.0392,
2399
+ "step": 9625
2400
+ },
2401
+ {
2402
+ "epoch": 5.02,
2403
+ "learning_rate": 7.68481012658228e-06,
2404
+ "loss": 0.0362,
2405
+ "step": 9650
2406
+ },
2407
+ {
2408
+ "epoch": 5.02,
2409
+ "learning_rate": 7.678481012658229e-06,
2410
+ "loss": 0.0358,
2411
+ "step": 9675
2412
+ },
2413
+ {
2414
+ "epoch": 5.02,
2415
+ "learning_rate": 7.672151898734178e-06,
2416
+ "loss": 0.045,
2417
+ "step": 9700
2418
+ },
2419
+ {
2420
+ "epoch": 5.02,
2421
+ "learning_rate": 7.665822784810128e-06,
2422
+ "loss": 0.0401,
2423
+ "step": 9725
2424
+ },
2425
+ {
2426
+ "epoch": 5.02,
2427
+ "learning_rate": 7.659493670886077e-06,
2428
+ "loss": 0.0357,
2429
+ "step": 9750
2430
+ },
2431
+ {
2432
+ "epoch": 5.03,
2433
+ "learning_rate": 7.653164556962027e-06,
2434
+ "loss": 0.0256,
2435
+ "step": 9775
2436
+ },
2437
+ {
2438
+ "epoch": 5.03,
2439
+ "learning_rate": 7.646835443037976e-06,
2440
+ "loss": 0.0358,
2441
+ "step": 9800
2442
+ },
2443
+ {
2444
+ "epoch": 5.03,
2445
+ "learning_rate": 7.640506329113925e-06,
2446
+ "loss": 0.0361,
2447
+ "step": 9825
2448
+ },
2449
+ {
2450
+ "epoch": 5.03,
2451
+ "learning_rate": 7.634177215189875e-06,
2452
+ "loss": 0.0365,
2453
+ "step": 9850
2454
+ },
2455
+ {
2456
+ "epoch": 5.03,
2457
+ "learning_rate": 7.6278481012658234e-06,
2458
+ "loss": 0.0344,
2459
+ "step": 9875
2460
+ },
2461
+ {
2462
+ "epoch": 5.03,
2463
+ "learning_rate": 7.621518987341773e-06,
2464
+ "loss": 0.0349,
2465
+ "step": 9900
2466
+ },
2467
+ {
2468
+ "epoch": 5.03,
2469
+ "learning_rate": 7.615189873417722e-06,
2470
+ "loss": 0.0289,
2471
+ "step": 9925
2472
+ },
2473
+ {
2474
+ "epoch": 5.03,
2475
+ "learning_rate": 7.608860759493672e-06,
2476
+ "loss": 0.0259,
2477
+ "step": 9950
2478
+ },
2479
+ {
2480
+ "epoch": 5.03,
2481
+ "learning_rate": 7.602531645569621e-06,
2482
+ "loss": 0.0289,
2483
+ "step": 9975
2484
+ },
2485
+ {
2486
+ "epoch": 5.03,
2487
+ "learning_rate": 7.59620253164557e-06,
2488
+ "loss": 0.0269,
2489
+ "step": 10000
2490
+ },
2491
+ {
2492
+ "epoch": 5.03,
2493
+ "eval_loss": 0.2615405321121216,
2494
+ "eval_runtime": 1420.7197,
2495
+ "eval_samples_per_second": 7.328,
2496
+ "eval_steps_per_second": 0.458,
2497
+ "eval_wer": 36.23310948145498,
2498
+ "step": 10000
2499
+ },
2500
+ {
2501
+ "epoch": 5.03,
2502
+ "learning_rate": 7.589873417721519e-06,
2503
+ "loss": 0.0261,
2504
+ "step": 10025
2505
+ },
2506
+ {
2507
+ "epoch": 5.03,
2508
+ "learning_rate": 7.583544303797469e-06,
2509
+ "loss": 0.0278,
2510
+ "step": 10050
2511
+ },
2512
+ {
2513
+ "epoch": 5.03,
2514
+ "learning_rate": 7.577215189873418e-06,
2515
+ "loss": 0.0248,
2516
+ "step": 10075
2517
+ },
2518
+ {
2519
+ "epoch": 5.03,
2520
+ "learning_rate": 7.5708860759493674e-06,
2521
+ "loss": 0.0263,
2522
+ "step": 10100
2523
+ },
2524
+ {
2525
+ "epoch": 5.03,
2526
+ "learning_rate": 7.564556962025317e-06,
2527
+ "loss": 0.0253,
2528
+ "step": 10125
2529
+ },
2530
+ {
2531
+ "epoch": 5.03,
2532
+ "learning_rate": 7.558227848101266e-06,
2533
+ "loss": 0.0248,
2534
+ "step": 10150
2535
+ },
2536
+ {
2537
+ "epoch": 5.04,
2538
+ "learning_rate": 7.551898734177216e-06,
2539
+ "loss": 0.0281,
2540
+ "step": 10175
2541
+ },
2542
+ {
2543
+ "epoch": 5.04,
2544
+ "learning_rate": 7.545569620253165e-06,
2545
+ "loss": 0.0276,
2546
+ "step": 10200
2547
+ },
2548
+ {
2549
+ "epoch": 5.04,
2550
+ "learning_rate": 7.539240506329114e-06,
2551
+ "loss": 0.0285,
2552
+ "step": 10225
2553
+ },
2554
+ {
2555
+ "epoch": 5.04,
2556
+ "learning_rate": 7.532911392405063e-06,
2557
+ "loss": 0.0265,
2558
+ "step": 10250
2559
+ },
2560
+ {
2561
+ "epoch": 5.04,
2562
+ "learning_rate": 7.526582278481013e-06,
2563
+ "loss": 0.0288,
2564
+ "step": 10275
2565
+ },
2566
+ {
2567
+ "epoch": 5.04,
2568
+ "learning_rate": 7.520253164556963e-06,
2569
+ "loss": 0.0247,
2570
+ "step": 10300
2571
+ },
2572
+ {
2573
+ "epoch": 5.04,
2574
+ "learning_rate": 7.513924050632912e-06,
2575
+ "loss": 0.0281,
2576
+ "step": 10325
2577
+ },
2578
+ {
2579
+ "epoch": 5.04,
2580
+ "learning_rate": 7.507594936708862e-06,
2581
+ "loss": 0.0262,
2582
+ "step": 10350
2583
+ },
2584
+ {
2585
+ "epoch": 5.04,
2586
+ "learning_rate": 7.501265822784811e-06,
2587
+ "loss": 0.0243,
2588
+ "step": 10375
2589
+ },
2590
+ {
2591
+ "epoch": 5.04,
2592
+ "learning_rate": 7.494936708860761e-06,
2593
+ "loss": 0.0331,
2594
+ "step": 10400
2595
+ },
2596
+ {
2597
+ "epoch": 5.04,
2598
+ "learning_rate": 7.48860759493671e-06,
2599
+ "loss": 0.0346,
2600
+ "step": 10425
2601
+ },
2602
+ {
2603
+ "epoch": 5.04,
2604
+ "learning_rate": 7.4822784810126594e-06,
2605
+ "loss": 0.0348,
2606
+ "step": 10450
2607
+ },
2608
+ {
2609
+ "epoch": 5.04,
2610
+ "learning_rate": 7.475949367088609e-06,
2611
+ "loss": 0.0272,
2612
+ "step": 10475
2613
+ },
2614
+ {
2615
+ "epoch": 5.04,
2616
+ "learning_rate": 7.4696202531645574e-06,
2617
+ "loss": 0.0326,
2618
+ "step": 10500
2619
+ },
2620
+ {
2621
+ "epoch": 6.0,
2622
+ "learning_rate": 7.463291139240507e-06,
2623
+ "loss": 0.033,
2624
+ "step": 10525
2625
+ },
2626
+ {
2627
+ "epoch": 6.0,
2628
+ "learning_rate": 7.456962025316456e-06,
2629
+ "loss": 0.0333,
2630
+ "step": 10550
2631
+ },
2632
+ {
2633
+ "epoch": 6.0,
2634
+ "learning_rate": 7.450632911392406e-06,
2635
+ "loss": 0.0371,
2636
+ "step": 10575
2637
+ },
2638
+ {
2639
+ "epoch": 6.0,
2640
+ "learning_rate": 7.444303797468355e-06,
2641
+ "loss": 0.0359,
2642
+ "step": 10600
2643
+ },
2644
+ {
2645
+ "epoch": 6.0,
2646
+ "learning_rate": 7.437974683544305e-06,
2647
+ "loss": 0.0438,
2648
+ "step": 10625
2649
+ },
2650
+ {
2651
+ "epoch": 6.0,
2652
+ "learning_rate": 7.431645569620254e-06,
2653
+ "loss": 0.0382,
2654
+ "step": 10650
2655
+ },
2656
+ {
2657
+ "epoch": 6.0,
2658
+ "learning_rate": 7.4253164556962034e-06,
2659
+ "loss": 0.0337,
2660
+ "step": 10675
2661
+ },
2662
+ {
2663
+ "epoch": 6.0,
2664
+ "learning_rate": 7.418987341772153e-06,
2665
+ "loss": 0.0422,
2666
+ "step": 10700
2667
+ },
2668
+ {
2669
+ "epoch": 6.01,
2670
+ "learning_rate": 7.4126582278481014e-06,
2671
+ "loss": 0.0467,
2672
+ "step": 10725
2673
+ },
2674
+ {
2675
+ "epoch": 6.01,
2676
+ "learning_rate": 7.406329113924051e-06,
2677
+ "loss": 0.0437,
2678
+ "step": 10750
2679
+ },
2680
+ {
2681
+ "epoch": 6.01,
2682
+ "learning_rate": 7.4e-06,
2683
+ "loss": 0.0328,
2684
+ "step": 10775
2685
+ },
2686
+ {
2687
+ "epoch": 6.01,
2688
+ "learning_rate": 7.39367088607595e-06,
2689
+ "loss": 0.0295,
2690
+ "step": 10800
2691
+ },
2692
+ {
2693
+ "epoch": 6.01,
2694
+ "learning_rate": 7.387341772151899e-06,
2695
+ "loss": 0.0431,
2696
+ "step": 10825
2697
+ },
2698
+ {
2699
+ "epoch": 6.01,
2700
+ "learning_rate": 7.381012658227849e-06,
2701
+ "loss": 0.0456,
2702
+ "step": 10850
2703
+ },
2704
+ {
2705
+ "epoch": 6.01,
2706
+ "learning_rate": 7.374683544303798e-06,
2707
+ "loss": 0.0267,
2708
+ "step": 10875
2709
+ },
2710
+ {
2711
+ "epoch": 6.01,
2712
+ "learning_rate": 7.3683544303797474e-06,
2713
+ "loss": 0.0334,
2714
+ "step": 10900
2715
+ },
2716
+ {
2717
+ "epoch": 6.01,
2718
+ "learning_rate": 7.362025316455697e-06,
2719
+ "loss": 0.0296,
2720
+ "step": 10925
2721
+ },
2722
+ {
2723
+ "epoch": 6.01,
2724
+ "learning_rate": 7.3556962025316454e-06,
2725
+ "loss": 0.0261,
2726
+ "step": 10950
2727
+ },
2728
+ {
2729
+ "epoch": 6.01,
2730
+ "learning_rate": 7.349367088607595e-06,
2731
+ "loss": 0.0299,
2732
+ "step": 10975
2733
+ },
2734
+ {
2735
+ "epoch": 6.01,
2736
+ "learning_rate": 7.343037974683544e-06,
2737
+ "loss": 0.0379,
2738
+ "step": 11000
2739
+ },
2740
+ {
2741
+ "epoch": 6.01,
2742
+ "eval_loss": 0.23971551656723022,
2743
+ "eval_runtime": 1435.6772,
2744
+ "eval_samples_per_second": 7.252,
2745
+ "eval_steps_per_second": 0.453,
2746
+ "eval_wer": 31.7712483129807,
2747
+ "step": 11000
2748
+ },
2749
+ {
2750
+ "epoch": 6.01,
2751
+ "learning_rate": 7.336708860759494e-06,
2752
+ "loss": 0.0457,
2753
+ "step": 11025
2754
+ },
2755
+ {
2756
+ "epoch": 6.01,
2757
+ "learning_rate": 7.330379746835443e-06,
2758
+ "loss": 0.0275,
2759
+ "step": 11050
2760
+ },
2761
+ {
2762
+ "epoch": 6.01,
2763
+ "learning_rate": 7.324050632911393e-06,
2764
+ "loss": 0.0254,
2765
+ "step": 11075
2766
+ },
2767
+ {
2768
+ "epoch": 6.01,
2769
+ "learning_rate": 7.317721518987342e-06,
2770
+ "loss": 0.035,
2771
+ "step": 11100
2772
+ },
2773
+ {
2774
+ "epoch": 6.02,
2775
+ "learning_rate": 7.311392405063292e-06,
2776
+ "loss": 0.0273,
2777
+ "step": 11125
2778
+ },
2779
+ {
2780
+ "epoch": 6.02,
2781
+ "learning_rate": 7.305063291139242e-06,
2782
+ "loss": 0.0227,
2783
+ "step": 11150
2784
+ },
2785
+ {
2786
+ "epoch": 6.02,
2787
+ "learning_rate": 7.298734177215191e-06,
2788
+ "loss": 0.0263,
2789
+ "step": 11175
2790
+ },
2791
+ {
2792
+ "epoch": 6.02,
2793
+ "learning_rate": 7.2924050632911406e-06,
2794
+ "loss": 0.0294,
2795
+ "step": 11200
2796
+ },
2797
+ {
2798
+ "epoch": 6.02,
2799
+ "learning_rate": 7.28607594936709e-06,
2800
+ "loss": 0.026,
2801
+ "step": 11225
2802
+ },
2803
+ {
2804
+ "epoch": 6.02,
2805
+ "learning_rate": 7.279746835443039e-06,
2806
+ "loss": 0.0251,
2807
+ "step": 11250
2808
+ },
2809
+ {
2810
+ "epoch": 6.02,
2811
+ "learning_rate": 7.273417721518988e-06,
2812
+ "loss": 0.024,
2813
+ "step": 11275
2814
+ },
2815
+ {
2816
+ "epoch": 6.02,
2817
+ "learning_rate": 7.2670886075949374e-06,
2818
+ "loss": 0.0344,
2819
+ "step": 11300
2820
+ },
2821
+ {
2822
+ "epoch": 6.02,
2823
+ "learning_rate": 7.260759493670887e-06,
2824
+ "loss": 0.0338,
2825
+ "step": 11325
2826
+ },
2827
+ {
2828
+ "epoch": 6.02,
2829
+ "learning_rate": 7.254430379746836e-06,
2830
+ "loss": 0.0266,
2831
+ "step": 11350
2832
+ },
2833
+ {
2834
+ "epoch": 6.02,
2835
+ "learning_rate": 7.248101265822786e-06,
2836
+ "loss": 0.0271,
2837
+ "step": 11375
2838
+ },
2839
+ {
2840
+ "epoch": 6.02,
2841
+ "learning_rate": 7.241772151898735e-06,
2842
+ "loss": 0.0338,
2843
+ "step": 11400
2844
+ },
2845
+ {
2846
+ "epoch": 6.02,
2847
+ "learning_rate": 7.2354430379746846e-06,
2848
+ "loss": 0.0321,
2849
+ "step": 11425
2850
+ },
2851
+ {
2852
+ "epoch": 6.02,
2853
+ "learning_rate": 7.229113924050634e-06,
2854
+ "loss": 0.0385,
2855
+ "step": 11450
2856
+ },
2857
+ {
2858
+ "epoch": 6.02,
2859
+ "learning_rate": 7.222784810126583e-06,
2860
+ "loss": 0.0284,
2861
+ "step": 11475
2862
+ },
2863
+ {
2864
+ "epoch": 6.02,
2865
+ "learning_rate": 7.216455696202532e-06,
2866
+ "loss": 0.0268,
2867
+ "step": 11500
2868
+ },
2869
+ {
2870
+ "epoch": 6.03,
2871
+ "learning_rate": 7.2101265822784814e-06,
2872
+ "loss": 0.0191,
2873
+ "step": 11525
2874
+ },
2875
+ {
2876
+ "epoch": 6.03,
2877
+ "learning_rate": 7.203797468354431e-06,
2878
+ "loss": 0.0243,
2879
+ "step": 11550
2880
+ },
2881
+ {
2882
+ "epoch": 6.03,
2883
+ "learning_rate": 7.19746835443038e-06,
2884
+ "loss": 0.0228,
2885
+ "step": 11575
2886
+ },
2887
+ {
2888
+ "epoch": 6.03,
2889
+ "learning_rate": 7.19113924050633e-06,
2890
+ "loss": 0.0187,
2891
+ "step": 11600
2892
+ },
2893
+ {
2894
+ "epoch": 6.03,
2895
+ "learning_rate": 7.184810126582279e-06,
2896
+ "loss": 0.0255,
2897
+ "step": 11625
2898
+ },
2899
+ {
2900
+ "epoch": 6.03,
2901
+ "learning_rate": 7.1784810126582286e-06,
2902
+ "loss": 0.0218,
2903
+ "step": 11650
2904
+ },
2905
+ {
2906
+ "epoch": 6.03,
2907
+ "learning_rate": 7.172151898734178e-06,
2908
+ "loss": 0.0249,
2909
+ "step": 11675
2910
+ },
2911
+ {
2912
+ "epoch": 6.03,
2913
+ "learning_rate": 7.165822784810127e-06,
2914
+ "loss": 0.0254,
2915
+ "step": 11700
2916
+ },
2917
+ {
2918
+ "epoch": 6.03,
2919
+ "learning_rate": 7.159493670886076e-06,
2920
+ "loss": 0.0323,
2921
+ "step": 11725
2922
+ },
2923
+ {
2924
+ "epoch": 6.03,
2925
+ "learning_rate": 7.1531645569620254e-06,
2926
+ "loss": 0.021,
2927
+ "step": 11750
2928
+ },
2929
+ {
2930
+ "epoch": 6.03,
2931
+ "learning_rate": 7.146835443037975e-06,
2932
+ "loss": 0.0242,
2933
+ "step": 11775
2934
+ },
2935
+ {
2936
+ "epoch": 6.03,
2937
+ "learning_rate": 7.140506329113924e-06,
2938
+ "loss": 0.0233,
2939
+ "step": 11800
2940
+ },
2941
+ {
2942
+ "epoch": 6.03,
2943
+ "learning_rate": 7.134177215189874e-06,
2944
+ "loss": 0.0237,
2945
+ "step": 11825
2946
+ },
2947
+ {
2948
+ "epoch": 6.03,
2949
+ "learning_rate": 7.127848101265823e-06,
2950
+ "loss": 0.0257,
2951
+ "step": 11850
2952
+ },
2953
+ {
2954
+ "epoch": 6.03,
2955
+ "learning_rate": 7.121518987341773e-06,
2956
+ "loss": 0.0256,
2957
+ "step": 11875
2958
+ },
2959
+ {
2960
+ "epoch": 6.03,
2961
+ "learning_rate": 7.115189873417722e-06,
2962
+ "loss": 0.0264,
2963
+ "step": 11900
2964
+ },
2965
+ {
2966
+ "epoch": 6.04,
2967
+ "learning_rate": 7.108860759493671e-06,
2968
+ "loss": 0.0254,
2969
+ "step": 11925
2970
+ },
2971
+ {
2972
+ "epoch": 6.04,
2973
+ "learning_rate": 7.102531645569622e-06,
2974
+ "loss": 0.025,
2975
+ "step": 11950
2976
+ },
2977
+ {
2978
+ "epoch": 6.04,
2979
+ "learning_rate": 7.096202531645571e-06,
2980
+ "loss": 0.0219,
2981
+ "step": 11975
2982
+ },
2983
+ {
2984
+ "epoch": 6.04,
2985
+ "learning_rate": 7.08987341772152e-06,
2986
+ "loss": 0.0241,
2987
+ "step": 12000
2988
+ },
2989
+ {
2990
+ "epoch": 6.04,
2991
+ "eval_loss": 0.26222535967826843,
2992
+ "eval_runtime": 1460.2321,
2993
+ "eval_samples_per_second": 7.13,
2994
+ "eval_steps_per_second": 0.446,
2995
+ "eval_wer": 32.87370526350022,
2996
+ "step": 12000
2997
+ }
2998
+ ],
2999
+ "max_steps": 40000,
3000
+ "num_train_epochs": 9223372036854775807,
3001
+ "total_flos": 5.53910817374208e+19,
3002
+ "trial_name": null,
3003
+ "trial_params": null
3004
+ }
checkpoint-12000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0289e1bf8fd1a33f7c3c9d3e15cc9bdd13d10e9774e0ada934c0b9d71501a993
3
+ size 4091
checkpoint-13000/config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/whisper-small",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": true,
6
+ "architectures": [
7
+ "WhisperForConditionalGeneration"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 768,
17
+ "decoder_attention_heads": 12,
18
+ "decoder_ffn_dim": 3072,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 12,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 12,
24
+ "encoder_ffn_dim": 3072,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 12,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": null,
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "mask_feature_length": 10,
32
+ "mask_feature_min_masks": 0,
33
+ "mask_feature_prob": 0.05,
34
+ "mask_time_length": 10,
35
+ "mask_time_min_masks": 2,
36
+ "mask_time_prob": 0.05,
37
+ "max_length": 448,
38
+ "max_source_positions": 1500,
39
+ "max_target_positions": 448,
40
+ "model_type": "whisper",
41
+ "num_hidden_layers": 12,
42
+ "num_mel_bins": 80,
43
+ "pad_token_id": 50257,
44
+ "scale_embedding": false,
45
+ "suppress_tokens": [],
46
+ "torch_dtype": "float32",
47
+ "transformers_version": "4.30.0.dev0",
48
+ "use_cache": true,
49
+ "use_weighted_layer_sum": false,
50
+ "vocab_size": 51865
51
+ }
checkpoint-13000/generation_config.json ADDED
@@ -0,0 +1,221 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "begin_suppress_tokens": [
3
+ 220,
4
+ 50257
5
+ ],
6
+ "bos_token_id": 50257,
7
+ "decoder_start_token_id": 50258,
8
+ "eos_token_id": 50257,
9
+ "forced_decoder_ids": [
10
+ [
11
+ 1,
12
+ null
13
+ ],
14
+ [
15
+ 2,
16
+ 50359
17
+ ]
18
+ ],
19
+ "is_multilingual": true,
20
+ "lang_to_id": {
21
+ "<|af|>": 50327,
22
+ "<|am|>": 50334,
23
+ "<|ar|>": 50272,
24
+ "<|as|>": 50350,
25
+ "<|az|>": 50304,
26
+ "<|ba|>": 50355,
27
+ "<|be|>": 50330,
28
+ "<|bg|>": 50292,
29
+ "<|bn|>": 50302,
30
+ "<|bo|>": 50347,
31
+ "<|br|>": 50309,
32
+ "<|bs|>": 50315,
33
+ "<|ca|>": 50270,
34
+ "<|cs|>": 50283,
35
+ "<|cy|>": 50297,
36
+ "<|da|>": 50285,
37
+ "<|de|>": 50261,
38
+ "<|el|>": 50281,
39
+ "<|en|>": 50259,
40
+ "<|es|>": 50262,
41
+ "<|et|>": 50307,
42
+ "<|eu|>": 50310,
43
+ "<|fa|>": 50300,
44
+ "<|fi|>": 50277,
45
+ "<|fo|>": 50338,
46
+ "<|fr|>": 50265,
47
+ "<|gl|>": 50319,
48
+ "<|gu|>": 50333,
49
+ "<|haw|>": 50352,
50
+ "<|ha|>": 50354,
51
+ "<|he|>": 50279,
52
+ "<|hi|>": 50276,
53
+ "<|hr|>": 50291,
54
+ "<|ht|>": 50339,
55
+ "<|hu|>": 50286,
56
+ "<|hy|>": 50312,
57
+ "<|id|>": 50275,
58
+ "<|is|>": 50311,
59
+ "<|it|>": 50274,
60
+ "<|ja|>": 50266,
61
+ "<|jw|>": 50356,
62
+ "<|ka|>": 50329,
63
+ "<|kk|>": 50316,
64
+ "<|km|>": 50323,
65
+ "<|kn|>": 50306,
66
+ "<|ko|>": 50264,
67
+ "<|la|>": 50294,
68
+ "<|lb|>": 50345,
69
+ "<|ln|>": 50353,
70
+ "<|lo|>": 50336,
71
+ "<|lt|>": 50293,
72
+ "<|lv|>": 50301,
73
+ "<|mg|>": 50349,
74
+ "<|mi|>": 50295,
75
+ "<|mk|>": 50308,
76
+ "<|ml|>": 50296,
77
+ "<|mn|>": 50314,
78
+ "<|mr|>": 50320,
79
+ "<|ms|>": 50282,
80
+ "<|mt|>": 50343,
81
+ "<|my|>": 50346,
82
+ "<|ne|>": 50313,
83
+ "<|nl|>": 50271,
84
+ "<|nn|>": 50342,
85
+ "<|no|>": 50288,
86
+ "<|oc|>": 50328,
87
+ "<|pa|>": 50321,
88
+ "<|pl|>": 50269,
89
+ "<|ps|>": 50340,
90
+ "<|pt|>": 50267,
91
+ "<|ro|>": 50284,
92
+ "<|ru|>": 50263,
93
+ "<|sa|>": 50344,
94
+ "<|sd|>": 50332,
95
+ "<|si|>": 50322,
96
+ "<|sk|>": 50298,
97
+ "<|sl|>": 50305,
98
+ "<|sn|>": 50324,
99
+ "<|so|>": 50326,
100
+ "<|sq|>": 50317,
101
+ "<|sr|>": 50303,
102
+ "<|su|>": 50357,
103
+ "<|sv|>": 50273,
104
+ "<|sw|>": 50318,
105
+ "<|ta|>": 50287,
106
+ "<|te|>": 50299,
107
+ "<|tg|>": 50331,
108
+ "<|th|>": 50289,
109
+ "<|tk|>": 50341,
110
+ "<|tl|>": 50348,
111
+ "<|tr|>": 50268,
112
+ "<|tt|>": 50351,
113
+ "<|uk|>": 50280,
114
+ "<|ur|>": 50290,
115
+ "<|uz|>": 50337,
116
+ "<|vi|>": 50278,
117
+ "<|yi|>": 50335,
118
+ "<|yo|>": 50325,
119
+ "<|zh|>": 50260
120
+ },
121
+ "max_initial_timestamp_index": 1,
122
+ "max_length": 448,
123
+ "no_timestamps_token_id": 50363,
124
+ "pad_token_id": 50257,
125
+ "return_timestamps": false,
126
+ "suppress_tokens": [
127
+ 1,
128
+ 2,
129
+ 7,
130
+ 8,
131
+ 9,
132
+ 10,
133
+ 14,
134
+ 25,
135
+ 26,
136
+ 27,
137
+ 28,
138
+ 29,
139
+ 31,
140
+ 58,
141
+ 59,
142
+ 60,
143
+ 61,
144
+ 62,
145
+ 63,
146
+ 90,
147
+ 91,
148
+ 92,
149
+ 93,
150
+ 359,
151
+ 503,
152
+ 522,
153
+ 542,
154
+ 873,
155
+ 893,
156
+ 902,
157
+ 918,
158
+ 922,
159
+ 931,
160
+ 1350,
161
+ 1853,
162
+ 1982,
163
+ 2460,
164
+ 2627,
165
+ 3246,
166
+ 3253,
167
+ 3268,
168
+ 3536,
169
+ 3846,
170
+ 3961,
171
+ 4183,
172
+ 4667,
173
+ 6585,
174
+ 6647,
175
+ 7273,
176
+ 9061,
177
+ 9383,
178
+ 10428,
179
+ 10929,
180
+ 11938,
181
+ 12033,
182
+ 12331,
183
+ 12562,
184
+ 13793,
185
+ 14157,
186
+ 14635,
187
+ 15265,
188
+ 15618,
189
+ 16553,
190
+ 16604,
191
+ 18362,
192
+ 18956,
193
+ 20075,
194
+ 21675,
195
+ 22520,
196
+ 26130,
197
+ 26161,
198
+ 26435,
199
+ 28279,
200
+ 29464,
201
+ 31650,
202
+ 32302,
203
+ 32470,
204
+ 36865,
205
+ 42863,
206
+ 47425,
207
+ 49870,
208
+ 50254,
209
+ 50258,
210
+ 50358,
211
+ 50359,
212
+ 50360,
213
+ 50361,
214
+ 50362
215
+ ],
216
+ "task_to_id": {
217
+ "transcribe": 50359,
218
+ "translate": 50358
219
+ },
220
+ "transformers_version": "4.30.0.dev0"
221
+ }
checkpoint-13000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca0307a5ac957394d12917783450a1a36dfb78a489e3c19c508da9731c864e51
3
+ size 1934161093
checkpoint-13000/preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
checkpoint-13000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e67a8b1d3e0f7bf51629723dd88a387b19d607a0669ee2acc20a9c14be05b85
3
+ size 967102601
checkpoint-13000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15cd7a103721023beab2179ac18954faf15caa4e1cb6377a37ce555a07f66e53
3
+ size 14575
checkpoint-13000/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ba0bdea6f52c5e2b58651f8fe7c517a709a88dc947af37831e8d1ee42deb4b5f
3
+ size 557
checkpoint-13000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:affdb4a0764216c9f0dc59a9078d6c986f9bd648222f67a55a33342eaed0c1e7
3
+ size 627
checkpoint-13000/trainer_state.json ADDED
@@ -0,0 +1,3253 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 27.36142051090262,
3
+ "best_model_checkpoint": "./whisper-small-fa-aug/checkpoint-13000",
4
+ "epoch": 7.018575,
5
+ "global_step": 13000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 4.2000000000000006e-07,
13
+ "loss": 3.8,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 9.200000000000001e-07,
19
+ "loss": 3.0497,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.0,
24
+ "learning_rate": 1.42e-06,
25
+ "loss": 2.0102,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.0,
30
+ "learning_rate": 1.9200000000000003e-06,
31
+ "loss": 1.5948,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.0,
36
+ "learning_rate": 2.42e-06,
37
+ "loss": 1.4094,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.0,
42
+ "learning_rate": 2.92e-06,
43
+ "loss": 1.2611,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.0,
48
+ "learning_rate": 3.4200000000000007e-06,
49
+ "loss": 1.1366,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.01,
54
+ "learning_rate": 3.920000000000001e-06,
55
+ "loss": 1.0817,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.01,
60
+ "learning_rate": 4.42e-06,
61
+ "loss": 1.0324,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.01,
66
+ "learning_rate": 4.92e-06,
67
+ "loss": 0.9176,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.01,
72
+ "learning_rate": 5.420000000000001e-06,
73
+ "loss": 0.7975,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.01,
78
+ "learning_rate": 5.92e-06,
79
+ "loss": 0.7079,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.01,
84
+ "learning_rate": 6.42e-06,
85
+ "loss": 0.5832,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.01,
90
+ "learning_rate": 6.92e-06,
91
+ "loss": 0.4469,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.01,
96
+ "learning_rate": 7.420000000000001e-06,
97
+ "loss": 0.3913,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.01,
102
+ "learning_rate": 7.92e-06,
103
+ "loss": 0.4189,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.01,
108
+ "learning_rate": 8.42e-06,
109
+ "loss": 0.3756,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.01,
114
+ "learning_rate": 8.920000000000001e-06,
115
+ "loss": 0.3876,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.01,
120
+ "learning_rate": 9.42e-06,
121
+ "loss": 0.3802,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.01,
126
+ "learning_rate": 9.920000000000002e-06,
127
+ "loss": 0.4061,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.01,
132
+ "learning_rate": 9.994683544303798e-06,
133
+ "loss": 0.3885,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.01,
138
+ "learning_rate": 9.988354430379748e-06,
139
+ "loss": 0.3449,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.01,
144
+ "learning_rate": 9.982025316455697e-06,
145
+ "loss": 0.3292,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.01,
150
+ "learning_rate": 9.975696202531647e-06,
151
+ "loss": 0.3439,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.02,
156
+ "learning_rate": 9.969367088607596e-06,
157
+ "loss": 0.3031,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.02,
162
+ "learning_rate": 9.963037974683545e-06,
163
+ "loss": 0.3159,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.02,
168
+ "learning_rate": 9.956708860759495e-06,
169
+ "loss": 0.3304,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.02,
174
+ "learning_rate": 9.950379746835444e-06,
175
+ "loss": 0.2918,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.02,
180
+ "learning_rate": 9.944050632911392e-06,
181
+ "loss": 0.2785,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.02,
186
+ "learning_rate": 9.937721518987341e-06,
187
+ "loss": 0.2618,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.02,
192
+ "learning_rate": 9.931392405063291e-06,
193
+ "loss": 0.2862,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.02,
198
+ "learning_rate": 9.92506329113924e-06,
199
+ "loss": 0.2962,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.02,
204
+ "learning_rate": 9.91873417721519e-06,
205
+ "loss": 0.2935,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.02,
210
+ "learning_rate": 9.912405063291141e-06,
211
+ "loss": 0.2827,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.02,
216
+ "learning_rate": 9.90607594936709e-06,
217
+ "loss": 0.2702,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 0.02,
222
+ "learning_rate": 9.89974683544304e-06,
223
+ "loss": 0.2987,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 0.02,
228
+ "learning_rate": 9.89341772151899e-06,
229
+ "loss": 0.3067,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 0.02,
234
+ "learning_rate": 9.887088607594939e-06,
235
+ "loss": 0.2962,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 0.02,
240
+ "learning_rate": 9.880759493670888e-06,
241
+ "loss": 0.2807,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 0.03,
246
+ "learning_rate": 9.874430379746836e-06,
247
+ "loss": 0.2533,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 0.03,
252
+ "eval_loss": 0.3274388611316681,
253
+ "eval_runtime": 1481.9106,
254
+ "eval_samples_per_second": 7.025,
255
+ "eval_steps_per_second": 0.439,
256
+ "eval_wer": 40.657571667832,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 0.03,
261
+ "learning_rate": 9.868101265822785e-06,
262
+ "loss": 0.21,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 0.03,
267
+ "learning_rate": 9.861772151898735e-06,
268
+ "loss": 0.2463,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 0.03,
273
+ "learning_rate": 9.855443037974684e-06,
274
+ "loss": 0.2688,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 0.03,
279
+ "learning_rate": 9.849113924050633e-06,
280
+ "loss": 0.2293,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 0.03,
285
+ "learning_rate": 9.842784810126583e-06,
286
+ "loss": 0.2283,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 0.03,
291
+ "learning_rate": 9.836455696202532e-06,
292
+ "loss": 0.228,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 0.03,
297
+ "learning_rate": 9.830126582278482e-06,
298
+ "loss": 0.1919,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 0.03,
303
+ "learning_rate": 9.823797468354431e-06,
304
+ "loss": 0.2173,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 0.03,
309
+ "learning_rate": 9.81746835443038e-06,
310
+ "loss": 0.2066,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 0.03,
315
+ "learning_rate": 9.81113924050633e-06,
316
+ "loss": 0.1826,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 0.03,
321
+ "learning_rate": 9.80481012658228e-06,
322
+ "loss": 0.1855,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 0.03,
327
+ "learning_rate": 9.798481012658229e-06,
328
+ "loss": 0.1702,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 0.03,
333
+ "learning_rate": 9.792151898734178e-06,
334
+ "loss": 0.1714,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 0.03,
339
+ "learning_rate": 9.785822784810128e-06,
340
+ "loss": 0.1686,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 0.03,
345
+ "learning_rate": 9.779493670886077e-06,
346
+ "loss": 0.1762,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 0.04,
351
+ "learning_rate": 9.773164556962027e-06,
352
+ "loss": 0.1722,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 0.04,
357
+ "learning_rate": 9.766835443037976e-06,
358
+ "loss": 0.2051,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 0.04,
363
+ "learning_rate": 9.760506329113924e-06,
364
+ "loss": 0.2071,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 0.04,
369
+ "learning_rate": 9.754177215189873e-06,
370
+ "loss": 0.1664,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 0.04,
375
+ "learning_rate": 9.747848101265823e-06,
376
+ "loss": 0.1871,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 0.04,
381
+ "learning_rate": 9.741518987341772e-06,
382
+ "loss": 0.1893,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 0.04,
387
+ "learning_rate": 9.735189873417721e-06,
388
+ "loss": 0.1756,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 0.04,
393
+ "learning_rate": 9.728860759493671e-06,
394
+ "loss": 0.1687,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 0.04,
399
+ "learning_rate": 9.72253164556962e-06,
400
+ "loss": 0.1685,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 0.04,
405
+ "learning_rate": 9.71620253164557e-06,
406
+ "loss": 0.1805,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 0.04,
411
+ "learning_rate": 9.70987341772152e-06,
412
+ "loss": 0.1868,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 0.04,
417
+ "learning_rate": 9.70354430379747e-06,
418
+ "loss": 0.212,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 0.04,
423
+ "learning_rate": 9.69721518987342e-06,
424
+ "loss": 0.2019,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 0.04,
429
+ "learning_rate": 9.69088607594937e-06,
430
+ "loss": 0.2076,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 0.04,
435
+ "learning_rate": 9.684556962025317e-06,
436
+ "loss": 0.1971,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 1.0,
441
+ "learning_rate": 9.678227848101266e-06,
442
+ "loss": 0.2424,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 1.0,
447
+ "learning_rate": 9.671898734177216e-06,
448
+ "loss": 0.2346,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 1.0,
453
+ "learning_rate": 9.665569620253165e-06,
454
+ "loss": 0.2436,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 1.0,
459
+ "learning_rate": 9.659240506329115e-06,
460
+ "loss": 0.2613,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 1.0,
465
+ "learning_rate": 9.652911392405064e-06,
466
+ "loss": 0.2781,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 1.0,
471
+ "learning_rate": 9.646582278481013e-06,
472
+ "loss": 0.2465,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 1.0,
477
+ "learning_rate": 9.640253164556963e-06,
478
+ "loss": 0.2386,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 1.0,
483
+ "learning_rate": 9.633924050632912e-06,
484
+ "loss": 0.2298,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 1.01,
489
+ "learning_rate": 9.627594936708862e-06,
490
+ "loss": 0.264,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 1.01,
495
+ "learning_rate": 9.621265822784811e-06,
496
+ "loss": 0.2077,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 1.01,
501
+ "eval_loss": 0.25397753715515137,
502
+ "eval_runtime": 1477.9319,
503
+ "eval_samples_per_second": 7.044,
504
+ "eval_steps_per_second": 0.44,
505
+ "eval_wer": 35.12252231743606,
506
+ "step": 2000
507
+ },
508
+ {
509
+ "epoch": 1.01,
510
+ "learning_rate": 9.61493670886076e-06,
511
+ "loss": 0.192,
512
+ "step": 2025
513
+ },
514
+ {
515
+ "epoch": 1.01,
516
+ "learning_rate": 9.60860759493671e-06,
517
+ "loss": 0.1903,
518
+ "step": 2050
519
+ },
520
+ {
521
+ "epoch": 1.01,
522
+ "learning_rate": 9.60227848101266e-06,
523
+ "loss": 0.2135,
524
+ "step": 2075
525
+ },
526
+ {
527
+ "epoch": 1.01,
528
+ "learning_rate": 9.595949367088609e-06,
529
+ "loss": 0.1917,
530
+ "step": 2100
531
+ },
532
+ {
533
+ "epoch": 1.01,
534
+ "learning_rate": 9.589620253164558e-06,
535
+ "loss": 0.1579,
536
+ "step": 2125
537
+ },
538
+ {
539
+ "epoch": 1.01,
540
+ "learning_rate": 9.583291139240508e-06,
541
+ "loss": 0.1779,
542
+ "step": 2150
543
+ },
544
+ {
545
+ "epoch": 1.01,
546
+ "learning_rate": 9.576962025316457e-06,
547
+ "loss": 0.1504,
548
+ "step": 2175
549
+ },
550
+ {
551
+ "epoch": 1.01,
552
+ "learning_rate": 9.570632911392405e-06,
553
+ "loss": 0.1574,
554
+ "step": 2200
555
+ },
556
+ {
557
+ "epoch": 1.01,
558
+ "learning_rate": 9.564303797468354e-06,
559
+ "loss": 0.1757,
560
+ "step": 2225
561
+ },
562
+ {
563
+ "epoch": 1.01,
564
+ "learning_rate": 9.557974683544304e-06,
565
+ "loss": 0.1996,
566
+ "step": 2250
567
+ },
568
+ {
569
+ "epoch": 1.01,
570
+ "learning_rate": 9.551645569620253e-06,
571
+ "loss": 0.1716,
572
+ "step": 2275
573
+ },
574
+ {
575
+ "epoch": 1.01,
576
+ "learning_rate": 9.545316455696203e-06,
577
+ "loss": 0.1509,
578
+ "step": 2300
579
+ },
580
+ {
581
+ "epoch": 1.01,
582
+ "learning_rate": 9.538987341772152e-06,
583
+ "loss": 0.1402,
584
+ "step": 2325
585
+ },
586
+ {
587
+ "epoch": 1.01,
588
+ "learning_rate": 9.532658227848101e-06,
589
+ "loss": 0.1612,
590
+ "step": 2350
591
+ },
592
+ {
593
+ "epoch": 1.02,
594
+ "learning_rate": 9.526329113924051e-06,
595
+ "loss": 0.1357,
596
+ "step": 2375
597
+ },
598
+ {
599
+ "epoch": 1.02,
600
+ "learning_rate": 9.52e-06,
601
+ "loss": 0.1426,
602
+ "step": 2400
603
+ },
604
+ {
605
+ "epoch": 1.02,
606
+ "learning_rate": 9.51367088607595e-06,
607
+ "loss": 0.1644,
608
+ "step": 2425
609
+ },
610
+ {
611
+ "epoch": 1.02,
612
+ "learning_rate": 9.5073417721519e-06,
613
+ "loss": 0.1196,
614
+ "step": 2450
615
+ },
616
+ {
617
+ "epoch": 1.02,
618
+ "learning_rate": 9.501012658227849e-06,
619
+ "loss": 0.1332,
620
+ "step": 2475
621
+ },
622
+ {
623
+ "epoch": 1.02,
624
+ "learning_rate": 9.494683544303798e-06,
625
+ "loss": 0.1096,
626
+ "step": 2500
627
+ },
628
+ {
629
+ "epoch": 1.02,
630
+ "learning_rate": 9.488354430379747e-06,
631
+ "loss": 0.1263,
632
+ "step": 2525
633
+ },
634
+ {
635
+ "epoch": 1.02,
636
+ "learning_rate": 9.482025316455697e-06,
637
+ "loss": 0.1467,
638
+ "step": 2550
639
+ },
640
+ {
641
+ "epoch": 1.02,
642
+ "learning_rate": 9.475696202531646e-06,
643
+ "loss": 0.1372,
644
+ "step": 2575
645
+ },
646
+ {
647
+ "epoch": 1.02,
648
+ "learning_rate": 9.469367088607596e-06,
649
+ "loss": 0.1256,
650
+ "step": 2600
651
+ },
652
+ {
653
+ "epoch": 1.02,
654
+ "learning_rate": 9.463037974683545e-06,
655
+ "loss": 0.1319,
656
+ "step": 2625
657
+ },
658
+ {
659
+ "epoch": 1.02,
660
+ "learning_rate": 9.456708860759495e-06,
661
+ "loss": 0.1385,
662
+ "step": 2650
663
+ },
664
+ {
665
+ "epoch": 1.02,
666
+ "learning_rate": 9.450379746835444e-06,
667
+ "loss": 0.147,
668
+ "step": 2675
669
+ },
670
+ {
671
+ "epoch": 1.02,
672
+ "learning_rate": 9.444050632911393e-06,
673
+ "loss": 0.1527,
674
+ "step": 2700
675
+ },
676
+ {
677
+ "epoch": 1.02,
678
+ "learning_rate": 9.437721518987343e-06,
679
+ "loss": 0.1357,
680
+ "step": 2725
681
+ },
682
+ {
683
+ "epoch": 1.02,
684
+ "learning_rate": 9.431392405063292e-06,
685
+ "loss": 0.122,
686
+ "step": 2750
687
+ },
688
+ {
689
+ "epoch": 1.03,
690
+ "learning_rate": 9.425063291139242e-06,
691
+ "loss": 0.0918,
692
+ "step": 2775
693
+ },
694
+ {
695
+ "epoch": 1.03,
696
+ "learning_rate": 9.418734177215191e-06,
697
+ "loss": 0.1218,
698
+ "step": 2800
699
+ },
700
+ {
701
+ "epoch": 1.03,
702
+ "learning_rate": 9.41240506329114e-06,
703
+ "loss": 0.139,
704
+ "step": 2825
705
+ },
706
+ {
707
+ "epoch": 1.03,
708
+ "learning_rate": 9.40607594936709e-06,
709
+ "loss": 0.1205,
710
+ "step": 2850
711
+ },
712
+ {
713
+ "epoch": 1.03,
714
+ "learning_rate": 9.39974683544304e-06,
715
+ "loss": 0.1158,
716
+ "step": 2875
717
+ },
718
+ {
719
+ "epoch": 1.03,
720
+ "learning_rate": 9.393417721518989e-06,
721
+ "loss": 0.1156,
722
+ "step": 2900
723
+ },
724
+ {
725
+ "epoch": 1.03,
726
+ "learning_rate": 9.387088607594937e-06,
727
+ "loss": 0.1044,
728
+ "step": 2925
729
+ },
730
+ {
731
+ "epoch": 1.03,
732
+ "learning_rate": 9.380759493670886e-06,
733
+ "loss": 0.112,
734
+ "step": 2950
735
+ },
736
+ {
737
+ "epoch": 1.03,
738
+ "learning_rate": 9.374430379746835e-06,
739
+ "loss": 0.1023,
740
+ "step": 2975
741
+ },
742
+ {
743
+ "epoch": 1.03,
744
+ "learning_rate": 9.368101265822785e-06,
745
+ "loss": 0.0966,
746
+ "step": 3000
747
+ },
748
+ {
749
+ "epoch": 1.03,
750
+ "eval_loss": 0.2638459801673889,
751
+ "eval_runtime": 1500.936,
752
+ "eval_samples_per_second": 6.936,
753
+ "eval_steps_per_second": 0.434,
754
+ "eval_wer": 37.93069806013106,
755
+ "step": 3000
756
+ },
757
+ {
758
+ "epoch": 1.03,
759
+ "learning_rate": 9.361772151898734e-06,
760
+ "loss": 0.0942,
761
+ "step": 3025
762
+ },
763
+ {
764
+ "epoch": 1.03,
765
+ "learning_rate": 9.355443037974684e-06,
766
+ "loss": 0.0826,
767
+ "step": 3050
768
+ },
769
+ {
770
+ "epoch": 1.03,
771
+ "learning_rate": 9.349113924050633e-06,
772
+ "loss": 0.086,
773
+ "step": 3075
774
+ },
775
+ {
776
+ "epoch": 1.03,
777
+ "learning_rate": 9.342784810126583e-06,
778
+ "loss": 0.0956,
779
+ "step": 3100
780
+ },
781
+ {
782
+ "epoch": 1.03,
783
+ "learning_rate": 9.336455696202532e-06,
784
+ "loss": 0.0935,
785
+ "step": 3125
786
+ },
787
+ {
788
+ "epoch": 1.03,
789
+ "learning_rate": 9.330126582278481e-06,
790
+ "loss": 0.0923,
791
+ "step": 3150
792
+ },
793
+ {
794
+ "epoch": 1.04,
795
+ "learning_rate": 9.323797468354431e-06,
796
+ "loss": 0.1163,
797
+ "step": 3175
798
+ },
799
+ {
800
+ "epoch": 1.04,
801
+ "learning_rate": 9.31746835443038e-06,
802
+ "loss": 0.1099,
803
+ "step": 3200
804
+ },
805
+ {
806
+ "epoch": 1.04,
807
+ "learning_rate": 9.31113924050633e-06,
808
+ "loss": 0.0879,
809
+ "step": 3225
810
+ },
811
+ {
812
+ "epoch": 1.04,
813
+ "learning_rate": 9.30481012658228e-06,
814
+ "loss": 0.0994,
815
+ "step": 3250
816
+ },
817
+ {
818
+ "epoch": 1.04,
819
+ "learning_rate": 9.298481012658229e-06,
820
+ "loss": 0.1069,
821
+ "step": 3275
822
+ },
823
+ {
824
+ "epoch": 1.04,
825
+ "learning_rate": 9.292151898734178e-06,
826
+ "loss": 0.0957,
827
+ "step": 3300
828
+ },
829
+ {
830
+ "epoch": 1.04,
831
+ "learning_rate": 9.285822784810127e-06,
832
+ "loss": 0.0866,
833
+ "step": 3325
834
+ },
835
+ {
836
+ "epoch": 1.04,
837
+ "learning_rate": 9.279493670886077e-06,
838
+ "loss": 0.0965,
839
+ "step": 3350
840
+ },
841
+ {
842
+ "epoch": 1.04,
843
+ "learning_rate": 9.273164556962026e-06,
844
+ "loss": 0.0952,
845
+ "step": 3375
846
+ },
847
+ {
848
+ "epoch": 1.04,
849
+ "learning_rate": 9.266835443037976e-06,
850
+ "loss": 0.1178,
851
+ "step": 3400
852
+ },
853
+ {
854
+ "epoch": 1.04,
855
+ "learning_rate": 9.260506329113925e-06,
856
+ "loss": 0.1154,
857
+ "step": 3425
858
+ },
859
+ {
860
+ "epoch": 1.04,
861
+ "learning_rate": 9.254177215189875e-06,
862
+ "loss": 0.1075,
863
+ "step": 3450
864
+ },
865
+ {
866
+ "epoch": 1.04,
867
+ "learning_rate": 9.247848101265824e-06,
868
+ "loss": 0.1161,
869
+ "step": 3475
870
+ },
871
+ {
872
+ "epoch": 1.04,
873
+ "learning_rate": 9.241518987341773e-06,
874
+ "loss": 0.1153,
875
+ "step": 3500
876
+ },
877
+ {
878
+ "epoch": 2.0,
879
+ "learning_rate": 9.235189873417723e-06,
880
+ "loss": 0.1343,
881
+ "step": 3525
882
+ },
883
+ {
884
+ "epoch": 2.0,
885
+ "learning_rate": 9.228860759493672e-06,
886
+ "loss": 0.1374,
887
+ "step": 3550
888
+ },
889
+ {
890
+ "epoch": 2.0,
891
+ "learning_rate": 9.222531645569622e-06,
892
+ "loss": 0.1436,
893
+ "step": 3575
894
+ },
895
+ {
896
+ "epoch": 2.0,
897
+ "learning_rate": 9.216202531645571e-06,
898
+ "loss": 0.1561,
899
+ "step": 3600
900
+ },
901
+ {
902
+ "epoch": 2.0,
903
+ "learning_rate": 9.20987341772152e-06,
904
+ "loss": 0.1709,
905
+ "step": 3625
906
+ },
907
+ {
908
+ "epoch": 2.0,
909
+ "learning_rate": 9.20354430379747e-06,
910
+ "loss": 0.142,
911
+ "step": 3650
912
+ },
913
+ {
914
+ "epoch": 2.0,
915
+ "learning_rate": 9.197215189873418e-06,
916
+ "loss": 0.1389,
917
+ "step": 3675
918
+ },
919
+ {
920
+ "epoch": 2.0,
921
+ "learning_rate": 9.190886075949367e-06,
922
+ "loss": 0.1398,
923
+ "step": 3700
924
+ },
925
+ {
926
+ "epoch": 2.01,
927
+ "learning_rate": 9.184556962025317e-06,
928
+ "loss": 0.1526,
929
+ "step": 3725
930
+ },
931
+ {
932
+ "epoch": 2.01,
933
+ "learning_rate": 9.178227848101266e-06,
934
+ "loss": 0.1273,
935
+ "step": 3750
936
+ },
937
+ {
938
+ "epoch": 2.01,
939
+ "learning_rate": 9.171898734177215e-06,
940
+ "loss": 0.1119,
941
+ "step": 3775
942
+ },
943
+ {
944
+ "epoch": 2.01,
945
+ "learning_rate": 9.165569620253165e-06,
946
+ "loss": 0.1088,
947
+ "step": 3800
948
+ },
949
+ {
950
+ "epoch": 2.01,
951
+ "learning_rate": 9.159493670886076e-06,
952
+ "loss": 0.1362,
953
+ "step": 3825
954
+ },
955
+ {
956
+ "epoch": 2.01,
957
+ "learning_rate": 9.153164556962026e-06,
958
+ "loss": 0.1247,
959
+ "step": 3850
960
+ },
961
+ {
962
+ "epoch": 2.01,
963
+ "learning_rate": 9.146835443037975e-06,
964
+ "loss": 0.0974,
965
+ "step": 3875
966
+ },
967
+ {
968
+ "epoch": 2.01,
969
+ "learning_rate": 9.140506329113925e-06,
970
+ "loss": 0.1218,
971
+ "step": 3900
972
+ },
973
+ {
974
+ "epoch": 2.01,
975
+ "learning_rate": 9.134177215189874e-06,
976
+ "loss": 0.0846,
977
+ "step": 3925
978
+ },
979
+ {
980
+ "epoch": 2.01,
981
+ "learning_rate": 9.127848101265823e-06,
982
+ "loss": 0.0964,
983
+ "step": 3950
984
+ },
985
+ {
986
+ "epoch": 2.01,
987
+ "learning_rate": 9.121518987341773e-06,
988
+ "loss": 0.1166,
989
+ "step": 3975
990
+ },
991
+ {
992
+ "epoch": 2.01,
993
+ "learning_rate": 9.115189873417722e-06,
994
+ "loss": 0.1212,
995
+ "step": 4000
996
+ },
997
+ {
998
+ "epoch": 2.01,
999
+ "eval_loss": 0.22517943382263184,
1000
+ "eval_runtime": 1537.556,
1001
+ "eval_samples_per_second": 6.771,
1002
+ "eval_steps_per_second": 0.423,
1003
+ "eval_wer": 69.04990325045935,
1004
+ "step": 4000
1005
+ },
1006
+ {
1007
+ "epoch": 2.01,
1008
+ "learning_rate": 9.108860759493672e-06,
1009
+ "loss": 0.1076,
1010
+ "step": 4025
1011
+ },
1012
+ {
1013
+ "epoch": 2.01,
1014
+ "learning_rate": 9.102531645569621e-06,
1015
+ "loss": 0.086,
1016
+ "step": 4050
1017
+ },
1018
+ {
1019
+ "epoch": 2.01,
1020
+ "learning_rate": 9.09620253164557e-06,
1021
+ "loss": 0.0773,
1022
+ "step": 4075
1023
+ },
1024
+ {
1025
+ "epoch": 2.01,
1026
+ "learning_rate": 9.08987341772152e-06,
1027
+ "loss": 0.098,
1028
+ "step": 4100
1029
+ },
1030
+ {
1031
+ "epoch": 2.02,
1032
+ "learning_rate": 9.08354430379747e-06,
1033
+ "loss": 0.0747,
1034
+ "step": 4125
1035
+ },
1036
+ {
1037
+ "epoch": 2.02,
1038
+ "learning_rate": 9.077215189873419e-06,
1039
+ "loss": 0.0895,
1040
+ "step": 4150
1041
+ },
1042
+ {
1043
+ "epoch": 2.02,
1044
+ "learning_rate": 9.070886075949368e-06,
1045
+ "loss": 0.0957,
1046
+ "step": 4175
1047
+ },
1048
+ {
1049
+ "epoch": 2.02,
1050
+ "learning_rate": 9.064556962025318e-06,
1051
+ "loss": 0.0789,
1052
+ "step": 4200
1053
+ },
1054
+ {
1055
+ "epoch": 2.02,
1056
+ "learning_rate": 9.058227848101267e-06,
1057
+ "loss": 0.0718,
1058
+ "step": 4225
1059
+ },
1060
+ {
1061
+ "epoch": 2.02,
1062
+ "learning_rate": 9.051898734177215e-06,
1063
+ "loss": 0.0737,
1064
+ "step": 4250
1065
+ },
1066
+ {
1067
+ "epoch": 2.02,
1068
+ "learning_rate": 9.045569620253164e-06,
1069
+ "loss": 0.0731,
1070
+ "step": 4275
1071
+ },
1072
+ {
1073
+ "epoch": 2.02,
1074
+ "learning_rate": 9.039240506329114e-06,
1075
+ "loss": 0.087,
1076
+ "step": 4300
1077
+ },
1078
+ {
1079
+ "epoch": 2.02,
1080
+ "learning_rate": 9.032911392405063e-06,
1081
+ "loss": 0.0875,
1082
+ "step": 4325
1083
+ },
1084
+ {
1085
+ "epoch": 2.02,
1086
+ "learning_rate": 9.026582278481013e-06,
1087
+ "loss": 0.0768,
1088
+ "step": 4350
1089
+ },
1090
+ {
1091
+ "epoch": 2.02,
1092
+ "learning_rate": 9.020253164556962e-06,
1093
+ "loss": 0.0842,
1094
+ "step": 4375
1095
+ },
1096
+ {
1097
+ "epoch": 2.02,
1098
+ "learning_rate": 9.013924050632911e-06,
1099
+ "loss": 0.092,
1100
+ "step": 4400
1101
+ },
1102
+ {
1103
+ "epoch": 2.02,
1104
+ "learning_rate": 9.007594936708861e-06,
1105
+ "loss": 0.0949,
1106
+ "step": 4425
1107
+ },
1108
+ {
1109
+ "epoch": 2.02,
1110
+ "learning_rate": 9.00126582278481e-06,
1111
+ "loss": 0.0989,
1112
+ "step": 4450
1113
+ },
1114
+ {
1115
+ "epoch": 2.02,
1116
+ "learning_rate": 8.99493670886076e-06,
1117
+ "loss": 0.0908,
1118
+ "step": 4475
1119
+ },
1120
+ {
1121
+ "epoch": 2.02,
1122
+ "learning_rate": 8.98860759493671e-06,
1123
+ "loss": 0.0783,
1124
+ "step": 4500
1125
+ },
1126
+ {
1127
+ "epoch": 2.03,
1128
+ "learning_rate": 8.982278481012659e-06,
1129
+ "loss": 0.0651,
1130
+ "step": 4525
1131
+ },
1132
+ {
1133
+ "epoch": 2.03,
1134
+ "learning_rate": 8.975949367088608e-06,
1135
+ "loss": 0.0791,
1136
+ "step": 4550
1137
+ },
1138
+ {
1139
+ "epoch": 2.03,
1140
+ "learning_rate": 8.969620253164557e-06,
1141
+ "loss": 0.0829,
1142
+ "step": 4575
1143
+ },
1144
+ {
1145
+ "epoch": 2.03,
1146
+ "learning_rate": 8.963291139240507e-06,
1147
+ "loss": 0.0797,
1148
+ "step": 4600
1149
+ },
1150
+ {
1151
+ "epoch": 2.03,
1152
+ "learning_rate": 8.956962025316456e-06,
1153
+ "loss": 0.0743,
1154
+ "step": 4625
1155
+ },
1156
+ {
1157
+ "epoch": 2.03,
1158
+ "learning_rate": 8.950632911392406e-06,
1159
+ "loss": 0.0814,
1160
+ "step": 4650
1161
+ },
1162
+ {
1163
+ "epoch": 2.03,
1164
+ "learning_rate": 8.944303797468355e-06,
1165
+ "loss": 0.0594,
1166
+ "step": 4675
1167
+ },
1168
+ {
1169
+ "epoch": 2.03,
1170
+ "learning_rate": 8.937974683544305e-06,
1171
+ "loss": 0.0709,
1172
+ "step": 4700
1173
+ },
1174
+ {
1175
+ "epoch": 2.03,
1176
+ "learning_rate": 8.931645569620254e-06,
1177
+ "loss": 0.0621,
1178
+ "step": 4725
1179
+ },
1180
+ {
1181
+ "epoch": 2.03,
1182
+ "learning_rate": 8.925316455696203e-06,
1183
+ "loss": 0.0599,
1184
+ "step": 4750
1185
+ },
1186
+ {
1187
+ "epoch": 2.03,
1188
+ "learning_rate": 8.918987341772153e-06,
1189
+ "loss": 0.0625,
1190
+ "step": 4775
1191
+ },
1192
+ {
1193
+ "epoch": 2.03,
1194
+ "learning_rate": 8.912658227848102e-06,
1195
+ "loss": 0.0541,
1196
+ "step": 4800
1197
+ },
1198
+ {
1199
+ "epoch": 2.03,
1200
+ "learning_rate": 8.906329113924052e-06,
1201
+ "loss": 0.059,
1202
+ "step": 4825
1203
+ },
1204
+ {
1205
+ "epoch": 2.03,
1206
+ "learning_rate": 8.900000000000001e-06,
1207
+ "loss": 0.0576,
1208
+ "step": 4850
1209
+ },
1210
+ {
1211
+ "epoch": 2.03,
1212
+ "learning_rate": 8.89367088607595e-06,
1213
+ "loss": 0.0623,
1214
+ "step": 4875
1215
+ },
1216
+ {
1217
+ "epoch": 2.03,
1218
+ "learning_rate": 8.8873417721519e-06,
1219
+ "loss": 0.0592,
1220
+ "step": 4900
1221
+ },
1222
+ {
1223
+ "epoch": 2.04,
1224
+ "learning_rate": 8.88101265822785e-06,
1225
+ "loss": 0.0692,
1226
+ "step": 4925
1227
+ },
1228
+ {
1229
+ "epoch": 2.04,
1230
+ "learning_rate": 8.874683544303799e-06,
1231
+ "loss": 0.077,
1232
+ "step": 4950
1233
+ },
1234
+ {
1235
+ "epoch": 2.04,
1236
+ "learning_rate": 8.868354430379748e-06,
1237
+ "loss": 0.0662,
1238
+ "step": 4975
1239
+ },
1240
+ {
1241
+ "epoch": 2.04,
1242
+ "learning_rate": 8.862025316455696e-06,
1243
+ "loss": 0.0605,
1244
+ "step": 5000
1245
+ },
1246
+ {
1247
+ "epoch": 2.04,
1248
+ "eval_loss": 0.2468421906232834,
1249
+ "eval_runtime": 1442.9116,
1250
+ "eval_samples_per_second": 7.215,
1251
+ "eval_steps_per_second": 0.451,
1252
+ "eval_wer": 33.46883689165677,
1253
+ "step": 5000
1254
+ },
1255
+ {
1256
+ "epoch": 2.04,
1257
+ "learning_rate": 8.855696202531645e-06,
1258
+ "loss": 0.0769,
1259
+ "step": 5025
1260
+ },
1261
+ {
1262
+ "epoch": 2.04,
1263
+ "learning_rate": 8.849367088607595e-06,
1264
+ "loss": 0.0601,
1265
+ "step": 5050
1266
+ },
1267
+ {
1268
+ "epoch": 2.04,
1269
+ "learning_rate": 8.843037974683544e-06,
1270
+ "loss": 0.0698,
1271
+ "step": 5075
1272
+ },
1273
+ {
1274
+ "epoch": 2.04,
1275
+ "learning_rate": 8.836708860759494e-06,
1276
+ "loss": 0.0593,
1277
+ "step": 5100
1278
+ },
1279
+ {
1280
+ "epoch": 2.04,
1281
+ "learning_rate": 8.830379746835443e-06,
1282
+ "loss": 0.0521,
1283
+ "step": 5125
1284
+ },
1285
+ {
1286
+ "epoch": 2.04,
1287
+ "learning_rate": 8.824050632911393e-06,
1288
+ "loss": 0.0687,
1289
+ "step": 5150
1290
+ },
1291
+ {
1292
+ "epoch": 2.04,
1293
+ "learning_rate": 8.817721518987342e-06,
1294
+ "loss": 0.0818,
1295
+ "step": 5175
1296
+ },
1297
+ {
1298
+ "epoch": 2.04,
1299
+ "learning_rate": 8.811392405063291e-06,
1300
+ "loss": 0.0706,
1301
+ "step": 5200
1302
+ },
1303
+ {
1304
+ "epoch": 2.04,
1305
+ "learning_rate": 8.805063291139241e-06,
1306
+ "loss": 0.0734,
1307
+ "step": 5225
1308
+ },
1309
+ {
1310
+ "epoch": 2.04,
1311
+ "learning_rate": 8.79873417721519e-06,
1312
+ "loss": 0.0795,
1313
+ "step": 5250
1314
+ },
1315
+ {
1316
+ "epoch": 3.0,
1317
+ "learning_rate": 8.79240506329114e-06,
1318
+ "loss": 0.0905,
1319
+ "step": 5275
1320
+ },
1321
+ {
1322
+ "epoch": 3.0,
1323
+ "learning_rate": 8.78607594936709e-06,
1324
+ "loss": 0.086,
1325
+ "step": 5300
1326
+ },
1327
+ {
1328
+ "epoch": 3.0,
1329
+ "learning_rate": 8.779746835443039e-06,
1330
+ "loss": 0.105,
1331
+ "step": 5325
1332
+ },
1333
+ {
1334
+ "epoch": 3.0,
1335
+ "learning_rate": 8.773417721518988e-06,
1336
+ "loss": 0.0981,
1337
+ "step": 5350
1338
+ },
1339
+ {
1340
+ "epoch": 3.0,
1341
+ "learning_rate": 8.767088607594937e-06,
1342
+ "loss": 0.1131,
1343
+ "step": 5375
1344
+ },
1345
+ {
1346
+ "epoch": 3.0,
1347
+ "learning_rate": 8.760759493670887e-06,
1348
+ "loss": 0.0916,
1349
+ "step": 5400
1350
+ },
1351
+ {
1352
+ "epoch": 3.0,
1353
+ "learning_rate": 8.754430379746836e-06,
1354
+ "loss": 0.0974,
1355
+ "step": 5425
1356
+ },
1357
+ {
1358
+ "epoch": 3.0,
1359
+ "learning_rate": 8.748101265822786e-06,
1360
+ "loss": 0.0899,
1361
+ "step": 5450
1362
+ },
1363
+ {
1364
+ "epoch": 3.01,
1365
+ "learning_rate": 8.741772151898735e-06,
1366
+ "loss": 0.1104,
1367
+ "step": 5475
1368
+ },
1369
+ {
1370
+ "epoch": 3.01,
1371
+ "learning_rate": 8.735443037974685e-06,
1372
+ "loss": 0.0818,
1373
+ "step": 5500
1374
+ },
1375
+ {
1376
+ "epoch": 3.01,
1377
+ "learning_rate": 8.729113924050634e-06,
1378
+ "loss": 0.0764,
1379
+ "step": 5525
1380
+ },
1381
+ {
1382
+ "epoch": 3.01,
1383
+ "learning_rate": 8.722784810126583e-06,
1384
+ "loss": 0.0709,
1385
+ "step": 5550
1386
+ },
1387
+ {
1388
+ "epoch": 3.01,
1389
+ "learning_rate": 8.716455696202533e-06,
1390
+ "loss": 0.0891,
1391
+ "step": 5575
1392
+ },
1393
+ {
1394
+ "epoch": 3.01,
1395
+ "learning_rate": 8.710126582278482e-06,
1396
+ "loss": 0.0869,
1397
+ "step": 5600
1398
+ },
1399
+ {
1400
+ "epoch": 3.01,
1401
+ "learning_rate": 8.703797468354432e-06,
1402
+ "loss": 0.0631,
1403
+ "step": 5625
1404
+ },
1405
+ {
1406
+ "epoch": 3.01,
1407
+ "learning_rate": 8.697468354430381e-06,
1408
+ "loss": 0.0795,
1409
+ "step": 5650
1410
+ },
1411
+ {
1412
+ "epoch": 3.01,
1413
+ "learning_rate": 8.69113924050633e-06,
1414
+ "loss": 0.0603,
1415
+ "step": 5675
1416
+ },
1417
+ {
1418
+ "epoch": 3.01,
1419
+ "learning_rate": 8.68481012658228e-06,
1420
+ "loss": 0.064,
1421
+ "step": 5700
1422
+ },
1423
+ {
1424
+ "epoch": 3.01,
1425
+ "learning_rate": 8.67848101265823e-06,
1426
+ "loss": 0.0738,
1427
+ "step": 5725
1428
+ },
1429
+ {
1430
+ "epoch": 3.01,
1431
+ "learning_rate": 8.672151898734177e-06,
1432
+ "loss": 0.0796,
1433
+ "step": 5750
1434
+ },
1435
+ {
1436
+ "epoch": 3.01,
1437
+ "learning_rate": 8.665822784810127e-06,
1438
+ "loss": 0.0723,
1439
+ "step": 5775
1440
+ },
1441
+ {
1442
+ "epoch": 3.01,
1443
+ "learning_rate": 8.659493670886076e-06,
1444
+ "loss": 0.0592,
1445
+ "step": 5800
1446
+ },
1447
+ {
1448
+ "epoch": 3.01,
1449
+ "learning_rate": 8.653164556962025e-06,
1450
+ "loss": 0.051,
1451
+ "step": 5825
1452
+ },
1453
+ {
1454
+ "epoch": 3.01,
1455
+ "learning_rate": 8.646835443037975e-06,
1456
+ "loss": 0.0716,
1457
+ "step": 5850
1458
+ },
1459
+ {
1460
+ "epoch": 3.02,
1461
+ "learning_rate": 8.640506329113924e-06,
1462
+ "loss": 0.0598,
1463
+ "step": 5875
1464
+ },
1465
+ {
1466
+ "epoch": 3.02,
1467
+ "learning_rate": 8.634177215189874e-06,
1468
+ "loss": 0.0524,
1469
+ "step": 5900
1470
+ },
1471
+ {
1472
+ "epoch": 3.02,
1473
+ "learning_rate": 8.627848101265823e-06,
1474
+ "loss": 0.0623,
1475
+ "step": 5925
1476
+ },
1477
+ {
1478
+ "epoch": 3.02,
1479
+ "learning_rate": 8.621518987341773e-06,
1480
+ "loss": 0.054,
1481
+ "step": 5950
1482
+ },
1483
+ {
1484
+ "epoch": 3.02,
1485
+ "learning_rate": 8.615189873417722e-06,
1486
+ "loss": 0.057,
1487
+ "step": 5975
1488
+ },
1489
+ {
1490
+ "epoch": 3.02,
1491
+ "learning_rate": 8.608860759493671e-06,
1492
+ "loss": 0.0479,
1493
+ "step": 6000
1494
+ },
1495
+ {
1496
+ "epoch": 3.02,
1497
+ "eval_loss": 0.24428367614746094,
1498
+ "eval_runtime": 1457.5568,
1499
+ "eval_samples_per_second": 7.143,
1500
+ "eval_steps_per_second": 0.447,
1501
+ "eval_wer": 36.51929299663409,
1502
+ "step": 6000
1503
+ },
1504
+ {
1505
+ "epoch": 3.02,
1506
+ "learning_rate": 8.602531645569621e-06,
1507
+ "loss": 0.0501,
1508
+ "step": 6025
1509
+ },
1510
+ {
1511
+ "epoch": 3.02,
1512
+ "learning_rate": 8.59620253164557e-06,
1513
+ "loss": 0.0527,
1514
+ "step": 6050
1515
+ },
1516
+ {
1517
+ "epoch": 3.02,
1518
+ "learning_rate": 8.58987341772152e-06,
1519
+ "loss": 0.0558,
1520
+ "step": 6075
1521
+ },
1522
+ {
1523
+ "epoch": 3.02,
1524
+ "learning_rate": 8.58354430379747e-06,
1525
+ "loss": 0.0548,
1526
+ "step": 6100
1527
+ },
1528
+ {
1529
+ "epoch": 3.02,
1530
+ "learning_rate": 8.577215189873419e-06,
1531
+ "loss": 0.0503,
1532
+ "step": 6125
1533
+ },
1534
+ {
1535
+ "epoch": 3.02,
1536
+ "learning_rate": 8.570886075949368e-06,
1537
+ "loss": 0.0641,
1538
+ "step": 6150
1539
+ },
1540
+ {
1541
+ "epoch": 3.02,
1542
+ "learning_rate": 8.564556962025316e-06,
1543
+ "loss": 0.0748,
1544
+ "step": 6175
1545
+ },
1546
+ {
1547
+ "epoch": 3.02,
1548
+ "learning_rate": 8.558227848101267e-06,
1549
+ "loss": 0.0775,
1550
+ "step": 6200
1551
+ },
1552
+ {
1553
+ "epoch": 3.02,
1554
+ "learning_rate": 8.551898734177216e-06,
1555
+ "loss": 0.0662,
1556
+ "step": 6225
1557
+ },
1558
+ {
1559
+ "epoch": 3.02,
1560
+ "learning_rate": 8.545569620253166e-06,
1561
+ "loss": 0.0591,
1562
+ "step": 6250
1563
+ },
1564
+ {
1565
+ "epoch": 3.03,
1566
+ "learning_rate": 8.539240506329115e-06,
1567
+ "loss": 0.0421,
1568
+ "step": 6275
1569
+ },
1570
+ {
1571
+ "epoch": 3.03,
1572
+ "learning_rate": 8.532911392405065e-06,
1573
+ "loss": 0.0565,
1574
+ "step": 6300
1575
+ },
1576
+ {
1577
+ "epoch": 3.03,
1578
+ "learning_rate": 8.526582278481014e-06,
1579
+ "loss": 0.0574,
1580
+ "step": 6325
1581
+ },
1582
+ {
1583
+ "epoch": 3.03,
1584
+ "learning_rate": 8.520253164556963e-06,
1585
+ "loss": 0.0466,
1586
+ "step": 6350
1587
+ },
1588
+ {
1589
+ "epoch": 3.03,
1590
+ "learning_rate": 8.513924050632913e-06,
1591
+ "loss": 0.0486,
1592
+ "step": 6375
1593
+ },
1594
+ {
1595
+ "epoch": 3.03,
1596
+ "learning_rate": 8.507594936708862e-06,
1597
+ "loss": 0.0502,
1598
+ "step": 6400
1599
+ },
1600
+ {
1601
+ "epoch": 3.03,
1602
+ "learning_rate": 8.501265822784812e-06,
1603
+ "loss": 0.0415,
1604
+ "step": 6425
1605
+ },
1606
+ {
1607
+ "epoch": 3.03,
1608
+ "learning_rate": 8.494936708860761e-06,
1609
+ "loss": 0.0449,
1610
+ "step": 6450
1611
+ },
1612
+ {
1613
+ "epoch": 3.03,
1614
+ "learning_rate": 8.488607594936709e-06,
1615
+ "loss": 0.0525,
1616
+ "step": 6475
1617
+ },
1618
+ {
1619
+ "epoch": 3.03,
1620
+ "learning_rate": 8.482278481012658e-06,
1621
+ "loss": 0.0401,
1622
+ "step": 6500
1623
+ },
1624
+ {
1625
+ "epoch": 3.03,
1626
+ "learning_rate": 8.475949367088608e-06,
1627
+ "loss": 0.0502,
1628
+ "step": 6525
1629
+ },
1630
+ {
1631
+ "epoch": 3.03,
1632
+ "learning_rate": 8.469620253164557e-06,
1633
+ "loss": 0.0453,
1634
+ "step": 6550
1635
+ },
1636
+ {
1637
+ "epoch": 3.03,
1638
+ "learning_rate": 8.463291139240507e-06,
1639
+ "loss": 0.0442,
1640
+ "step": 6575
1641
+ },
1642
+ {
1643
+ "epoch": 3.03,
1644
+ "learning_rate": 8.456962025316456e-06,
1645
+ "loss": 0.0423,
1646
+ "step": 6600
1647
+ },
1648
+ {
1649
+ "epoch": 3.03,
1650
+ "learning_rate": 8.450632911392405e-06,
1651
+ "loss": 0.0518,
1652
+ "step": 6625
1653
+ },
1654
+ {
1655
+ "epoch": 3.03,
1656
+ "learning_rate": 8.444303797468355e-06,
1657
+ "loss": 0.0479,
1658
+ "step": 6650
1659
+ },
1660
+ {
1661
+ "epoch": 3.04,
1662
+ "learning_rate": 8.437974683544304e-06,
1663
+ "loss": 0.0521,
1664
+ "step": 6675
1665
+ },
1666
+ {
1667
+ "epoch": 3.04,
1668
+ "learning_rate": 8.431645569620254e-06,
1669
+ "loss": 0.0436,
1670
+ "step": 6700
1671
+ },
1672
+ {
1673
+ "epoch": 3.04,
1674
+ "learning_rate": 8.425316455696203e-06,
1675
+ "loss": 0.0539,
1676
+ "step": 6725
1677
+ },
1678
+ {
1679
+ "epoch": 3.04,
1680
+ "learning_rate": 8.418987341772153e-06,
1681
+ "loss": 0.0446,
1682
+ "step": 6750
1683
+ },
1684
+ {
1685
+ "epoch": 3.04,
1686
+ "learning_rate": 8.412658227848102e-06,
1687
+ "loss": 0.0521,
1688
+ "step": 6775
1689
+ },
1690
+ {
1691
+ "epoch": 3.04,
1692
+ "learning_rate": 8.406329113924051e-06,
1693
+ "loss": 0.0391,
1694
+ "step": 6800
1695
+ },
1696
+ {
1697
+ "epoch": 3.04,
1698
+ "learning_rate": 8.400000000000001e-06,
1699
+ "loss": 0.0513,
1700
+ "step": 6825
1701
+ },
1702
+ {
1703
+ "epoch": 3.04,
1704
+ "learning_rate": 8.39367088607595e-06,
1705
+ "loss": 0.0449,
1706
+ "step": 6850
1707
+ },
1708
+ {
1709
+ "epoch": 3.04,
1710
+ "learning_rate": 8.3873417721519e-06,
1711
+ "loss": 0.0463,
1712
+ "step": 6875
1713
+ },
1714
+ {
1715
+ "epoch": 3.04,
1716
+ "learning_rate": 8.38101265822785e-06,
1717
+ "loss": 0.0495,
1718
+ "step": 6900
1719
+ },
1720
+ {
1721
+ "epoch": 3.04,
1722
+ "learning_rate": 8.374683544303797e-06,
1723
+ "loss": 0.0628,
1724
+ "step": 6925
1725
+ },
1726
+ {
1727
+ "epoch": 3.04,
1728
+ "learning_rate": 8.368354430379746e-06,
1729
+ "loss": 0.0594,
1730
+ "step": 6950
1731
+ },
1732
+ {
1733
+ "epoch": 3.04,
1734
+ "learning_rate": 8.362025316455696e-06,
1735
+ "loss": 0.0573,
1736
+ "step": 6975
1737
+ },
1738
+ {
1739
+ "epoch": 3.04,
1740
+ "learning_rate": 8.355696202531645e-06,
1741
+ "loss": 0.057,
1742
+ "step": 7000
1743
+ },
1744
+ {
1745
+ "epoch": 3.04,
1746
+ "eval_loss": 0.25129273533821106,
1747
+ "eval_runtime": 1479.2043,
1748
+ "eval_samples_per_second": 7.038,
1749
+ "eval_steps_per_second": 0.44,
1750
+ "eval_wer": 57.010683100538216,
1751
+ "step": 7000
1752
+ },
1753
+ {
1754
+ "epoch": 4.0,
1755
+ "learning_rate": 8.349367088607596e-06,
1756
+ "loss": 0.0534,
1757
+ "step": 7025
1758
+ },
1759
+ {
1760
+ "epoch": 4.0,
1761
+ "learning_rate": 8.343037974683546e-06,
1762
+ "loss": 0.0663,
1763
+ "step": 7050
1764
+ },
1765
+ {
1766
+ "epoch": 4.0,
1767
+ "learning_rate": 8.336708860759495e-06,
1768
+ "loss": 0.0698,
1769
+ "step": 7075
1770
+ },
1771
+ {
1772
+ "epoch": 4.0,
1773
+ "learning_rate": 8.330379746835445e-06,
1774
+ "loss": 0.0633,
1775
+ "step": 7100
1776
+ },
1777
+ {
1778
+ "epoch": 4.0,
1779
+ "learning_rate": 8.324050632911394e-06,
1780
+ "loss": 0.0808,
1781
+ "step": 7125
1782
+ },
1783
+ {
1784
+ "epoch": 4.0,
1785
+ "learning_rate": 8.317721518987343e-06,
1786
+ "loss": 0.0699,
1787
+ "step": 7150
1788
+ },
1789
+ {
1790
+ "epoch": 4.0,
1791
+ "learning_rate": 8.311392405063293e-06,
1792
+ "loss": 0.0736,
1793
+ "step": 7175
1794
+ },
1795
+ {
1796
+ "epoch": 4.0,
1797
+ "learning_rate": 8.305063291139242e-06,
1798
+ "loss": 0.069,
1799
+ "step": 7200
1800
+ },
1801
+ {
1802
+ "epoch": 4.01,
1803
+ "learning_rate": 8.29873417721519e-06,
1804
+ "loss": 0.0718,
1805
+ "step": 7225
1806
+ },
1807
+ {
1808
+ "epoch": 4.01,
1809
+ "learning_rate": 8.29240506329114e-06,
1810
+ "loss": 0.0674,
1811
+ "step": 7250
1812
+ },
1813
+ {
1814
+ "epoch": 4.01,
1815
+ "learning_rate": 8.286075949367089e-06,
1816
+ "loss": 0.0568,
1817
+ "step": 7275
1818
+ },
1819
+ {
1820
+ "epoch": 4.01,
1821
+ "learning_rate": 8.279746835443038e-06,
1822
+ "loss": 0.0497,
1823
+ "step": 7300
1824
+ },
1825
+ {
1826
+ "epoch": 4.01,
1827
+ "learning_rate": 8.273417721518988e-06,
1828
+ "loss": 0.0563,
1829
+ "step": 7325
1830
+ },
1831
+ {
1832
+ "epoch": 4.01,
1833
+ "learning_rate": 8.267088607594937e-06,
1834
+ "loss": 0.0649,
1835
+ "step": 7350
1836
+ },
1837
+ {
1838
+ "epoch": 4.01,
1839
+ "learning_rate": 8.260759493670887e-06,
1840
+ "loss": 0.042,
1841
+ "step": 7375
1842
+ },
1843
+ {
1844
+ "epoch": 4.01,
1845
+ "learning_rate": 8.254430379746836e-06,
1846
+ "loss": 0.0559,
1847
+ "step": 7400
1848
+ },
1849
+ {
1850
+ "epoch": 4.01,
1851
+ "learning_rate": 8.248101265822785e-06,
1852
+ "loss": 0.0476,
1853
+ "step": 7425
1854
+ },
1855
+ {
1856
+ "epoch": 4.01,
1857
+ "learning_rate": 8.241772151898735e-06,
1858
+ "loss": 0.0427,
1859
+ "step": 7450
1860
+ },
1861
+ {
1862
+ "epoch": 4.01,
1863
+ "learning_rate": 8.235443037974684e-06,
1864
+ "loss": 0.0533,
1865
+ "step": 7475
1866
+ },
1867
+ {
1868
+ "epoch": 4.01,
1869
+ "learning_rate": 8.229113924050634e-06,
1870
+ "loss": 0.0633,
1871
+ "step": 7500
1872
+ },
1873
+ {
1874
+ "epoch": 4.01,
1875
+ "learning_rate": 8.222784810126583e-06,
1876
+ "loss": 0.055,
1877
+ "step": 7525
1878
+ },
1879
+ {
1880
+ "epoch": 4.01,
1881
+ "learning_rate": 8.216455696202533e-06,
1882
+ "loss": 0.043,
1883
+ "step": 7550
1884
+ },
1885
+ {
1886
+ "epoch": 4.01,
1887
+ "learning_rate": 8.210126582278482e-06,
1888
+ "loss": 0.0421,
1889
+ "step": 7575
1890
+ },
1891
+ {
1892
+ "epoch": 4.01,
1893
+ "learning_rate": 8.203797468354431e-06,
1894
+ "loss": 0.0519,
1895
+ "step": 7600
1896
+ },
1897
+ {
1898
+ "epoch": 4.02,
1899
+ "learning_rate": 8.197468354430381e-06,
1900
+ "loss": 0.0398,
1901
+ "step": 7625
1902
+ },
1903
+ {
1904
+ "epoch": 4.02,
1905
+ "learning_rate": 8.19113924050633e-06,
1906
+ "loss": 0.041,
1907
+ "step": 7650
1908
+ },
1909
+ {
1910
+ "epoch": 4.02,
1911
+ "learning_rate": 8.184810126582278e-06,
1912
+ "loss": 0.0477,
1913
+ "step": 7675
1914
+ },
1915
+ {
1916
+ "epoch": 4.02,
1917
+ "learning_rate": 8.178481012658227e-06,
1918
+ "loss": 0.046,
1919
+ "step": 7700
1920
+ },
1921
+ {
1922
+ "epoch": 4.02,
1923
+ "learning_rate": 8.172151898734177e-06,
1924
+ "loss": 0.0465,
1925
+ "step": 7725
1926
+ },
1927
+ {
1928
+ "epoch": 4.02,
1929
+ "learning_rate": 8.165822784810126e-06,
1930
+ "loss": 0.0359,
1931
+ "step": 7750
1932
+ },
1933
+ {
1934
+ "epoch": 4.02,
1935
+ "learning_rate": 8.159493670886076e-06,
1936
+ "loss": 0.0362,
1937
+ "step": 7775
1938
+ },
1939
+ {
1940
+ "epoch": 4.02,
1941
+ "learning_rate": 8.153164556962025e-06,
1942
+ "loss": 0.04,
1943
+ "step": 7800
1944
+ },
1945
+ {
1946
+ "epoch": 4.02,
1947
+ "learning_rate": 8.146835443037975e-06,
1948
+ "loss": 0.0433,
1949
+ "step": 7825
1950
+ },
1951
+ {
1952
+ "epoch": 4.02,
1953
+ "learning_rate": 8.140506329113926e-06,
1954
+ "loss": 0.0399,
1955
+ "step": 7850
1956
+ },
1957
+ {
1958
+ "epoch": 4.02,
1959
+ "learning_rate": 8.134177215189875e-06,
1960
+ "loss": 0.0396,
1961
+ "step": 7875
1962
+ },
1963
+ {
1964
+ "epoch": 4.02,
1965
+ "learning_rate": 8.127848101265825e-06,
1966
+ "loss": 0.0472,
1967
+ "step": 7900
1968
+ },
1969
+ {
1970
+ "epoch": 4.02,
1971
+ "learning_rate": 8.121518987341774e-06,
1972
+ "loss": 0.0533,
1973
+ "step": 7925
1974
+ },
1975
+ {
1976
+ "epoch": 4.02,
1977
+ "learning_rate": 8.115189873417722e-06,
1978
+ "loss": 0.0564,
1979
+ "step": 7950
1980
+ },
1981
+ {
1982
+ "epoch": 4.02,
1983
+ "learning_rate": 8.108860759493671e-06,
1984
+ "loss": 0.0508,
1985
+ "step": 7975
1986
+ },
1987
+ {
1988
+ "epoch": 4.02,
1989
+ "learning_rate": 8.10253164556962e-06,
1990
+ "loss": 0.0375,
1991
+ "step": 8000
1992
+ },
1993
+ {
1994
+ "epoch": 4.02,
1995
+ "eval_loss": 0.252152681350708,
1996
+ "eval_runtime": 1456.2607,
1997
+ "eval_samples_per_second": 7.149,
1998
+ "eval_steps_per_second": 0.447,
1999
+ "eval_wer": 44.38120945056017,
2000
+ "step": 8000
2001
+ },
2002
+ {
2003
+ "epoch": 4.03,
2004
+ "learning_rate": 8.09620253164557e-06,
2005
+ "loss": 0.0296,
2006
+ "step": 8025
2007
+ },
2008
+ {
2009
+ "epoch": 4.03,
2010
+ "learning_rate": 8.08987341772152e-06,
2011
+ "loss": 0.0395,
2012
+ "step": 8050
2013
+ },
2014
+ {
2015
+ "epoch": 4.03,
2016
+ "learning_rate": 8.083544303797469e-06,
2017
+ "loss": 0.0409,
2018
+ "step": 8075
2019
+ },
2020
+ {
2021
+ "epoch": 4.03,
2022
+ "learning_rate": 8.077215189873418e-06,
2023
+ "loss": 0.0343,
2024
+ "step": 8100
2025
+ },
2026
+ {
2027
+ "epoch": 4.03,
2028
+ "learning_rate": 8.070886075949368e-06,
2029
+ "loss": 0.0321,
2030
+ "step": 8125
2031
+ },
2032
+ {
2033
+ "epoch": 4.03,
2034
+ "learning_rate": 8.064556962025317e-06,
2035
+ "loss": 0.0333,
2036
+ "step": 8150
2037
+ },
2038
+ {
2039
+ "epoch": 4.03,
2040
+ "learning_rate": 8.058227848101267e-06,
2041
+ "loss": 0.0327,
2042
+ "step": 8175
2043
+ },
2044
+ {
2045
+ "epoch": 4.03,
2046
+ "learning_rate": 8.051898734177216e-06,
2047
+ "loss": 0.0318,
2048
+ "step": 8200
2049
+ },
2050
+ {
2051
+ "epoch": 4.03,
2052
+ "learning_rate": 8.045569620253165e-06,
2053
+ "loss": 0.0405,
2054
+ "step": 8225
2055
+ },
2056
+ {
2057
+ "epoch": 4.03,
2058
+ "learning_rate": 8.039240506329115e-06,
2059
+ "loss": 0.0382,
2060
+ "step": 8250
2061
+ },
2062
+ {
2063
+ "epoch": 4.03,
2064
+ "learning_rate": 8.032911392405064e-06,
2065
+ "loss": 0.0317,
2066
+ "step": 8275
2067
+ },
2068
+ {
2069
+ "epoch": 4.03,
2070
+ "learning_rate": 8.026582278481014e-06,
2071
+ "loss": 0.0285,
2072
+ "step": 8300
2073
+ },
2074
+ {
2075
+ "epoch": 4.03,
2076
+ "learning_rate": 8.020253164556963e-06,
2077
+ "loss": 0.0285,
2078
+ "step": 8325
2079
+ },
2080
+ {
2081
+ "epoch": 4.03,
2082
+ "learning_rate": 8.013924050632913e-06,
2083
+ "loss": 0.0269,
2084
+ "step": 8350
2085
+ },
2086
+ {
2087
+ "epoch": 4.03,
2088
+ "learning_rate": 8.007594936708862e-06,
2089
+ "loss": 0.0334,
2090
+ "step": 8375
2091
+ },
2092
+ {
2093
+ "epoch": 4.03,
2094
+ "learning_rate": 8.00126582278481e-06,
2095
+ "loss": 0.0406,
2096
+ "step": 8400
2097
+ },
2098
+ {
2099
+ "epoch": 4.04,
2100
+ "learning_rate": 7.99493670886076e-06,
2101
+ "loss": 0.038,
2102
+ "step": 8425
2103
+ },
2104
+ {
2105
+ "epoch": 4.04,
2106
+ "learning_rate": 7.988607594936709e-06,
2107
+ "loss": 0.0315,
2108
+ "step": 8450
2109
+ },
2110
+ {
2111
+ "epoch": 4.04,
2112
+ "learning_rate": 7.982278481012658e-06,
2113
+ "loss": 0.0329,
2114
+ "step": 8475
2115
+ },
2116
+ {
2117
+ "epoch": 4.04,
2118
+ "learning_rate": 7.975949367088607e-06,
2119
+ "loss": 0.0373,
2120
+ "step": 8500
2121
+ },
2122
+ {
2123
+ "epoch": 4.04,
2124
+ "learning_rate": 7.969620253164557e-06,
2125
+ "loss": 0.0514,
2126
+ "step": 8525
2127
+ },
2128
+ {
2129
+ "epoch": 4.04,
2130
+ "learning_rate": 7.963291139240506e-06,
2131
+ "loss": 0.0349,
2132
+ "step": 8550
2133
+ },
2134
+ {
2135
+ "epoch": 4.04,
2136
+ "learning_rate": 7.956962025316456e-06,
2137
+ "loss": 0.0385,
2138
+ "step": 8575
2139
+ },
2140
+ {
2141
+ "epoch": 4.04,
2142
+ "learning_rate": 7.950632911392405e-06,
2143
+ "loss": 0.0367,
2144
+ "step": 8600
2145
+ },
2146
+ {
2147
+ "epoch": 4.04,
2148
+ "learning_rate": 7.944303797468355e-06,
2149
+ "loss": 0.0317,
2150
+ "step": 8625
2151
+ },
2152
+ {
2153
+ "epoch": 4.04,
2154
+ "learning_rate": 7.937974683544304e-06,
2155
+ "loss": 0.0368,
2156
+ "step": 8650
2157
+ },
2158
+ {
2159
+ "epoch": 4.04,
2160
+ "learning_rate": 7.931645569620255e-06,
2161
+ "loss": 0.045,
2162
+ "step": 8675
2163
+ },
2164
+ {
2165
+ "epoch": 4.04,
2166
+ "learning_rate": 7.925316455696203e-06,
2167
+ "loss": 0.0456,
2168
+ "step": 8700
2169
+ },
2170
+ {
2171
+ "epoch": 4.04,
2172
+ "learning_rate": 7.918987341772152e-06,
2173
+ "loss": 0.0421,
2174
+ "step": 8725
2175
+ },
2176
+ {
2177
+ "epoch": 4.04,
2178
+ "learning_rate": 7.912658227848102e-06,
2179
+ "loss": 0.043,
2180
+ "step": 8750
2181
+ },
2182
+ {
2183
+ "epoch": 5.0,
2184
+ "learning_rate": 7.906329113924051e-06,
2185
+ "loss": 0.0428,
2186
+ "step": 8775
2187
+ },
2188
+ {
2189
+ "epoch": 5.0,
2190
+ "learning_rate": 7.9e-06,
2191
+ "loss": 0.0459,
2192
+ "step": 8800
2193
+ },
2194
+ {
2195
+ "epoch": 5.0,
2196
+ "learning_rate": 7.89367088607595e-06,
2197
+ "loss": 0.0467,
2198
+ "step": 8825
2199
+ },
2200
+ {
2201
+ "epoch": 5.0,
2202
+ "learning_rate": 7.8873417721519e-06,
2203
+ "loss": 0.0462,
2204
+ "step": 8850
2205
+ },
2206
+ {
2207
+ "epoch": 5.0,
2208
+ "learning_rate": 7.881012658227849e-06,
2209
+ "loss": 0.0578,
2210
+ "step": 8875
2211
+ },
2212
+ {
2213
+ "epoch": 5.0,
2214
+ "learning_rate": 7.874683544303798e-06,
2215
+ "loss": 0.0422,
2216
+ "step": 8900
2217
+ },
2218
+ {
2219
+ "epoch": 5.0,
2220
+ "learning_rate": 7.868354430379748e-06,
2221
+ "loss": 0.0546,
2222
+ "step": 8925
2223
+ },
2224
+ {
2225
+ "epoch": 5.0,
2226
+ "learning_rate": 7.862025316455697e-06,
2227
+ "loss": 0.0515,
2228
+ "step": 8950
2229
+ },
2230
+ {
2231
+ "epoch": 5.01,
2232
+ "learning_rate": 7.855696202531647e-06,
2233
+ "loss": 0.0541,
2234
+ "step": 8975
2235
+ },
2236
+ {
2237
+ "epoch": 5.01,
2238
+ "learning_rate": 7.849367088607596e-06,
2239
+ "loss": 0.0479,
2240
+ "step": 9000
2241
+ },
2242
+ {
2243
+ "epoch": 5.01,
2244
+ "eval_loss": 0.2375001162290573,
2245
+ "eval_runtime": 1456.1088,
2246
+ "eval_samples_per_second": 7.15,
2247
+ "eval_steps_per_second": 0.447,
2248
+ "eval_wer": 31.58262735979447,
2249
+ "step": 9000
2250
+ },
2251
+ {
2252
+ "epoch": 5.01,
2253
+ "learning_rate": 7.843037974683545e-06,
2254
+ "loss": 0.0434,
2255
+ "step": 9025
2256
+ },
2257
+ {
2258
+ "epoch": 5.01,
2259
+ "learning_rate": 7.836708860759495e-06,
2260
+ "loss": 0.033,
2261
+ "step": 9050
2262
+ },
2263
+ {
2264
+ "epoch": 5.01,
2265
+ "learning_rate": 7.830379746835444e-06,
2266
+ "loss": 0.0494,
2267
+ "step": 9075
2268
+ },
2269
+ {
2270
+ "epoch": 5.01,
2271
+ "learning_rate": 7.824050632911394e-06,
2272
+ "loss": 0.0522,
2273
+ "step": 9100
2274
+ },
2275
+ {
2276
+ "epoch": 5.01,
2277
+ "learning_rate": 7.817721518987343e-06,
2278
+ "loss": 0.0338,
2279
+ "step": 9125
2280
+ },
2281
+ {
2282
+ "epoch": 5.01,
2283
+ "learning_rate": 7.811392405063291e-06,
2284
+ "loss": 0.0395,
2285
+ "step": 9150
2286
+ },
2287
+ {
2288
+ "epoch": 5.01,
2289
+ "learning_rate": 7.80506329113924e-06,
2290
+ "loss": 0.0332,
2291
+ "step": 9175
2292
+ },
2293
+ {
2294
+ "epoch": 5.01,
2295
+ "learning_rate": 7.79873417721519e-06,
2296
+ "loss": 0.0381,
2297
+ "step": 9200
2298
+ },
2299
+ {
2300
+ "epoch": 5.01,
2301
+ "learning_rate": 7.79240506329114e-06,
2302
+ "loss": 0.0409,
2303
+ "step": 9225
2304
+ },
2305
+ {
2306
+ "epoch": 5.01,
2307
+ "learning_rate": 7.786075949367089e-06,
2308
+ "loss": 0.0529,
2309
+ "step": 9250
2310
+ },
2311
+ {
2312
+ "epoch": 5.01,
2313
+ "learning_rate": 7.779746835443038e-06,
2314
+ "loss": 0.0498,
2315
+ "step": 9275
2316
+ },
2317
+ {
2318
+ "epoch": 5.01,
2319
+ "learning_rate": 7.773417721518987e-06,
2320
+ "loss": 0.0379,
2321
+ "step": 9300
2322
+ },
2323
+ {
2324
+ "epoch": 5.01,
2325
+ "learning_rate": 7.767088607594937e-06,
2326
+ "loss": 0.0316,
2327
+ "step": 9325
2328
+ },
2329
+ {
2330
+ "epoch": 5.01,
2331
+ "learning_rate": 7.760759493670886e-06,
2332
+ "loss": 0.0398,
2333
+ "step": 9350
2334
+ },
2335
+ {
2336
+ "epoch": 5.02,
2337
+ "learning_rate": 7.754430379746836e-06,
2338
+ "loss": 0.0427,
2339
+ "step": 9375
2340
+ },
2341
+ {
2342
+ "epoch": 5.02,
2343
+ "learning_rate": 7.748101265822785e-06,
2344
+ "loss": 0.0363,
2345
+ "step": 9400
2346
+ },
2347
+ {
2348
+ "epoch": 5.02,
2349
+ "learning_rate": 7.741772151898735e-06,
2350
+ "loss": 0.0393,
2351
+ "step": 9425
2352
+ },
2353
+ {
2354
+ "epoch": 5.02,
2355
+ "learning_rate": 7.735443037974684e-06,
2356
+ "loss": 0.0377,
2357
+ "step": 9450
2358
+ },
2359
+ {
2360
+ "epoch": 5.02,
2361
+ "learning_rate": 7.729113924050633e-06,
2362
+ "loss": 0.0313,
2363
+ "step": 9475
2364
+ },
2365
+ {
2366
+ "epoch": 5.02,
2367
+ "learning_rate": 7.722784810126583e-06,
2368
+ "loss": 0.0359,
2369
+ "step": 9500
2370
+ },
2371
+ {
2372
+ "epoch": 5.02,
2373
+ "learning_rate": 7.716455696202532e-06,
2374
+ "loss": 0.0289,
2375
+ "step": 9525
2376
+ },
2377
+ {
2378
+ "epoch": 5.02,
2379
+ "learning_rate": 7.710126582278482e-06,
2380
+ "loss": 0.0354,
2381
+ "step": 9550
2382
+ },
2383
+ {
2384
+ "epoch": 5.02,
2385
+ "learning_rate": 7.703797468354431e-06,
2386
+ "loss": 0.0309,
2387
+ "step": 9575
2388
+ },
2389
+ {
2390
+ "epoch": 5.02,
2391
+ "learning_rate": 7.69746835443038e-06,
2392
+ "loss": 0.0321,
2393
+ "step": 9600
2394
+ },
2395
+ {
2396
+ "epoch": 5.02,
2397
+ "learning_rate": 7.69113924050633e-06,
2398
+ "loss": 0.0392,
2399
+ "step": 9625
2400
+ },
2401
+ {
2402
+ "epoch": 5.02,
2403
+ "learning_rate": 7.68481012658228e-06,
2404
+ "loss": 0.0362,
2405
+ "step": 9650
2406
+ },
2407
+ {
2408
+ "epoch": 5.02,
2409
+ "learning_rate": 7.678481012658229e-06,
2410
+ "loss": 0.0358,
2411
+ "step": 9675
2412
+ },
2413
+ {
2414
+ "epoch": 5.02,
2415
+ "learning_rate": 7.672151898734178e-06,
2416
+ "loss": 0.045,
2417
+ "step": 9700
2418
+ },
2419
+ {
2420
+ "epoch": 5.02,
2421
+ "learning_rate": 7.665822784810128e-06,
2422
+ "loss": 0.0401,
2423
+ "step": 9725
2424
+ },
2425
+ {
2426
+ "epoch": 5.02,
2427
+ "learning_rate": 7.659493670886077e-06,
2428
+ "loss": 0.0357,
2429
+ "step": 9750
2430
+ },
2431
+ {
2432
+ "epoch": 5.03,
2433
+ "learning_rate": 7.653164556962027e-06,
2434
+ "loss": 0.0256,
2435
+ "step": 9775
2436
+ },
2437
+ {
2438
+ "epoch": 5.03,
2439
+ "learning_rate": 7.646835443037976e-06,
2440
+ "loss": 0.0358,
2441
+ "step": 9800
2442
+ },
2443
+ {
2444
+ "epoch": 5.03,
2445
+ "learning_rate": 7.640506329113925e-06,
2446
+ "loss": 0.0361,
2447
+ "step": 9825
2448
+ },
2449
+ {
2450
+ "epoch": 5.03,
2451
+ "learning_rate": 7.634177215189875e-06,
2452
+ "loss": 0.0365,
2453
+ "step": 9850
2454
+ },
2455
+ {
2456
+ "epoch": 5.03,
2457
+ "learning_rate": 7.6278481012658234e-06,
2458
+ "loss": 0.0344,
2459
+ "step": 9875
2460
+ },
2461
+ {
2462
+ "epoch": 5.03,
2463
+ "learning_rate": 7.621518987341773e-06,
2464
+ "loss": 0.0349,
2465
+ "step": 9900
2466
+ },
2467
+ {
2468
+ "epoch": 5.03,
2469
+ "learning_rate": 7.615189873417722e-06,
2470
+ "loss": 0.0289,
2471
+ "step": 9925
2472
+ },
2473
+ {
2474
+ "epoch": 5.03,
2475
+ "learning_rate": 7.608860759493672e-06,
2476
+ "loss": 0.0259,
2477
+ "step": 9950
2478
+ },
2479
+ {
2480
+ "epoch": 5.03,
2481
+ "learning_rate": 7.602531645569621e-06,
2482
+ "loss": 0.0289,
2483
+ "step": 9975
2484
+ },
2485
+ {
2486
+ "epoch": 5.03,
2487
+ "learning_rate": 7.59620253164557e-06,
2488
+ "loss": 0.0269,
2489
+ "step": 10000
2490
+ },
2491
+ {
2492
+ "epoch": 5.03,
2493
+ "eval_loss": 0.2615405321121216,
2494
+ "eval_runtime": 1420.7197,
2495
+ "eval_samples_per_second": 7.328,
2496
+ "eval_steps_per_second": 0.458,
2497
+ "eval_wer": 36.23310948145498,
2498
+ "step": 10000
2499
+ },
2500
+ {
2501
+ "epoch": 5.03,
2502
+ "learning_rate": 7.589873417721519e-06,
2503
+ "loss": 0.0261,
2504
+ "step": 10025
2505
+ },
2506
+ {
2507
+ "epoch": 5.03,
2508
+ "learning_rate": 7.583544303797469e-06,
2509
+ "loss": 0.0278,
2510
+ "step": 10050
2511
+ },
2512
+ {
2513
+ "epoch": 5.03,
2514
+ "learning_rate": 7.577215189873418e-06,
2515
+ "loss": 0.0248,
2516
+ "step": 10075
2517
+ },
2518
+ {
2519
+ "epoch": 5.03,
2520
+ "learning_rate": 7.5708860759493674e-06,
2521
+ "loss": 0.0263,
2522
+ "step": 10100
2523
+ },
2524
+ {
2525
+ "epoch": 5.03,
2526
+ "learning_rate": 7.564556962025317e-06,
2527
+ "loss": 0.0253,
2528
+ "step": 10125
2529
+ },
2530
+ {
2531
+ "epoch": 5.03,
2532
+ "learning_rate": 7.558227848101266e-06,
2533
+ "loss": 0.0248,
2534
+ "step": 10150
2535
+ },
2536
+ {
2537
+ "epoch": 5.04,
2538
+ "learning_rate": 7.551898734177216e-06,
2539
+ "loss": 0.0281,
2540
+ "step": 10175
2541
+ },
2542
+ {
2543
+ "epoch": 5.04,
2544
+ "learning_rate": 7.545569620253165e-06,
2545
+ "loss": 0.0276,
2546
+ "step": 10200
2547
+ },
2548
+ {
2549
+ "epoch": 5.04,
2550
+ "learning_rate": 7.539240506329114e-06,
2551
+ "loss": 0.0285,
2552
+ "step": 10225
2553
+ },
2554
+ {
2555
+ "epoch": 5.04,
2556
+ "learning_rate": 7.532911392405063e-06,
2557
+ "loss": 0.0265,
2558
+ "step": 10250
2559
+ },
2560
+ {
2561
+ "epoch": 5.04,
2562
+ "learning_rate": 7.526582278481013e-06,
2563
+ "loss": 0.0288,
2564
+ "step": 10275
2565
+ },
2566
+ {
2567
+ "epoch": 5.04,
2568
+ "learning_rate": 7.520253164556963e-06,
2569
+ "loss": 0.0247,
2570
+ "step": 10300
2571
+ },
2572
+ {
2573
+ "epoch": 5.04,
2574
+ "learning_rate": 7.513924050632912e-06,
2575
+ "loss": 0.0281,
2576
+ "step": 10325
2577
+ },
2578
+ {
2579
+ "epoch": 5.04,
2580
+ "learning_rate": 7.507594936708862e-06,
2581
+ "loss": 0.0262,
2582
+ "step": 10350
2583
+ },
2584
+ {
2585
+ "epoch": 5.04,
2586
+ "learning_rate": 7.501265822784811e-06,
2587
+ "loss": 0.0243,
2588
+ "step": 10375
2589
+ },
2590
+ {
2591
+ "epoch": 5.04,
2592
+ "learning_rate": 7.494936708860761e-06,
2593
+ "loss": 0.0331,
2594
+ "step": 10400
2595
+ },
2596
+ {
2597
+ "epoch": 5.04,
2598
+ "learning_rate": 7.48860759493671e-06,
2599
+ "loss": 0.0346,
2600
+ "step": 10425
2601
+ },
2602
+ {
2603
+ "epoch": 5.04,
2604
+ "learning_rate": 7.4822784810126594e-06,
2605
+ "loss": 0.0348,
2606
+ "step": 10450
2607
+ },
2608
+ {
2609
+ "epoch": 5.04,
2610
+ "learning_rate": 7.475949367088609e-06,
2611
+ "loss": 0.0272,
2612
+ "step": 10475
2613
+ },
2614
+ {
2615
+ "epoch": 5.04,
2616
+ "learning_rate": 7.4696202531645574e-06,
2617
+ "loss": 0.0326,
2618
+ "step": 10500
2619
+ },
2620
+ {
2621
+ "epoch": 6.0,
2622
+ "learning_rate": 7.463291139240507e-06,
2623
+ "loss": 0.033,
2624
+ "step": 10525
2625
+ },
2626
+ {
2627
+ "epoch": 6.0,
2628
+ "learning_rate": 7.456962025316456e-06,
2629
+ "loss": 0.0333,
2630
+ "step": 10550
2631
+ },
2632
+ {
2633
+ "epoch": 6.0,
2634
+ "learning_rate": 7.450632911392406e-06,
2635
+ "loss": 0.0371,
2636
+ "step": 10575
2637
+ },
2638
+ {
2639
+ "epoch": 6.0,
2640
+ "learning_rate": 7.444303797468355e-06,
2641
+ "loss": 0.0359,
2642
+ "step": 10600
2643
+ },
2644
+ {
2645
+ "epoch": 6.0,
2646
+ "learning_rate": 7.437974683544305e-06,
2647
+ "loss": 0.0438,
2648
+ "step": 10625
2649
+ },
2650
+ {
2651
+ "epoch": 6.0,
2652
+ "learning_rate": 7.431645569620254e-06,
2653
+ "loss": 0.0382,
2654
+ "step": 10650
2655
+ },
2656
+ {
2657
+ "epoch": 6.0,
2658
+ "learning_rate": 7.4253164556962034e-06,
2659
+ "loss": 0.0337,
2660
+ "step": 10675
2661
+ },
2662
+ {
2663
+ "epoch": 6.0,
2664
+ "learning_rate": 7.418987341772153e-06,
2665
+ "loss": 0.0422,
2666
+ "step": 10700
2667
+ },
2668
+ {
2669
+ "epoch": 6.01,
2670
+ "learning_rate": 7.4126582278481014e-06,
2671
+ "loss": 0.0467,
2672
+ "step": 10725
2673
+ },
2674
+ {
2675
+ "epoch": 6.01,
2676
+ "learning_rate": 7.406329113924051e-06,
2677
+ "loss": 0.0437,
2678
+ "step": 10750
2679
+ },
2680
+ {
2681
+ "epoch": 6.01,
2682
+ "learning_rate": 7.4e-06,
2683
+ "loss": 0.0328,
2684
+ "step": 10775
2685
+ },
2686
+ {
2687
+ "epoch": 6.01,
2688
+ "learning_rate": 7.39367088607595e-06,
2689
+ "loss": 0.0295,
2690
+ "step": 10800
2691
+ },
2692
+ {
2693
+ "epoch": 6.01,
2694
+ "learning_rate": 7.387341772151899e-06,
2695
+ "loss": 0.0431,
2696
+ "step": 10825
2697
+ },
2698
+ {
2699
+ "epoch": 6.01,
2700
+ "learning_rate": 7.381012658227849e-06,
2701
+ "loss": 0.0456,
2702
+ "step": 10850
2703
+ },
2704
+ {
2705
+ "epoch": 6.01,
2706
+ "learning_rate": 7.374683544303798e-06,
2707
+ "loss": 0.0267,
2708
+ "step": 10875
2709
+ },
2710
+ {
2711
+ "epoch": 6.01,
2712
+ "learning_rate": 7.3683544303797474e-06,
2713
+ "loss": 0.0334,
2714
+ "step": 10900
2715
+ },
2716
+ {
2717
+ "epoch": 6.01,
2718
+ "learning_rate": 7.362025316455697e-06,
2719
+ "loss": 0.0296,
2720
+ "step": 10925
2721
+ },
2722
+ {
2723
+ "epoch": 6.01,
2724
+ "learning_rate": 7.3556962025316454e-06,
2725
+ "loss": 0.0261,
2726
+ "step": 10950
2727
+ },
2728
+ {
2729
+ "epoch": 6.01,
2730
+ "learning_rate": 7.349367088607595e-06,
2731
+ "loss": 0.0299,
2732
+ "step": 10975
2733
+ },
2734
+ {
2735
+ "epoch": 6.01,
2736
+ "learning_rate": 7.343037974683544e-06,
2737
+ "loss": 0.0379,
2738
+ "step": 11000
2739
+ },
2740
+ {
2741
+ "epoch": 6.01,
2742
+ "eval_loss": 0.23971551656723022,
2743
+ "eval_runtime": 1435.6772,
2744
+ "eval_samples_per_second": 7.252,
2745
+ "eval_steps_per_second": 0.453,
2746
+ "eval_wer": 31.7712483129807,
2747
+ "step": 11000
2748
+ },
2749
+ {
2750
+ "epoch": 6.01,
2751
+ "learning_rate": 7.336708860759494e-06,
2752
+ "loss": 0.0457,
2753
+ "step": 11025
2754
+ },
2755
+ {
2756
+ "epoch": 6.01,
2757
+ "learning_rate": 7.330379746835443e-06,
2758
+ "loss": 0.0275,
2759
+ "step": 11050
2760
+ },
2761
+ {
2762
+ "epoch": 6.01,
2763
+ "learning_rate": 7.324050632911393e-06,
2764
+ "loss": 0.0254,
2765
+ "step": 11075
2766
+ },
2767
+ {
2768
+ "epoch": 6.01,
2769
+ "learning_rate": 7.317721518987342e-06,
2770
+ "loss": 0.035,
2771
+ "step": 11100
2772
+ },
2773
+ {
2774
+ "epoch": 6.02,
2775
+ "learning_rate": 7.311392405063292e-06,
2776
+ "loss": 0.0273,
2777
+ "step": 11125
2778
+ },
2779
+ {
2780
+ "epoch": 6.02,
2781
+ "learning_rate": 7.305063291139242e-06,
2782
+ "loss": 0.0227,
2783
+ "step": 11150
2784
+ },
2785
+ {
2786
+ "epoch": 6.02,
2787
+ "learning_rate": 7.298734177215191e-06,
2788
+ "loss": 0.0263,
2789
+ "step": 11175
2790
+ },
2791
+ {
2792
+ "epoch": 6.02,
2793
+ "learning_rate": 7.2924050632911406e-06,
2794
+ "loss": 0.0294,
2795
+ "step": 11200
2796
+ },
2797
+ {
2798
+ "epoch": 6.02,
2799
+ "learning_rate": 7.28607594936709e-06,
2800
+ "loss": 0.026,
2801
+ "step": 11225
2802
+ },
2803
+ {
2804
+ "epoch": 6.02,
2805
+ "learning_rate": 7.279746835443039e-06,
2806
+ "loss": 0.0251,
2807
+ "step": 11250
2808
+ },
2809
+ {
2810
+ "epoch": 6.02,
2811
+ "learning_rate": 7.273417721518988e-06,
2812
+ "loss": 0.024,
2813
+ "step": 11275
2814
+ },
2815
+ {
2816
+ "epoch": 6.02,
2817
+ "learning_rate": 7.2670886075949374e-06,
2818
+ "loss": 0.0344,
2819
+ "step": 11300
2820
+ },
2821
+ {
2822
+ "epoch": 6.02,
2823
+ "learning_rate": 7.260759493670887e-06,
2824
+ "loss": 0.0338,
2825
+ "step": 11325
2826
+ },
2827
+ {
2828
+ "epoch": 6.02,
2829
+ "learning_rate": 7.254430379746836e-06,
2830
+ "loss": 0.0266,
2831
+ "step": 11350
2832
+ },
2833
+ {
2834
+ "epoch": 6.02,
2835
+ "learning_rate": 7.248101265822786e-06,
2836
+ "loss": 0.0271,
2837
+ "step": 11375
2838
+ },
2839
+ {
2840
+ "epoch": 6.02,
2841
+ "learning_rate": 7.241772151898735e-06,
2842
+ "loss": 0.0338,
2843
+ "step": 11400
2844
+ },
2845
+ {
2846
+ "epoch": 6.02,
2847
+ "learning_rate": 7.2354430379746846e-06,
2848
+ "loss": 0.0321,
2849
+ "step": 11425
2850
+ },
2851
+ {
2852
+ "epoch": 6.02,
2853
+ "learning_rate": 7.229113924050634e-06,
2854
+ "loss": 0.0385,
2855
+ "step": 11450
2856
+ },
2857
+ {
2858
+ "epoch": 6.02,
2859
+ "learning_rate": 7.222784810126583e-06,
2860
+ "loss": 0.0284,
2861
+ "step": 11475
2862
+ },
2863
+ {
2864
+ "epoch": 6.02,
2865
+ "learning_rate": 7.216455696202532e-06,
2866
+ "loss": 0.0268,
2867
+ "step": 11500
2868
+ },
2869
+ {
2870
+ "epoch": 6.03,
2871
+ "learning_rate": 7.2101265822784814e-06,
2872
+ "loss": 0.0191,
2873
+ "step": 11525
2874
+ },
2875
+ {
2876
+ "epoch": 6.03,
2877
+ "learning_rate": 7.203797468354431e-06,
2878
+ "loss": 0.0243,
2879
+ "step": 11550
2880
+ },
2881
+ {
2882
+ "epoch": 6.03,
2883
+ "learning_rate": 7.19746835443038e-06,
2884
+ "loss": 0.0228,
2885
+ "step": 11575
2886
+ },
2887
+ {
2888
+ "epoch": 6.03,
2889
+ "learning_rate": 7.19113924050633e-06,
2890
+ "loss": 0.0187,
2891
+ "step": 11600
2892
+ },
2893
+ {
2894
+ "epoch": 6.03,
2895
+ "learning_rate": 7.184810126582279e-06,
2896
+ "loss": 0.0255,
2897
+ "step": 11625
2898
+ },
2899
+ {
2900
+ "epoch": 6.03,
2901
+ "learning_rate": 7.1784810126582286e-06,
2902
+ "loss": 0.0218,
2903
+ "step": 11650
2904
+ },
2905
+ {
2906
+ "epoch": 6.03,
2907
+ "learning_rate": 7.172151898734178e-06,
2908
+ "loss": 0.0249,
2909
+ "step": 11675
2910
+ },
2911
+ {
2912
+ "epoch": 6.03,
2913
+ "learning_rate": 7.165822784810127e-06,
2914
+ "loss": 0.0254,
2915
+ "step": 11700
2916
+ },
2917
+ {
2918
+ "epoch": 6.03,
2919
+ "learning_rate": 7.159493670886076e-06,
2920
+ "loss": 0.0323,
2921
+ "step": 11725
2922
+ },
2923
+ {
2924
+ "epoch": 6.03,
2925
+ "learning_rate": 7.1531645569620254e-06,
2926
+ "loss": 0.021,
2927
+ "step": 11750
2928
+ },
2929
+ {
2930
+ "epoch": 6.03,
2931
+ "learning_rate": 7.146835443037975e-06,
2932
+ "loss": 0.0242,
2933
+ "step": 11775
2934
+ },
2935
+ {
2936
+ "epoch": 6.03,
2937
+ "learning_rate": 7.140506329113924e-06,
2938
+ "loss": 0.0233,
2939
+ "step": 11800
2940
+ },
2941
+ {
2942
+ "epoch": 6.03,
2943
+ "learning_rate": 7.134177215189874e-06,
2944
+ "loss": 0.0237,
2945
+ "step": 11825
2946
+ },
2947
+ {
2948
+ "epoch": 6.03,
2949
+ "learning_rate": 7.127848101265823e-06,
2950
+ "loss": 0.0257,
2951
+ "step": 11850
2952
+ },
2953
+ {
2954
+ "epoch": 6.03,
2955
+ "learning_rate": 7.121518987341773e-06,
2956
+ "loss": 0.0256,
2957
+ "step": 11875
2958
+ },
2959
+ {
2960
+ "epoch": 6.03,
2961
+ "learning_rate": 7.115189873417722e-06,
2962
+ "loss": 0.0264,
2963
+ "step": 11900
2964
+ },
2965
+ {
2966
+ "epoch": 6.04,
2967
+ "learning_rate": 7.108860759493671e-06,
2968
+ "loss": 0.0254,
2969
+ "step": 11925
2970
+ },
2971
+ {
2972
+ "epoch": 6.04,
2973
+ "learning_rate": 7.102531645569622e-06,
2974
+ "loss": 0.025,
2975
+ "step": 11950
2976
+ },
2977
+ {
2978
+ "epoch": 6.04,
2979
+ "learning_rate": 7.096202531645571e-06,
2980
+ "loss": 0.0219,
2981
+ "step": 11975
2982
+ },
2983
+ {
2984
+ "epoch": 6.04,
2985
+ "learning_rate": 7.08987341772152e-06,
2986
+ "loss": 0.0241,
2987
+ "step": 12000
2988
+ },
2989
+ {
2990
+ "epoch": 6.04,
2991
+ "eval_loss": 0.26222535967826843,
2992
+ "eval_runtime": 1460.2321,
2993
+ "eval_samples_per_second": 7.13,
2994
+ "eval_steps_per_second": 0.446,
2995
+ "eval_wer": 32.87370526350022,
2996
+ "step": 12000
2997
+ },
2998
+ {
2999
+ "epoch": 6.04,
3000
+ "learning_rate": 7.083544303797469e-06,
3001
+ "loss": 0.0337,
3002
+ "step": 12025
3003
+ },
3004
+ {
3005
+ "epoch": 6.04,
3006
+ "learning_rate": 7.0772151898734186e-06,
3007
+ "loss": 0.0179,
3008
+ "step": 12050
3009
+ },
3010
+ {
3011
+ "epoch": 6.04,
3012
+ "learning_rate": 7.070886075949368e-06,
3013
+ "loss": 0.0174,
3014
+ "step": 12075
3015
+ },
3016
+ {
3017
+ "epoch": 6.04,
3018
+ "learning_rate": 7.0645569620253174e-06,
3019
+ "loss": 0.0211,
3020
+ "step": 12100
3021
+ },
3022
+ {
3023
+ "epoch": 6.04,
3024
+ "learning_rate": 7.058227848101267e-06,
3025
+ "loss": 0.0224,
3026
+ "step": 12125
3027
+ },
3028
+ {
3029
+ "epoch": 6.04,
3030
+ "learning_rate": 7.051898734177216e-06,
3031
+ "loss": 0.0226,
3032
+ "step": 12150
3033
+ },
3034
+ {
3035
+ "epoch": 6.04,
3036
+ "learning_rate": 7.045569620253166e-06,
3037
+ "loss": 0.0271,
3038
+ "step": 12175
3039
+ },
3040
+ {
3041
+ "epoch": 6.04,
3042
+ "learning_rate": 7.039240506329114e-06,
3043
+ "loss": 0.0264,
3044
+ "step": 12200
3045
+ },
3046
+ {
3047
+ "epoch": 6.04,
3048
+ "learning_rate": 7.032911392405064e-06,
3049
+ "loss": 0.031,
3050
+ "step": 12225
3051
+ },
3052
+ {
3053
+ "epoch": 6.04,
3054
+ "learning_rate": 7.026582278481013e-06,
3055
+ "loss": 0.022,
3056
+ "step": 12250
3057
+ },
3058
+ {
3059
+ "epoch": 7.0,
3060
+ "learning_rate": 7.0202531645569626e-06,
3061
+ "loss": 0.0266,
3062
+ "step": 12275
3063
+ },
3064
+ {
3065
+ "epoch": 7.0,
3066
+ "learning_rate": 7.013924050632912e-06,
3067
+ "loss": 0.0269,
3068
+ "step": 12300
3069
+ },
3070
+ {
3071
+ "epoch": 7.0,
3072
+ "learning_rate": 7.0075949367088614e-06,
3073
+ "loss": 0.028,
3074
+ "step": 12325
3075
+ },
3076
+ {
3077
+ "epoch": 7.0,
3078
+ "learning_rate": 7.001265822784811e-06,
3079
+ "loss": 0.0361,
3080
+ "step": 12350
3081
+ },
3082
+ {
3083
+ "epoch": 7.0,
3084
+ "learning_rate": 6.99493670886076e-06,
3085
+ "loss": 0.0383,
3086
+ "step": 12375
3087
+ },
3088
+ {
3089
+ "epoch": 7.0,
3090
+ "learning_rate": 6.98860759493671e-06,
3091
+ "loss": 0.0312,
3092
+ "step": 12400
3093
+ },
3094
+ {
3095
+ "epoch": 7.0,
3096
+ "learning_rate": 6.982278481012658e-06,
3097
+ "loss": 0.0275,
3098
+ "step": 12425
3099
+ },
3100
+ {
3101
+ "epoch": 7.0,
3102
+ "learning_rate": 6.975949367088608e-06,
3103
+ "loss": 0.0407,
3104
+ "step": 12450
3105
+ },
3106
+ {
3107
+ "epoch": 7.01,
3108
+ "learning_rate": 6.969620253164557e-06,
3109
+ "loss": 0.0436,
3110
+ "step": 12475
3111
+ },
3112
+ {
3113
+ "epoch": 7.01,
3114
+ "learning_rate": 6.9632911392405066e-06,
3115
+ "loss": 0.0359,
3116
+ "step": 12500
3117
+ },
3118
+ {
3119
+ "epoch": 7.01,
3120
+ "learning_rate": 6.956962025316456e-06,
3121
+ "loss": 0.0281,
3122
+ "step": 12525
3123
+ },
3124
+ {
3125
+ "epoch": 7.01,
3126
+ "learning_rate": 6.9506329113924054e-06,
3127
+ "loss": 0.0283,
3128
+ "step": 12550
3129
+ },
3130
+ {
3131
+ "epoch": 7.01,
3132
+ "learning_rate": 6.944303797468355e-06,
3133
+ "loss": 0.0248,
3134
+ "step": 12575
3135
+ },
3136
+ {
3137
+ "epoch": 7.01,
3138
+ "learning_rate": 6.937974683544304e-06,
3139
+ "loss": 0.0295,
3140
+ "step": 12600
3141
+ },
3142
+ {
3143
+ "epoch": 7.01,
3144
+ "learning_rate": 6.931645569620254e-06,
3145
+ "loss": 0.0287,
3146
+ "step": 12625
3147
+ },
3148
+ {
3149
+ "epoch": 7.01,
3150
+ "learning_rate": 6.925316455696202e-06,
3151
+ "loss": 0.0261,
3152
+ "step": 12650
3153
+ },
3154
+ {
3155
+ "epoch": 7.01,
3156
+ "learning_rate": 6.918987341772152e-06,
3157
+ "loss": 0.0259,
3158
+ "step": 12675
3159
+ },
3160
+ {
3161
+ "epoch": 7.01,
3162
+ "learning_rate": 6.912658227848101e-06,
3163
+ "loss": 0.0265,
3164
+ "step": 12700
3165
+ },
3166
+ {
3167
+ "epoch": 7.01,
3168
+ "learning_rate": 6.906329113924051e-06,
3169
+ "loss": 0.0239,
3170
+ "step": 12725
3171
+ },
3172
+ {
3173
+ "epoch": 7.01,
3174
+ "learning_rate": 6.9e-06,
3175
+ "loss": 0.0342,
3176
+ "step": 12750
3177
+ },
3178
+ {
3179
+ "epoch": 7.01,
3180
+ "learning_rate": 6.89367088607595e-06,
3181
+ "loss": 0.0253,
3182
+ "step": 12775
3183
+ },
3184
+ {
3185
+ "epoch": 7.01,
3186
+ "learning_rate": 6.8873417721519e-06,
3187
+ "loss": 0.0317,
3188
+ "step": 12800
3189
+ },
3190
+ {
3191
+ "epoch": 7.01,
3192
+ "learning_rate": 6.881012658227849e-06,
3193
+ "loss": 0.0233,
3194
+ "step": 12825
3195
+ },
3196
+ {
3197
+ "epoch": 7.01,
3198
+ "learning_rate": 6.8746835443037986e-06,
3199
+ "loss": 0.0198,
3200
+ "step": 12850
3201
+ },
3202
+ {
3203
+ "epoch": 7.02,
3204
+ "learning_rate": 6.868354430379748e-06,
3205
+ "loss": 0.018,
3206
+ "step": 12875
3207
+ },
3208
+ {
3209
+ "epoch": 7.02,
3210
+ "learning_rate": 6.862025316455697e-06,
3211
+ "loss": 0.0182,
3212
+ "step": 12900
3213
+ },
3214
+ {
3215
+ "epoch": 7.02,
3216
+ "learning_rate": 6.855949367088608e-06,
3217
+ "loss": 0.0279,
3218
+ "step": 12925
3219
+ },
3220
+ {
3221
+ "epoch": 7.02,
3222
+ "learning_rate": 6.8496202531645574e-06,
3223
+ "loss": 0.0201,
3224
+ "step": 12950
3225
+ },
3226
+ {
3227
+ "epoch": 7.02,
3228
+ "learning_rate": 6.843291139240507e-06,
3229
+ "loss": 0.0222,
3230
+ "step": 12975
3231
+ },
3232
+ {
3233
+ "epoch": 7.02,
3234
+ "learning_rate": 6.836962025316456e-06,
3235
+ "loss": 0.0236,
3236
+ "step": 13000
3237
+ },
3238
+ {
3239
+ "epoch": 7.02,
3240
+ "eval_loss": 0.2616036534309387,
3241
+ "eval_runtime": 1403.0214,
3242
+ "eval_samples_per_second": 7.42,
3243
+ "eval_steps_per_second": 0.464,
3244
+ "eval_wer": 27.36142051090262,
3245
+ "step": 13000
3246
+ }
3247
+ ],
3248
+ "max_steps": 40000,
3249
+ "num_train_epochs": 9223372036854775807,
3250
+ "total_flos": 6.00055622885376e+19,
3251
+ "trial_name": null,
3252
+ "trial_params": null
3253
+ }
checkpoint-13000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0289e1bf8fd1a33f7c3c9d3e15cc9bdd13d10e9774e0ada934c0b9d71501a993
3
+ size 4091