Hugging Face's logo

CharisK
/
whisper-small-el

Automatic Speech Recognition
Transformers
PyTorch
Greek
whisper
whisper-small
mozilla-foundation/common_voice_13_0
greek
whisper-event
Inference Endpoints
Model card Files Community
charisk commited on
Commit
c415067
•
1 Parent(s): 0465cc9

Upload 10 files

Browse files
Files changed (10) hide show
  1. config.json +51 -0
  2. generation_config.json +221 -0
  3. optimizer.pt +3 -0
  4. preprocessor_config.json +14 -0
  5. pytorch_model.bin +3 -0
  6. rng_state.pth +3 -0
  7. scaler.pt +3 -0
  8. scheduler.pt +3 -0
  9. trainer_state.json +3253 -0
  10. training_args.bin +3 -0
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": false,
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": 0.0,
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": "4.29.0.dev0",
48
+ "use_cache": true,
49
+ "use_weighted_layer_sum": false,
50
+ "vocab_size": 51865
51
+ }
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": "4.29.0.dev0"
221
+ }
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:09fd8168ae39e05103cb8c73491f47d54183939c987a1ec9f7b747fe9b533778
3
+ size 1934161093
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
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3dbd57322bd7589f6386ed63c3bb3814410323295fada74acf2a72c2b2ac654
3
+ size 967102601
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9eea8e1d15f7edb6e02c3613bfc4e72d3dbd3cf7733d66f1d0c0df3938d0b14c
3
+ size 14575
scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a29b374b5f3e1132a9ebb87246ed2aa5d17dfed2cb80ce7e7130ec2456aa1169
3
+ size 557
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad04f8d57796f99fbf1ed435f8929ca37532061bd5efab72f6bb9e231965266d
3
+ size 627
trainer_state.json ADDED
@@ -0,0 +1,3253 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 56.720179204778795,
3
+ "best_model_checkpoint": "CharisMsensis/whisper-small-el/checkpoint-13000",
4
+ "epoch": 14.412416851441241,
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": 0.03,
12
+ "learning_rate": 4.2000000000000006e-07,
13
+ "loss": 3.1562,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.06,
18
+ "learning_rate": 9.200000000000001e-07,
19
+ "loss": 2.4611,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.08,
24
+ "learning_rate": 1.42e-06,
25
+ "loss": 1.9694,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.11,
30
+ "learning_rate": 1.9200000000000003e-06,
31
+ "loss": 1.4359,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.14,
36
+ "learning_rate": 2.42e-06,
37
+ "loss": 1.0483,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.17,
42
+ "learning_rate": 2.92e-06,
43
+ "loss": 1.0062,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.19,
48
+ "learning_rate": 3.4200000000000007e-06,
49
+ "loss": 0.9075,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.22,
54
+ "learning_rate": 3.920000000000001e-06,
55
+ "loss": 0.844,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.25,
60
+ "learning_rate": 4.42e-06,
61
+ "loss": 0.7819,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.28,
66
+ "learning_rate": 4.92e-06,
67
+ "loss": 0.7264,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.3,
72
+ "learning_rate": 5.420000000000001e-06,
73
+ "loss": 0.6921,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.33,
78
+ "learning_rate": 5.92e-06,
79
+ "loss": 0.6049,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.36,
84
+ "learning_rate": 6.42e-06,
85
+ "loss": 0.4845,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.39,
90
+ "learning_rate": 6.92e-06,
91
+ "loss": 0.2998,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.42,
96
+ "learning_rate": 7.420000000000001e-06,
97
+ "loss": 0.3535,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.44,
102
+ "learning_rate": 7.92e-06,
103
+ "loss": 0.3455,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.47,
108
+ "learning_rate": 8.42e-06,
109
+ "loss": 0.3208,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.5,
114
+ "learning_rate": 8.920000000000001e-06,
115
+ "loss": 0.2778,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.53,
120
+ "learning_rate": 9.42e-06,
121
+ "loss": 0.2778,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.55,
126
+ "learning_rate": 9.920000000000002e-06,
127
+ "loss": 0.3278,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.58,
132
+ "learning_rate": 9.98551724137931e-06,
133
+ "loss": 0.274,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.61,
138
+ "learning_rate": 9.968275862068967e-06,
139
+ "loss": 0.2801,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.64,
144
+ "learning_rate": 9.951034482758622e-06,
145
+ "loss": 0.3386,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.67,
150
+ "learning_rate": 9.933793103448277e-06,
151
+ "loss": 0.3055,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.69,
156
+ "learning_rate": 9.916551724137932e-06,
157
+ "loss": 0.3042,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.72,
162
+ "learning_rate": 9.899310344827588e-06,
163
+ "loss": 0.2838,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.75,
168
+ "learning_rate": 9.882068965517241e-06,
169
+ "loss": 0.3243,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.78,
174
+ "learning_rate": 9.864827586206898e-06,
175
+ "loss": 0.2682,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.8,
180
+ "learning_rate": 9.847586206896553e-06,
181
+ "loss": 0.3168,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.83,
186
+ "learning_rate": 9.830344827586208e-06,
187
+ "loss": 0.3085,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.86,
192
+ "learning_rate": 9.813103448275862e-06,
193
+ "loss": 0.2691,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.89,
198
+ "learning_rate": 9.795862068965517e-06,
199
+ "loss": 0.2721,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.91,
204
+ "learning_rate": 9.778620689655172e-06,
205
+ "loss": 0.2691,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.94,
210
+ "learning_rate": 9.761379310344829e-06,
211
+ "loss": 0.2797,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.97,
216
+ "learning_rate": 9.744137931034484e-06,
217
+ "loss": 0.257,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 1.0,
222
+ "learning_rate": 9.726896551724139e-06,
223
+ "loss": 0.2606,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 1.03,
228
+ "learning_rate": 9.709655172413795e-06,
229
+ "loss": 0.1628,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 1.05,
234
+ "learning_rate": 9.692413793103448e-06,
235
+ "loss": 0.1529,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 1.08,
240
+ "learning_rate": 9.675172413793105e-06,
241
+ "loss": 0.162,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 1.11,
246
+ "learning_rate": 9.65793103448276e-06,
247
+ "loss": 0.1399,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 1.11,
252
+ "eval_loss": 0.29191553592681885,
253
+ "eval_runtime": 3943.1183,
254
+ "eval_samples_per_second": 0.432,
255
+ "eval_steps_per_second": 0.054,
256
+ "eval_wer": 243.45715885756954,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 1.14,
261
+ "learning_rate": 9.640689655172415e-06,
262
+ "loss": 0.1552,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 1.16,
267
+ "learning_rate": 9.62344827586207e-06,
268
+ "loss": 0.147,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 1.19,
273
+ "learning_rate": 9.606206896551726e-06,
274
+ "loss": 0.1647,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 1.22,
279
+ "learning_rate": 9.58896551724138e-06,
280
+ "loss": 0.165,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 1.25,
285
+ "learning_rate": 9.571724137931036e-06,
286
+ "loss": 0.1502,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 1.27,
291
+ "learning_rate": 9.55448275862069e-06,
292
+ "loss": 0.1293,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 1.3,
297
+ "learning_rate": 9.537241379310345e-06,
298
+ "loss": 0.1541,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 1.33,
303
+ "learning_rate": 9.52e-06,
304
+ "loss": 0.1492,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 1.36,
309
+ "learning_rate": 9.502758620689655e-06,
310
+ "loss": 0.1632,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 1.39,
315
+ "learning_rate": 9.48551724137931e-06,
316
+ "loss": 0.1591,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 1.41,
321
+ "learning_rate": 9.468275862068967e-06,
322
+ "loss": 0.1542,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 1.44,
327
+ "learning_rate": 9.451034482758622e-06,
328
+ "loss": 0.1511,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 1.47,
333
+ "learning_rate": 9.433793103448276e-06,
334
+ "loss": 0.15,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 1.5,
339
+ "learning_rate": 9.416551724137933e-06,
340
+ "loss": 0.155,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 1.52,
345
+ "learning_rate": 9.399310344827586e-06,
346
+ "loss": 0.1381,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 1.55,
351
+ "learning_rate": 9.382068965517243e-06,
352
+ "loss": 0.1636,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 1.58,
357
+ "learning_rate": 9.364827586206898e-06,
358
+ "loss": 0.1611,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 1.61,
363
+ "learning_rate": 9.347586206896552e-06,
364
+ "loss": 0.1522,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 1.64,
369
+ "learning_rate": 9.330344827586207e-06,
370
+ "loss": 0.1359,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 1.66,
375
+ "learning_rate": 9.313103448275864e-06,
376
+ "loss": 0.1388,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 1.69,
381
+ "learning_rate": 9.295862068965517e-06,
382
+ "loss": 0.1512,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 1.72,
387
+ "learning_rate": 9.278620689655174e-06,
388
+ "loss": 0.1284,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 1.75,
393
+ "learning_rate": 9.261379310344828e-06,
394
+ "loss": 0.1484,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 1.77,
399
+ "learning_rate": 9.244137931034483e-06,
400
+ "loss": 0.1303,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 1.8,
405
+ "learning_rate": 9.226896551724138e-06,
406
+ "loss": 0.1743,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 1.83,
411
+ "learning_rate": 9.209655172413793e-06,
412
+ "loss": 0.1655,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 1.86,
417
+ "learning_rate": 9.192413793103448e-06,
418
+ "loss": 0.1327,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 1.88,
423
+ "learning_rate": 9.175172413793105e-06,
424
+ "loss": 0.1412,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 1.91,
429
+ "learning_rate": 9.15793103448276e-06,
430
+ "loss": 0.1346,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 1.94,
435
+ "learning_rate": 9.140689655172414e-06,
436
+ "loss": 0.1258,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 1.97,
441
+ "learning_rate": 9.12344827586207e-06,
442
+ "loss": 0.1702,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 2.0,
447
+ "learning_rate": 9.106206896551724e-06,
448
+ "loss": 0.1444,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 2.02,
453
+ "learning_rate": 9.08896551724138e-06,
454
+ "loss": 0.0881,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 2.05,
459
+ "learning_rate": 9.071724137931035e-06,
460
+ "loss": 0.0722,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 2.08,
465
+ "learning_rate": 9.05448275862069e-06,
466
+ "loss": 0.0726,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 2.11,
471
+ "learning_rate": 9.037241379310345e-06,
472
+ "loss": 0.0635,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 2.13,
477
+ "learning_rate": 9.020000000000002e-06,
478
+ "loss": 0.0673,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 2.16,
483
+ "learning_rate": 9.002758620689655e-06,
484
+ "loss": 0.0658,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 2.19,
489
+ "learning_rate": 8.985517241379311e-06,
490
+ "loss": 0.0817,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 2.22,
495
+ "learning_rate": 8.968275862068966e-06,
496
+ "loss": 0.0726,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 2.22,
501
+ "eval_loss": 0.2807031571865082,
502
+ "eval_runtime": 3432.3667,
503
+ "eval_samples_per_second": 0.496,
504
+ "eval_steps_per_second": 0.062,
505
+ "eval_wer": 110.51894717192458,
506
+ "step": 2000
507
+ },
508
+ {
509
+ "epoch": 2.25,
510
+ "learning_rate": 8.951034482758621e-06,
511
+ "loss": 0.0802,
512
+ "step": 2025
513
+ },
514
+ {
515
+ "epoch": 2.27,
516
+ "learning_rate": 8.933793103448276e-06,
517
+ "loss": 0.0702,
518
+ "step": 2050
519
+ },
520
+ {
521
+ "epoch": 2.3,
522
+ "learning_rate": 8.916551724137931e-06,
523
+ "loss": 0.0821,
524
+ "step": 2075
525
+ },
526
+ {
527
+ "epoch": 2.33,
528
+ "learning_rate": 8.899310344827588e-06,
529
+ "loss": 0.0825,
530
+ "step": 2100
531
+ },
532
+ {
533
+ "epoch": 2.36,
534
+ "learning_rate": 8.882068965517242e-06,
535
+ "loss": 0.0718,
536
+ "step": 2125
537
+ },
538
+ {
539
+ "epoch": 2.38,
540
+ "learning_rate": 8.864827586206897e-06,
541
+ "loss": 0.075,
542
+ "step": 2150
543
+ },
544
+ {
545
+ "epoch": 2.41,
546
+ "learning_rate": 8.847586206896552e-06,
547
+ "loss": 0.0818,
548
+ "step": 2175
549
+ },
550
+ {
551
+ "epoch": 2.44,
552
+ "learning_rate": 8.830344827586209e-06,
553
+ "loss": 0.0735,
554
+ "step": 2200
555
+ },
556
+ {
557
+ "epoch": 2.47,
558
+ "learning_rate": 8.813103448275862e-06,
559
+ "loss": 0.0636,
560
+ "step": 2225
561
+ },
562
+ {
563
+ "epoch": 2.49,
564
+ "learning_rate": 8.795862068965518e-06,
565
+ "loss": 0.0683,
566
+ "step": 2250
567
+ },
568
+ {
569
+ "epoch": 2.52,
570
+ "learning_rate": 8.778620689655173e-06,
571
+ "loss": 0.0775,
572
+ "step": 2275
573
+ },
574
+ {
575
+ "epoch": 2.55,
576
+ "learning_rate": 8.761379310344828e-06,
577
+ "loss": 0.072,
578
+ "step": 2300
579
+ },
580
+ {
581
+ "epoch": 2.58,
582
+ "learning_rate": 8.744137931034483e-06,
583
+ "loss": 0.0724,
584
+ "step": 2325
585
+ },
586
+ {
587
+ "epoch": 2.61,
588
+ "learning_rate": 8.72689655172414e-06,
589
+ "loss": 0.0658,
590
+ "step": 2350
591
+ },
592
+ {
593
+ "epoch": 2.63,
594
+ "learning_rate": 8.709655172413793e-06,
595
+ "loss": 0.0762,
596
+ "step": 2375
597
+ },
598
+ {
599
+ "epoch": 2.66,
600
+ "learning_rate": 8.69241379310345e-06,
601
+ "loss": 0.0886,
602
+ "step": 2400
603
+ },
604
+ {
605
+ "epoch": 2.69,
606
+ "learning_rate": 8.675172413793104e-06,
607
+ "loss": 0.0792,
608
+ "step": 2425
609
+ },
610
+ {
611
+ "epoch": 2.72,
612
+ "learning_rate": 8.657931034482759e-06,
613
+ "loss": 0.0667,
614
+ "step": 2450
615
+ },
616
+ {
617
+ "epoch": 2.74,
618
+ "learning_rate": 8.640689655172414e-06,
619
+ "loss": 0.0753,
620
+ "step": 2475
621
+ },
622
+ {
623
+ "epoch": 2.77,
624
+ "learning_rate": 8.623448275862069e-06,
625
+ "loss": 0.0765,
626
+ "step": 2500
627
+ },
628
+ {
629
+ "epoch": 2.8,
630
+ "learning_rate": 8.606206896551725e-06,
631
+ "loss": 0.0729,
632
+ "step": 2525
633
+ },
634
+ {
635
+ "epoch": 2.83,
636
+ "learning_rate": 8.58896551724138e-06,
637
+ "loss": 0.0687,
638
+ "step": 2550
639
+ },
640
+ {
641
+ "epoch": 2.85,
642
+ "learning_rate": 8.571724137931035e-06,
643
+ "loss": 0.0715,
644
+ "step": 2575
645
+ },
646
+ {
647
+ "epoch": 2.88,
648
+ "learning_rate": 8.55448275862069e-06,
649
+ "loss": 0.0777,
650
+ "step": 2600
651
+ },
652
+ {
653
+ "epoch": 2.91,
654
+ "learning_rate": 8.537241379310347e-06,
655
+ "loss": 0.0787,
656
+ "step": 2625
657
+ },
658
+ {
659
+ "epoch": 2.94,
660
+ "learning_rate": 8.52e-06,
661
+ "loss": 0.0659,
662
+ "step": 2650
663
+ },
664
+ {
665
+ "epoch": 2.97,
666
+ "learning_rate": 8.502758620689656e-06,
667
+ "loss": 0.0759,
668
+ "step": 2675
669
+ },
670
+ {
671
+ "epoch": 2.99,
672
+ "learning_rate": 8.485517241379311e-06,
673
+ "loss": 0.0714,
674
+ "step": 2700
675
+ },
676
+ {
677
+ "epoch": 3.02,
678
+ "learning_rate": 8.468275862068966e-06,
679
+ "loss": 0.053,
680
+ "step": 2725
681
+ },
682
+ {
683
+ "epoch": 3.05,
684
+ "learning_rate": 8.451034482758621e-06,
685
+ "loss": 0.0387,
686
+ "step": 2750
687
+ },
688
+ {
689
+ "epoch": 3.08,
690
+ "learning_rate": 8.433793103448277e-06,
691
+ "loss": 0.0339,
692
+ "step": 2775
693
+ },
694
+ {
695
+ "epoch": 3.1,
696
+ "learning_rate": 8.41655172413793e-06,
697
+ "loss": 0.0331,
698
+ "step": 2800
699
+ },
700
+ {
701
+ "epoch": 3.13,
702
+ "learning_rate": 8.399310344827587e-06,
703
+ "loss": 0.0363,
704
+ "step": 2825
705
+ },
706
+ {
707
+ "epoch": 3.16,
708
+ "learning_rate": 8.382068965517242e-06,
709
+ "loss": 0.0434,
710
+ "step": 2850
711
+ },
712
+ {
713
+ "epoch": 3.19,
714
+ "learning_rate": 8.364827586206897e-06,
715
+ "loss": 0.0376,
716
+ "step": 2875
717
+ },
718
+ {
719
+ "epoch": 3.22,
720
+ "learning_rate": 8.347586206896552e-06,
721
+ "loss": 0.0423,
722
+ "step": 2900
723
+ },
724
+ {
725
+ "epoch": 3.24,
726
+ "learning_rate": 8.330344827586208e-06,
727
+ "loss": 0.0407,
728
+ "step": 2925
729
+ },
730
+ {
731
+ "epoch": 3.27,
732
+ "learning_rate": 8.313103448275863e-06,
733
+ "loss": 0.0415,
734
+ "step": 2950
735
+ },
736
+ {
737
+ "epoch": 3.3,
738
+ "learning_rate": 8.295862068965518e-06,
739
+ "loss": 0.0379,
740
+ "step": 2975
741
+ },
742
+ {
743
+ "epoch": 3.33,
744
+ "learning_rate": 8.278620689655173e-06,
745
+ "loss": 0.0404,
746
+ "step": 3000
747
+ },
748
+ {
749
+ "epoch": 3.33,
750
+ "eval_loss": 0.3070586323738098,
751
+ "eval_runtime": 3252.4732,
752
+ "eval_samples_per_second": 0.524,
753
+ "eval_steps_per_second": 0.065,
754
+ "eval_wer": 135.31827515400408,
755
+ "step": 3000
756
+ },
757
+ {
758
+ "epoch": 3.35,
759
+ "learning_rate": 8.261379310344828e-06,
760
+ "loss": 0.0406,
761
+ "step": 3025
762
+ },
763
+ {
764
+ "epoch": 3.38,
765
+ "learning_rate": 8.244137931034484e-06,
766
+ "loss": 0.042,
767
+ "step": 3050
768
+ },
769
+ {
770
+ "epoch": 3.41,
771
+ "learning_rate": 8.226896551724138e-06,
772
+ "loss": 0.0429,
773
+ "step": 3075
774
+ },
775
+ {
776
+ "epoch": 3.44,
777
+ "learning_rate": 8.209655172413794e-06,
778
+ "loss": 0.0436,
779
+ "step": 3100
780
+ },
781
+ {
782
+ "epoch": 3.46,
783
+ "learning_rate": 8.192413793103449e-06,
784
+ "loss": 0.0427,
785
+ "step": 3125
786
+ },
787
+ {
788
+ "epoch": 3.49,
789
+ "learning_rate": 8.175172413793104e-06,
790
+ "loss": 0.0462,
791
+ "step": 3150
792
+ },
793
+ {
794
+ "epoch": 3.52,
795
+ "learning_rate": 8.157931034482759e-06,
796
+ "loss": 0.0376,
797
+ "step": 3175
798
+ },
799
+ {
800
+ "epoch": 3.55,
801
+ "learning_rate": 8.140689655172415e-06,
802
+ "loss": 0.0503,
803
+ "step": 3200
804
+ },
805
+ {
806
+ "epoch": 3.58,
807
+ "learning_rate": 8.123448275862069e-06,
808
+ "loss": 0.0403,
809
+ "step": 3225
810
+ },
811
+ {
812
+ "epoch": 3.6,
813
+ "learning_rate": 8.106206896551725e-06,
814
+ "loss": 0.0386,
815
+ "step": 3250
816
+ },
817
+ {
818
+ "epoch": 3.63,
819
+ "learning_rate": 8.08896551724138e-06,
820
+ "loss": 0.0437,
821
+ "step": 3275
822
+ },
823
+ {
824
+ "epoch": 3.66,
825
+ "learning_rate": 8.071724137931035e-06,
826
+ "loss": 0.0383,
827
+ "step": 3300
828
+ },
829
+ {
830
+ "epoch": 3.69,
831
+ "learning_rate": 8.054482758620691e-06,
832
+ "loss": 0.034,
833
+ "step": 3325
834
+ },
835
+ {
836
+ "epoch": 3.71,
837
+ "learning_rate": 8.037241379310346e-06,
838
+ "loss": 0.0412,
839
+ "step": 3350
840
+ },
841
+ {
842
+ "epoch": 3.74,
843
+ "learning_rate": 8.020000000000001e-06,
844
+ "loss": 0.0399,
845
+ "step": 3375
846
+ },
847
+ {
848
+ "epoch": 3.77,
849
+ "learning_rate": 8.002758620689656e-06,
850
+ "loss": 0.0479,
851
+ "step": 3400
852
+ },
853
+ {
854
+ "epoch": 3.8,
855
+ "learning_rate": 7.985517241379311e-06,
856
+ "loss": 0.0539,
857
+ "step": 3425
858
+ },
859
+ {
860
+ "epoch": 3.82,
861
+ "learning_rate": 7.968275862068966e-06,
862
+ "loss": 0.0385,
863
+ "step": 3450
864
+ },
865
+ {
866
+ "epoch": 3.85,
867
+ "learning_rate": 7.951034482758622e-06,
868
+ "loss": 0.0463,
869
+ "step": 3475
870
+ },
871
+ {
872
+ "epoch": 3.88,
873
+ "learning_rate": 7.933793103448275e-06,
874
+ "loss": 0.0328,
875
+ "step": 3500
876
+ },
877
+ {
878
+ "epoch": 3.91,
879
+ "learning_rate": 7.916551724137932e-06,
880
+ "loss": 0.042,
881
+ "step": 3525
882
+ },
883
+ {
884
+ "epoch": 3.94,
885
+ "learning_rate": 7.899310344827587e-06,
886
+ "loss": 0.0413,
887
+ "step": 3550
888
+ },
889
+ {
890
+ "epoch": 3.96,
891
+ "learning_rate": 7.882068965517242e-06,
892
+ "loss": 0.0383,
893
+ "step": 3575
894
+ },
895
+ {
896
+ "epoch": 3.99,
897
+ "learning_rate": 7.864827586206897e-06,
898
+ "loss": 0.0487,
899
+ "step": 3600
900
+ },
901
+ {
902
+ "epoch": 4.02,
903
+ "learning_rate": 7.847586206896553e-06,
904
+ "loss": 0.0187,
905
+ "step": 3625
906
+ },
907
+ {
908
+ "epoch": 4.05,
909
+ "learning_rate": 7.830344827586206e-06,
910
+ "loss": 0.0174,
911
+ "step": 3650
912
+ },
913
+ {
914
+ "epoch": 4.07,
915
+ "learning_rate": 7.813103448275863e-06,
916
+ "loss": 0.0206,
917
+ "step": 3675
918
+ },
919
+ {
920
+ "epoch": 4.1,
921
+ "learning_rate": 7.795862068965518e-06,
922
+ "loss": 0.0293,
923
+ "step": 3700
924
+ },
925
+ {
926
+ "epoch": 4.13,
927
+ "learning_rate": 7.778620689655173e-06,
928
+ "loss": 0.02,
929
+ "step": 3725
930
+ },
931
+ {
932
+ "epoch": 4.16,
933
+ "learning_rate": 7.76137931034483e-06,
934
+ "loss": 0.0195,
935
+ "step": 3750
936
+ },
937
+ {
938
+ "epoch": 4.19,
939
+ "learning_rate": 7.744137931034484e-06,
940
+ "loss": 0.0205,
941
+ "step": 3775
942
+ },
943
+ {
944
+ "epoch": 4.21,
945
+ "learning_rate": 7.726896551724139e-06,
946
+ "loss": 0.0191,
947
+ "step": 3800
948
+ },
949
+ {
950
+ "epoch": 4.24,
951
+ "learning_rate": 7.709655172413794e-06,
952
+ "loss": 0.016,
953
+ "step": 3825
954
+ },
955
+ {
956
+ "epoch": 4.27,
957
+ "learning_rate": 7.692413793103449e-06,
958
+ "loss": 0.0232,
959
+ "step": 3850
960
+ },
961
+ {
962
+ "epoch": 4.3,
963
+ "learning_rate": 7.675172413793104e-06,
964
+ "loss": 0.0177,
965
+ "step": 3875
966
+ },
967
+ {
968
+ "epoch": 4.32,
969
+ "learning_rate": 7.65793103448276e-06,
970
+ "loss": 0.0233,
971
+ "step": 3900
972
+ },
973
+ {
974
+ "epoch": 4.35,
975
+ "learning_rate": 7.640689655172413e-06,
976
+ "loss": 0.0223,
977
+ "step": 3925
978
+ },
979
+ {
980
+ "epoch": 4.38,
981
+ "learning_rate": 7.62344827586207e-06,
982
+ "loss": 0.027,
983
+ "step": 3950
984
+ },
985
+ {
986
+ "epoch": 4.41,
987
+ "learning_rate": 7.606206896551725e-06,
988
+ "loss": 0.0246,
989
+ "step": 3975
990
+ },
991
+ {
992
+ "epoch": 4.43,
993
+ "learning_rate": 7.58896551724138e-06,
994
+ "loss": 0.0247,
995
+ "step": 4000
996
+ },
997
+ {
998
+ "epoch": 4.43,
999
+ "eval_loss": 0.32613393664360046,
1000
+ "eval_runtime": 3013.158,
1001
+ "eval_samples_per_second": 0.566,
1002
+ "eval_steps_per_second": 0.071,
1003
+ "eval_wer": 69.33918237819675,
1004
+ "step": 4000
1005
+ },
1006
+ {
1007
+ "epoch": 4.46,
1008
+ "learning_rate": 7.5717241379310345e-06,
1009
+ "loss": 0.0206,
1010
+ "step": 4025
1011
+ },
1012
+ {
1013
+ "epoch": 4.49,
1014
+ "learning_rate": 7.55448275862069e-06,
1015
+ "loss": 0.0214,
1016
+ "step": 4050
1017
+ },
1018
+ {
1019
+ "epoch": 4.52,
1020
+ "learning_rate": 7.537241379310345e-06,
1021
+ "loss": 0.0289,
1022
+ "step": 4075
1023
+ },
1024
+ {
1025
+ "epoch": 4.55,
1026
+ "learning_rate": 7.520000000000001e-06,
1027
+ "loss": 0.0182,
1028
+ "step": 4100
1029
+ },
1030
+ {
1031
+ "epoch": 4.57,
1032
+ "learning_rate": 7.5027586206896566e-06,
1033
+ "loss": 0.0269,
1034
+ "step": 4125
1035
+ },
1036
+ {
1037
+ "epoch": 4.6,
1038
+ "learning_rate": 7.485517241379311e-06,
1039
+ "loss": 0.0273,
1040
+ "step": 4150
1041
+ },
1042
+ {
1043
+ "epoch": 4.63,
1044
+ "learning_rate": 7.468275862068966e-06,
1045
+ "loss": 0.0319,
1046
+ "step": 4175
1047
+ },
1048
+ {
1049
+ "epoch": 4.66,
1050
+ "learning_rate": 7.451724137931035e-06,
1051
+ "loss": 0.0293,
1052
+ "step": 4200
1053
+ },
1054
+ {
1055
+ "epoch": 4.68,
1056
+ "learning_rate": 7.4344827586206906e-06,
1057
+ "loss": 0.0226,
1058
+ "step": 4225
1059
+ },
1060
+ {
1061
+ "epoch": 4.71,
1062
+ "learning_rate": 7.417241379310346e-06,
1063
+ "loss": 0.0231,
1064
+ "step": 4250
1065
+ },
1066
+ {
1067
+ "epoch": 4.74,
1068
+ "learning_rate": 7.4e-06,
1069
+ "loss": 0.0199,
1070
+ "step": 4275
1071
+ },
1072
+ {
1073
+ "epoch": 4.77,
1074
+ "learning_rate": 7.382758620689656e-06,
1075
+ "loss": 0.0228,
1076
+ "step": 4300
1077
+ },
1078
+ {
1079
+ "epoch": 4.79,
1080
+ "learning_rate": 7.365517241379311e-06,
1081
+ "loss": 0.0303,
1082
+ "step": 4325
1083
+ },
1084
+ {
1085
+ "epoch": 4.82,
1086
+ "learning_rate": 7.348275862068967e-06,
1087
+ "loss": 0.023,
1088
+ "step": 4350
1089
+ },
1090
+ {
1091
+ "epoch": 4.85,
1092
+ "learning_rate": 7.3310344827586215e-06,
1093
+ "loss": 0.0297,
1094
+ "step": 4375
1095
+ },
1096
+ {
1097
+ "epoch": 4.88,
1098
+ "learning_rate": 7.313793103448276e-06,
1099
+ "loss": 0.0409,
1100
+ "step": 4400
1101
+ },
1102
+ {
1103
+ "epoch": 4.91,
1104
+ "learning_rate": 7.296551724137931e-06,
1105
+ "loss": 0.0179,
1106
+ "step": 4425
1107
+ },
1108
+ {
1109
+ "epoch": 4.93,
1110
+ "learning_rate": 7.279310344827587e-06,
1111
+ "loss": 0.0288,
1112
+ "step": 4450
1113
+ },
1114
+ {
1115
+ "epoch": 4.96,
1116
+ "learning_rate": 7.262068965517242e-06,
1117
+ "loss": 0.0339,
1118
+ "step": 4475
1119
+ },
1120
+ {
1121
+ "epoch": 4.99,
1122
+ "learning_rate": 7.2448275862068975e-06,
1123
+ "loss": 0.0278,
1124
+ "step": 4500
1125
+ },
1126
+ {
1127
+ "epoch": 5.02,
1128
+ "learning_rate": 7.2275862068965515e-06,
1129
+ "loss": 0.0154,
1130
+ "step": 4525
1131
+ },
1132
+ {
1133
+ "epoch": 5.04,
1134
+ "learning_rate": 7.210344827586207e-06,
1135
+ "loss": 0.0091,
1136
+ "step": 4550
1137
+ },
1138
+ {
1139
+ "epoch": 5.07,
1140
+ "learning_rate": 7.193103448275862e-06,
1141
+ "loss": 0.0105,
1142
+ "step": 4575
1143
+ },
1144
+ {
1145
+ "epoch": 5.1,
1146
+ "learning_rate": 7.175862068965518e-06,
1147
+ "loss": 0.0122,
1148
+ "step": 4600
1149
+ },
1150
+ {
1151
+ "epoch": 5.13,
1152
+ "learning_rate": 7.1586206896551736e-06,
1153
+ "loss": 0.0102,
1154
+ "step": 4625
1155
+ },
1156
+ {
1157
+ "epoch": 5.16,
1158
+ "learning_rate": 7.1413793103448284e-06,
1159
+ "loss": 0.0111,
1160
+ "step": 4650
1161
+ },
1162
+ {
1163
+ "epoch": 5.18,
1164
+ "learning_rate": 7.124137931034484e-06,
1165
+ "loss": 0.0102,
1166
+ "step": 4675
1167
+ },
1168
+ {
1169
+ "epoch": 5.21,
1170
+ "learning_rate": 7.106896551724138e-06,
1171
+ "loss": 0.0154,
1172
+ "step": 4700
1173
+ },
1174
+ {
1175
+ "epoch": 5.24,
1176
+ "learning_rate": 7.089655172413794e-06,
1177
+ "loss": 0.0145,
1178
+ "step": 4725
1179
+ },
1180
+ {
1181
+ "epoch": 5.27,
1182
+ "learning_rate": 7.072413793103449e-06,
1183
+ "loss": 0.0159,
1184
+ "step": 4750
1185
+ },
1186
+ {
1187
+ "epoch": 5.29,
1188
+ "learning_rate": 7.0551724137931045e-06,
1189
+ "loss": 0.0073,
1190
+ "step": 4775
1191
+ },
1192
+ {
1193
+ "epoch": 5.32,
1194
+ "learning_rate": 7.037931034482759e-06,
1195
+ "loss": 0.008,
1196
+ "step": 4800
1197
+ },
1198
+ {
1199
+ "epoch": 5.35,
1200
+ "learning_rate": 7.020689655172414e-06,
1201
+ "loss": 0.0117,
1202
+ "step": 4825
1203
+ },
1204
+ {
1205
+ "epoch": 5.38,
1206
+ "learning_rate": 7.003448275862069e-06,
1207
+ "loss": 0.0109,
1208
+ "step": 4850
1209
+ },
1210
+ {
1211
+ "epoch": 5.4,
1212
+ "learning_rate": 6.986206896551725e-06,
1213
+ "loss": 0.0192,
1214
+ "step": 4875
1215
+ },
1216
+ {
1217
+ "epoch": 5.43,
1218
+ "learning_rate": 6.96896551724138e-06,
1219
+ "loss": 0.0077,
1220
+ "step": 4900
1221
+ },
1222
+ {
1223
+ "epoch": 5.46,
1224
+ "learning_rate": 6.951724137931035e-06,
1225
+ "loss": 0.0114,
1226
+ "step": 4925
1227
+ },
1228
+ {
1229
+ "epoch": 5.49,
1230
+ "learning_rate": 6.934482758620689e-06,
1231
+ "loss": 0.0118,
1232
+ "step": 4950
1233
+ },
1234
+ {
1235
+ "epoch": 5.52,
1236
+ "learning_rate": 6.917241379310345e-06,
1237
+ "loss": 0.0133,
1238
+ "step": 4975
1239
+ },
1240
+ {
1241
+ "epoch": 5.54,
1242
+ "learning_rate": 6.9e-06,
1243
+ "loss": 0.0102,
1244
+ "step": 5000
1245
+ },
1246
+ {
1247
+ "epoch": 5.54,
1248
+ "eval_loss": 0.37662339210510254,
1249
+ "eval_runtime": 3307.2582,
1250
+ "eval_samples_per_second": 0.515,
1251
+ "eval_steps_per_second": 0.064,
1252
+ "eval_wer": 99.71065895090535,
1253
+ "step": 5000
1254
+ },
1255
+ {
1256
+ "epoch": 5.57,
1257
+ "learning_rate": 6.882758620689656e-06,
1258
+ "loss": 0.0075,
1259
+ "step": 5025
1260
+ },
1261
+ {
1262
+ "epoch": 5.6,
1263
+ "learning_rate": 6.8655172413793114e-06,
1264
+ "loss": 0.0238,
1265
+ "step": 5050
1266
+ },
1267
+ {
1268
+ "epoch": 5.63,
1269
+ "learning_rate": 6.848275862068966e-06,
1270
+ "loss": 0.0124,
1271
+ "step": 5075
1272
+ },
1273
+ {
1274
+ "epoch": 5.65,
1275
+ "learning_rate": 6.831034482758622e-06,
1276
+ "loss": 0.0132,
1277
+ "step": 5100
1278
+ },
1279
+ {
1280
+ "epoch": 5.68,
1281
+ "learning_rate": 6.813793103448276e-06,
1282
+ "loss": 0.0257,
1283
+ "step": 5125
1284
+ },
1285
+ {
1286
+ "epoch": 5.71,
1287
+ "learning_rate": 6.796551724137932e-06,
1288
+ "loss": 0.0153,
1289
+ "step": 5150
1290
+ },
1291
+ {
1292
+ "epoch": 5.74,
1293
+ "learning_rate": 6.779310344827587e-06,
1294
+ "loss": 0.0154,
1295
+ "step": 5175
1296
+ },
1297
+ {
1298
+ "epoch": 5.76,
1299
+ "learning_rate": 6.762068965517242e-06,
1300
+ "loss": 0.0098,
1301
+ "step": 5200
1302
+ },
1303
+ {
1304
+ "epoch": 5.79,
1305
+ "learning_rate": 6.744827586206897e-06,
1306
+ "loss": 0.0175,
1307
+ "step": 5225
1308
+ },
1309
+ {
1310
+ "epoch": 5.82,
1311
+ "learning_rate": 6.727586206896552e-06,
1312
+ "loss": 0.0142,
1313
+ "step": 5250
1314
+ },
1315
+ {
1316
+ "epoch": 5.85,
1317
+ "learning_rate": 6.710344827586207e-06,
1318
+ "loss": 0.0193,
1319
+ "step": 5275
1320
+ },
1321
+ {
1322
+ "epoch": 5.88,
1323
+ "learning_rate": 6.693103448275863e-06,
1324
+ "loss": 0.0133,
1325
+ "step": 5300
1326
+ },
1327
+ {
1328
+ "epoch": 5.9,
1329
+ "learning_rate": 6.6758620689655176e-06,
1330
+ "loss": 0.0061,
1331
+ "step": 5325
1332
+ },
1333
+ {
1334
+ "epoch": 5.93,
1335
+ "learning_rate": 6.658620689655173e-06,
1336
+ "loss": 0.0127,
1337
+ "step": 5350
1338
+ },
1339
+ {
1340
+ "epoch": 5.96,
1341
+ "learning_rate": 6.641379310344827e-06,
1342
+ "loss": 0.0109,
1343
+ "step": 5375
1344
+ },
1345
+ {
1346
+ "epoch": 5.99,
1347
+ "learning_rate": 6.624137931034483e-06,
1348
+ "loss": 0.0135,
1349
+ "step": 5400
1350
+ },
1351
+ {
1352
+ "epoch": 6.01,
1353
+ "learning_rate": 6.606896551724139e-06,
1354
+ "loss": 0.0113,
1355
+ "step": 5425
1356
+ },
1357
+ {
1358
+ "epoch": 6.04,
1359
+ "learning_rate": 6.589655172413794e-06,
1360
+ "loss": 0.006,
1361
+ "step": 5450
1362
+ },
1363
+ {
1364
+ "epoch": 6.07,
1365
+ "learning_rate": 6.572413793103449e-06,
1366
+ "loss": 0.0089,
1367
+ "step": 5475
1368
+ },
1369
+ {
1370
+ "epoch": 6.1,
1371
+ "learning_rate": 6.555172413793104e-06,
1372
+ "loss": 0.007,
1373
+ "step": 5500
1374
+ },
1375
+ {
1376
+ "epoch": 6.13,
1377
+ "learning_rate": 6.53793103448276e-06,
1378
+ "loss": 0.0072,
1379
+ "step": 5525
1380
+ },
1381
+ {
1382
+ "epoch": 6.15,
1383
+ "learning_rate": 6.520689655172414e-06,
1384
+ "loss": 0.0063,
1385
+ "step": 5550
1386
+ },
1387
+ {
1388
+ "epoch": 6.18,
1389
+ "learning_rate": 6.50344827586207e-06,
1390
+ "loss": 0.008,
1391
+ "step": 5575
1392
+ },
1393
+ {
1394
+ "epoch": 6.21,
1395
+ "learning_rate": 6.4862068965517245e-06,
1396
+ "loss": 0.0111,
1397
+ "step": 5600
1398
+ },
1399
+ {
1400
+ "epoch": 6.24,
1401
+ "learning_rate": 6.46896551724138e-06,
1402
+ "loss": 0.0042,
1403
+ "step": 5625
1404
+ },
1405
+ {
1406
+ "epoch": 6.26,
1407
+ "learning_rate": 6.451724137931035e-06,
1408
+ "loss": 0.005,
1409
+ "step": 5650
1410
+ },
1411
+ {
1412
+ "epoch": 6.29,
1413
+ "learning_rate": 6.43448275862069e-06,
1414
+ "loss": 0.003,
1415
+ "step": 5675
1416
+ },
1417
+ {
1418
+ "epoch": 6.32,
1419
+ "learning_rate": 6.417241379310345e-06,
1420
+ "loss": 0.0048,
1421
+ "step": 5700
1422
+ },
1423
+ {
1424
+ "epoch": 6.35,
1425
+ "learning_rate": 6.4000000000000006e-06,
1426
+ "loss": 0.0034,
1427
+ "step": 5725
1428
+ },
1429
+ {
1430
+ "epoch": 6.37,
1431
+ "learning_rate": 6.3827586206896554e-06,
1432
+ "loss": 0.014,
1433
+ "step": 5750
1434
+ },
1435
+ {
1436
+ "epoch": 6.4,
1437
+ "learning_rate": 6.365517241379311e-06,
1438
+ "loss": 0.0035,
1439
+ "step": 5775
1440
+ },
1441
+ {
1442
+ "epoch": 6.43,
1443
+ "learning_rate": 6.348275862068965e-06,
1444
+ "loss": 0.005,
1445
+ "step": 5800
1446
+ },
1447
+ {
1448
+ "epoch": 6.46,
1449
+ "learning_rate": 6.331034482758621e-06,
1450
+ "loss": 0.0033,
1451
+ "step": 5825
1452
+ },
1453
+ {
1454
+ "epoch": 6.49,
1455
+ "learning_rate": 6.313793103448277e-06,
1456
+ "loss": 0.0057,
1457
+ "step": 5850
1458
+ },
1459
+ {
1460
+ "epoch": 6.51,
1461
+ "learning_rate": 6.2965517241379315e-06,
1462
+ "loss": 0.009,
1463
+ "step": 5875
1464
+ },
1465
+ {
1466
+ "epoch": 6.54,
1467
+ "learning_rate": 6.279310344827587e-06,
1468
+ "loss": 0.0075,
1469
+ "step": 5900
1470
+ },
1471
+ {
1472
+ "epoch": 6.57,
1473
+ "learning_rate": 6.262068965517242e-06,
1474
+ "loss": 0.0045,
1475
+ "step": 5925
1476
+ },
1477
+ {
1478
+ "epoch": 6.6,
1479
+ "learning_rate": 6.244827586206898e-06,
1480
+ "loss": 0.0021,
1481
+ "step": 5950
1482
+ },
1483
+ {
1484
+ "epoch": 6.62,
1485
+ "learning_rate": 6.227586206896552e-06,
1486
+ "loss": 0.0061,
1487
+ "step": 5975
1488
+ },
1489
+ {
1490
+ "epoch": 6.65,
1491
+ "learning_rate": 6.2103448275862075e-06,
1492
+ "loss": 0.0096,
1493
+ "step": 6000
1494
+ },
1495
+ {
1496
+ "epoch": 6.65,
1497
+ "eval_loss": 0.4006502330303192,
1498
+ "eval_runtime": 2903.7003,
1499
+ "eval_samples_per_second": 0.587,
1500
+ "eval_steps_per_second": 0.073,
1501
+ "eval_wer": 83.9835728952772,
1502
+ "step": 6000
1503
+ },
1504
+ {
1505
+ "epoch": 6.68,
1506
+ "learning_rate": 6.193103448275862e-06,
1507
+ "loss": 0.0041,
1508
+ "step": 6025
1509
+ },
1510
+ {
1511
+ "epoch": 6.71,
1512
+ "learning_rate": 6.175862068965518e-06,
1513
+ "loss": 0.0089,
1514
+ "step": 6050
1515
+ },
1516
+ {
1517
+ "epoch": 6.74,
1518
+ "learning_rate": 6.158620689655173e-06,
1519
+ "loss": 0.0055,
1520
+ "step": 6075
1521
+ },
1522
+ {
1523
+ "epoch": 6.76,
1524
+ "learning_rate": 6.141379310344828e-06,
1525
+ "loss": 0.0073,
1526
+ "step": 6100
1527
+ },
1528
+ {
1529
+ "epoch": 6.79,
1530
+ "learning_rate": 6.124137931034483e-06,
1531
+ "loss": 0.0069,
1532
+ "step": 6125
1533
+ },
1534
+ {
1535
+ "epoch": 6.82,
1536
+ "learning_rate": 6.1068965517241384e-06,
1537
+ "loss": 0.0059,
1538
+ "step": 6150
1539
+ },
1540
+ {
1541
+ "epoch": 6.85,
1542
+ "learning_rate": 6.089655172413793e-06,
1543
+ "loss": 0.0071,
1544
+ "step": 6175
1545
+ },
1546
+ {
1547
+ "epoch": 6.87,
1548
+ "learning_rate": 6.072413793103449e-06,
1549
+ "loss": 0.0053,
1550
+ "step": 6200
1551
+ },
1552
+ {
1553
+ "epoch": 6.9,
1554
+ "learning_rate": 6.055172413793105e-06,
1555
+ "loss": 0.0074,
1556
+ "step": 6225
1557
+ },
1558
+ {
1559
+ "epoch": 6.93,
1560
+ "learning_rate": 6.037931034482759e-06,
1561
+ "loss": 0.007,
1562
+ "step": 6250
1563
+ },
1564
+ {
1565
+ "epoch": 6.96,
1566
+ "learning_rate": 6.0206896551724145e-06,
1567
+ "loss": 0.0079,
1568
+ "step": 6275
1569
+ },
1570
+ {
1571
+ "epoch": 6.98,
1572
+ "learning_rate": 6.003448275862069e-06,
1573
+ "loss": 0.0111,
1574
+ "step": 6300
1575
+ },
1576
+ {
1577
+ "epoch": 7.01,
1578
+ "learning_rate": 5.986206896551725e-06,
1579
+ "loss": 0.0052,
1580
+ "step": 6325
1581
+ },
1582
+ {
1583
+ "epoch": 7.04,
1584
+ "learning_rate": 5.96896551724138e-06,
1585
+ "loss": 0.0055,
1586
+ "step": 6350
1587
+ },
1588
+ {
1589
+ "epoch": 7.07,
1590
+ "learning_rate": 5.951724137931036e-06,
1591
+ "loss": 0.0035,
1592
+ "step": 6375
1593
+ },
1594
+ {
1595
+ "epoch": 7.1,
1596
+ "learning_rate": 5.93448275862069e-06,
1597
+ "loss": 0.0031,
1598
+ "step": 6400
1599
+ },
1600
+ {
1601
+ "epoch": 7.12,
1602
+ "learning_rate": 5.917241379310345e-06,
1603
+ "loss": 0.0015,
1604
+ "step": 6425
1605
+ },
1606
+ {
1607
+ "epoch": 7.15,
1608
+ "learning_rate": 5.9e-06,
1609
+ "loss": 0.0055,
1610
+ "step": 6450
1611
+ },
1612
+ {
1613
+ "epoch": 7.18,
1614
+ "learning_rate": 5.882758620689656e-06,
1615
+ "loss": 0.0128,
1616
+ "step": 6475
1617
+ },
1618
+ {
1619
+ "epoch": 7.21,
1620
+ "learning_rate": 5.865517241379311e-06,
1621
+ "loss": 0.0028,
1622
+ "step": 6500
1623
+ },
1624
+ {
1625
+ "epoch": 7.23,
1626
+ "learning_rate": 5.848275862068966e-06,
1627
+ "loss": 0.0025,
1628
+ "step": 6525
1629
+ },
1630
+ {
1631
+ "epoch": 7.26,
1632
+ "learning_rate": 5.831034482758621e-06,
1633
+ "loss": 0.002,
1634
+ "step": 6550
1635
+ },
1636
+ {
1637
+ "epoch": 7.29,
1638
+ "learning_rate": 5.813793103448276e-06,
1639
+ "loss": 0.0061,
1640
+ "step": 6575
1641
+ },
1642
+ {
1643
+ "epoch": 7.32,
1644
+ "learning_rate": 5.796551724137931e-06,
1645
+ "loss": 0.001,
1646
+ "step": 6600
1647
+ },
1648
+ {
1649
+ "epoch": 7.34,
1650
+ "learning_rate": 5.779310344827587e-06,
1651
+ "loss": 0.0021,
1652
+ "step": 6625
1653
+ },
1654
+ {
1655
+ "epoch": 7.37,
1656
+ "learning_rate": 5.762068965517243e-06,
1657
+ "loss": 0.0008,
1658
+ "step": 6650
1659
+ },
1660
+ {
1661
+ "epoch": 7.4,
1662
+ "learning_rate": 5.744827586206897e-06,
1663
+ "loss": 0.0048,
1664
+ "step": 6675
1665
+ },
1666
+ {
1667
+ "epoch": 7.43,
1668
+ "learning_rate": 5.727586206896552e-06,
1669
+ "loss": 0.0017,
1670
+ "step": 6700
1671
+ },
1672
+ {
1673
+ "epoch": 7.46,
1674
+ "learning_rate": 5.710344827586207e-06,
1675
+ "loss": 0.0032,
1676
+ "step": 6725
1677
+ },
1678
+ {
1679
+ "epoch": 7.48,
1680
+ "learning_rate": 5.693103448275863e-06,
1681
+ "loss": 0.0033,
1682
+ "step": 6750
1683
+ },
1684
+ {
1685
+ "epoch": 7.51,
1686
+ "learning_rate": 5.675862068965518e-06,
1687
+ "loss": 0.0093,
1688
+ "step": 6775
1689
+ },
1690
+ {
1691
+ "epoch": 7.54,
1692
+ "learning_rate": 5.6586206896551735e-06,
1693
+ "loss": 0.002,
1694
+ "step": 6800
1695
+ },
1696
+ {
1697
+ "epoch": 7.57,
1698
+ "learning_rate": 5.6413793103448275e-06,
1699
+ "loss": 0.0036,
1700
+ "step": 6825
1701
+ },
1702
+ {
1703
+ "epoch": 7.59,
1704
+ "learning_rate": 5.624137931034483e-06,
1705
+ "loss": 0.0066,
1706
+ "step": 6850
1707
+ },
1708
+ {
1709
+ "epoch": 7.62,
1710
+ "learning_rate": 5.606896551724138e-06,
1711
+ "loss": 0.0009,
1712
+ "step": 6875
1713
+ },
1714
+ {
1715
+ "epoch": 7.65,
1716
+ "learning_rate": 5.589655172413794e-06,
1717
+ "loss": 0.0007,
1718
+ "step": 6900
1719
+ },
1720
+ {
1721
+ "epoch": 7.68,
1722
+ "learning_rate": 5.572413793103449e-06,
1723
+ "loss": 0.0035,
1724
+ "step": 6925
1725
+ },
1726
+ {
1727
+ "epoch": 7.71,
1728
+ "learning_rate": 5.555172413793104e-06,
1729
+ "loss": 0.0018,
1730
+ "step": 6950
1731
+ },
1732
+ {
1733
+ "epoch": 7.73,
1734
+ "learning_rate": 5.5379310344827585e-06,
1735
+ "loss": 0.0061,
1736
+ "step": 6975
1737
+ },
1738
+ {
1739
+ "epoch": 7.76,
1740
+ "learning_rate": 5.520689655172414e-06,
1741
+ "loss": 0.0098,
1742
+ "step": 7000
1743
+ },
1744
+ {
1745
+ "epoch": 7.76,
1746
+ "eval_loss": 0.42961812019348145,
1747
+ "eval_runtime": 2962.2311,
1748
+ "eval_samples_per_second": 0.575,
1749
+ "eval_steps_per_second": 0.072,
1750
+ "eval_wer": 97.61993653164083,
1751
+ "step": 7000
1752
+ },
1753
+ {
1754
+ "epoch": 7.79,
1755
+ "learning_rate": 5.503448275862069e-06,
1756
+ "loss": 0.0019,
1757
+ "step": 7025
1758
+ },
1759
+ {
1760
+ "epoch": 7.82,
1761
+ "learning_rate": 5.486206896551725e-06,
1762
+ "loss": 0.0106,
1763
+ "step": 7050
1764
+ },
1765
+ {
1766
+ "epoch": 7.84,
1767
+ "learning_rate": 5.4689655172413805e-06,
1768
+ "loss": 0.0074,
1769
+ "step": 7075
1770
+ },
1771
+ {
1772
+ "epoch": 7.87,
1773
+ "learning_rate": 5.4517241379310345e-06,
1774
+ "loss": 0.0044,
1775
+ "step": 7100
1776
+ },
1777
+ {
1778
+ "epoch": 7.9,
1779
+ "learning_rate": 5.43448275862069e-06,
1780
+ "loss": 0.0065,
1781
+ "step": 7125
1782
+ },
1783
+ {
1784
+ "epoch": 7.93,
1785
+ "learning_rate": 5.417241379310345e-06,
1786
+ "loss": 0.0072,
1787
+ "step": 7150
1788
+ },
1789
+ {
1790
+ "epoch": 7.95,
1791
+ "learning_rate": 5.400000000000001e-06,
1792
+ "loss": 0.0049,
1793
+ "step": 7175
1794
+ },
1795
+ {
1796
+ "epoch": 7.98,
1797
+ "learning_rate": 5.382758620689656e-06,
1798
+ "loss": 0.0041,
1799
+ "step": 7200
1800
+ },
1801
+ {
1802
+ "epoch": 8.01,
1803
+ "learning_rate": 5.365517241379311e-06,
1804
+ "loss": 0.0045,
1805
+ "step": 7225
1806
+ },
1807
+ {
1808
+ "epoch": 8.04,
1809
+ "learning_rate": 5.3482758620689654e-06,
1810
+ "loss": 0.0013,
1811
+ "step": 7250
1812
+ },
1813
+ {
1814
+ "epoch": 8.07,
1815
+ "learning_rate": 5.331034482758621e-06,
1816
+ "loss": 0.0011,
1817
+ "step": 7275
1818
+ },
1819
+ {
1820
+ "epoch": 8.09,
1821
+ "learning_rate": 5.313793103448276e-06,
1822
+ "loss": 0.0015,
1823
+ "step": 7300
1824
+ },
1825
+ {
1826
+ "epoch": 8.12,
1827
+ "learning_rate": 5.296551724137932e-06,
1828
+ "loss": 0.0026,
1829
+ "step": 7325
1830
+ },
1831
+ {
1832
+ "epoch": 8.15,
1833
+ "learning_rate": 5.279310344827587e-06,
1834
+ "loss": 0.0082,
1835
+ "step": 7350
1836
+ },
1837
+ {
1838
+ "epoch": 8.18,
1839
+ "learning_rate": 5.2620689655172415e-06,
1840
+ "loss": 0.0022,
1841
+ "step": 7375
1842
+ },
1843
+ {
1844
+ "epoch": 8.2,
1845
+ "learning_rate": 5.244827586206896e-06,
1846
+ "loss": 0.0026,
1847
+ "step": 7400
1848
+ },
1849
+ {
1850
+ "epoch": 8.23,
1851
+ "learning_rate": 5.227586206896552e-06,
1852
+ "loss": 0.0034,
1853
+ "step": 7425
1854
+ },
1855
+ {
1856
+ "epoch": 8.26,
1857
+ "learning_rate": 5.210344827586208e-06,
1858
+ "loss": 0.0017,
1859
+ "step": 7450
1860
+ },
1861
+ {
1862
+ "epoch": 8.29,
1863
+ "learning_rate": 5.193103448275863e-06,
1864
+ "loss": 0.0042,
1865
+ "step": 7475
1866
+ },
1867
+ {
1868
+ "epoch": 8.31,
1869
+ "learning_rate": 5.175862068965518e-06,
1870
+ "loss": 0.002,
1871
+ "step": 7500
1872
+ },
1873
+ {
1874
+ "epoch": 8.34,
1875
+ "learning_rate": 5.158620689655172e-06,
1876
+ "loss": 0.0032,
1877
+ "step": 7525
1878
+ },
1879
+ {
1880
+ "epoch": 8.37,
1881
+ "learning_rate": 5.141379310344828e-06,
1882
+ "loss": 0.0027,
1883
+ "step": 7550
1884
+ },
1885
+ {
1886
+ "epoch": 8.4,
1887
+ "learning_rate": 5.124137931034483e-06,
1888
+ "loss": 0.0013,
1889
+ "step": 7575
1890
+ },
1891
+ {
1892
+ "epoch": 8.43,
1893
+ "learning_rate": 5.106896551724139e-06,
1894
+ "loss": 0.0026,
1895
+ "step": 7600
1896
+ },
1897
+ {
1898
+ "epoch": 8.45,
1899
+ "learning_rate": 5.0896551724137936e-06,
1900
+ "loss": 0.0007,
1901
+ "step": 7625
1902
+ },
1903
+ {
1904
+ "epoch": 8.48,
1905
+ "learning_rate": 5.072413793103449e-06,
1906
+ "loss": 0.0014,
1907
+ "step": 7650
1908
+ },
1909
+ {
1910
+ "epoch": 8.51,
1911
+ "learning_rate": 5.055172413793103e-06,
1912
+ "loss": 0.0039,
1913
+ "step": 7675
1914
+ },
1915
+ {
1916
+ "epoch": 8.54,
1917
+ "learning_rate": 5.037931034482759e-06,
1918
+ "loss": 0.0012,
1919
+ "step": 7700
1920
+ },
1921
+ {
1922
+ "epoch": 8.56,
1923
+ "learning_rate": 5.020689655172414e-06,
1924
+ "loss": 0.0004,
1925
+ "step": 7725
1926
+ },
1927
+ {
1928
+ "epoch": 8.59,
1929
+ "learning_rate": 5.00344827586207e-06,
1930
+ "loss": 0.0009,
1931
+ "step": 7750
1932
+ },
1933
+ {
1934
+ "epoch": 8.62,
1935
+ "learning_rate": 4.9862068965517245e-06,
1936
+ "loss": 0.0044,
1937
+ "step": 7775
1938
+ },
1939
+ {
1940
+ "epoch": 8.65,
1941
+ "learning_rate": 4.968965517241379e-06,
1942
+ "loss": 0.0026,
1943
+ "step": 7800
1944
+ },
1945
+ {
1946
+ "epoch": 8.68,
1947
+ "learning_rate": 4.951724137931035e-06,
1948
+ "loss": 0.0011,
1949
+ "step": 7825
1950
+ },
1951
+ {
1952
+ "epoch": 8.7,
1953
+ "learning_rate": 4.93448275862069e-06,
1954
+ "loss": 0.005,
1955
+ "step": 7850
1956
+ },
1957
+ {
1958
+ "epoch": 8.73,
1959
+ "learning_rate": 4.917241379310345e-06,
1960
+ "loss": 0.0014,
1961
+ "step": 7875
1962
+ },
1963
+ {
1964
+ "epoch": 8.76,
1965
+ "learning_rate": 4.9000000000000005e-06,
1966
+ "loss": 0.0023,
1967
+ "step": 7900
1968
+ },
1969
+ {
1970
+ "epoch": 8.79,
1971
+ "learning_rate": 4.882758620689655e-06,
1972
+ "loss": 0.0009,
1973
+ "step": 7925
1974
+ },
1975
+ {
1976
+ "epoch": 8.81,
1977
+ "learning_rate": 4.86551724137931e-06,
1978
+ "loss": 0.0029,
1979
+ "step": 7950
1980
+ },
1981
+ {
1982
+ "epoch": 8.84,
1983
+ "learning_rate": 4.848275862068966e-06,
1984
+ "loss": 0.0014,
1985
+ "step": 7975
1986
+ },
1987
+ {
1988
+ "epoch": 8.87,
1989
+ "learning_rate": 4.831034482758621e-06,
1990
+ "loss": 0.0045,
1991
+ "step": 8000
1992
+ },
1993
+ {
1994
+ "epoch": 8.87,
1995
+ "eval_loss": 0.43131473660469055,
1996
+ "eval_runtime": 2955.3959,
1997
+ "eval_samples_per_second": 0.577,
1998
+ "eval_steps_per_second": 0.072,
1999
+ "eval_wer": 96.49990666417771,
2000
+ "step": 8000
2001
+ },
2002
+ {
2003
+ "epoch": 8.9,
2004
+ "learning_rate": 4.813793103448276e-06,
2005
+ "loss": 0.0098,
2006
+ "step": 8025
2007
+ },
2008
+ {
2009
+ "epoch": 8.92,
2010
+ "learning_rate": 4.7965517241379314e-06,
2011
+ "loss": 0.001,
2012
+ "step": 8050
2013
+ },
2014
+ {
2015
+ "epoch": 8.95,
2016
+ "learning_rate": 4.779310344827587e-06,
2017
+ "loss": 0.0032,
2018
+ "step": 8075
2019
+ },
2020
+ {
2021
+ "epoch": 8.98,
2022
+ "learning_rate": 4.762068965517242e-06,
2023
+ "loss": 0.0011,
2024
+ "step": 8100
2025
+ },
2026
+ {
2027
+ "epoch": 9.01,
2028
+ "learning_rate": 4.744827586206897e-06,
2029
+ "loss": 0.0028,
2030
+ "step": 8125
2031
+ },
2032
+ {
2033
+ "epoch": 9.04,
2034
+ "learning_rate": 4.727586206896553e-06,
2035
+ "loss": 0.0006,
2036
+ "step": 8150
2037
+ },
2038
+ {
2039
+ "epoch": 9.06,
2040
+ "learning_rate": 4.7103448275862075e-06,
2041
+ "loss": 0.0003,
2042
+ "step": 8175
2043
+ },
2044
+ {
2045
+ "epoch": 9.09,
2046
+ "learning_rate": 4.693103448275862e-06,
2047
+ "loss": 0.0006,
2048
+ "step": 8200
2049
+ },
2050
+ {
2051
+ "epoch": 9.12,
2052
+ "learning_rate": 4.675862068965517e-06,
2053
+ "loss": 0.0019,
2054
+ "step": 8225
2055
+ },
2056
+ {
2057
+ "epoch": 9.15,
2058
+ "learning_rate": 4.658620689655173e-06,
2059
+ "loss": 0.0011,
2060
+ "step": 8250
2061
+ },
2062
+ {
2063
+ "epoch": 9.17,
2064
+ "learning_rate": 4.641379310344828e-06,
2065
+ "loss": 0.0012,
2066
+ "step": 8275
2067
+ },
2068
+ {
2069
+ "epoch": 9.2,
2070
+ "learning_rate": 4.624137931034483e-06,
2071
+ "loss": 0.0018,
2072
+ "step": 8300
2073
+ },
2074
+ {
2075
+ "epoch": 9.23,
2076
+ "learning_rate": 4.606896551724138e-06,
2077
+ "loss": 0.0009,
2078
+ "step": 8325
2079
+ },
2080
+ {
2081
+ "epoch": 9.26,
2082
+ "learning_rate": 4.589655172413793e-06,
2083
+ "loss": 0.001,
2084
+ "step": 8350
2085
+ },
2086
+ {
2087
+ "epoch": 9.28,
2088
+ "learning_rate": 4.572413793103448e-06,
2089
+ "loss": 0.0022,
2090
+ "step": 8375
2091
+ },
2092
+ {
2093
+ "epoch": 9.31,
2094
+ "learning_rate": 4.555172413793104e-06,
2095
+ "loss": 0.0004,
2096
+ "step": 8400
2097
+ },
2098
+ {
2099
+ "epoch": 9.34,
2100
+ "learning_rate": 4.537931034482759e-06,
2101
+ "loss": 0.0006,
2102
+ "step": 8425
2103
+ },
2104
+ {
2105
+ "epoch": 9.37,
2106
+ "learning_rate": 4.5206896551724144e-06,
2107
+ "loss": 0.0028,
2108
+ "step": 8450
2109
+ },
2110
+ {
2111
+ "epoch": 9.4,
2112
+ "learning_rate": 4.503448275862069e-06,
2113
+ "loss": 0.0007,
2114
+ "step": 8475
2115
+ },
2116
+ {
2117
+ "epoch": 9.42,
2118
+ "learning_rate": 4.486896551724138e-06,
2119
+ "loss": 0.0066,
2120
+ "step": 8500
2121
+ },
2122
+ {
2123
+ "epoch": 9.45,
2124
+ "learning_rate": 4.4696551724137936e-06,
2125
+ "loss": 0.0026,
2126
+ "step": 8525
2127
+ },
2128
+ {
2129
+ "epoch": 9.48,
2130
+ "learning_rate": 4.452413793103449e-06,
2131
+ "loss": 0.0012,
2132
+ "step": 8550
2133
+ },
2134
+ {
2135
+ "epoch": 9.51,
2136
+ "learning_rate": 4.435172413793104e-06,
2137
+ "loss": 0.0007,
2138
+ "step": 8575
2139
+ },
2140
+ {
2141
+ "epoch": 9.53,
2142
+ "learning_rate": 4.417931034482759e-06,
2143
+ "loss": 0.0012,
2144
+ "step": 8600
2145
+ },
2146
+ {
2147
+ "epoch": 9.56,
2148
+ "learning_rate": 4.400689655172414e-06,
2149
+ "loss": 0.0004,
2150
+ "step": 8625
2151
+ },
2152
+ {
2153
+ "epoch": 9.59,
2154
+ "learning_rate": 4.38344827586207e-06,
2155
+ "loss": 0.0009,
2156
+ "step": 8650
2157
+ },
2158
+ {
2159
+ "epoch": 9.62,
2160
+ "learning_rate": 4.3662068965517245e-06,
2161
+ "loss": 0.0009,
2162
+ "step": 8675
2163
+ },
2164
+ {
2165
+ "epoch": 9.65,
2166
+ "learning_rate": 4.348965517241379e-06,
2167
+ "loss": 0.0015,
2168
+ "step": 8700
2169
+ },
2170
+ {
2171
+ "epoch": 9.67,
2172
+ "learning_rate": 4.331724137931035e-06,
2173
+ "loss": 0.003,
2174
+ "step": 8725
2175
+ },
2176
+ {
2177
+ "epoch": 9.7,
2178
+ "learning_rate": 4.31448275862069e-06,
2179
+ "loss": 0.0004,
2180
+ "step": 8750
2181
+ },
2182
+ {
2183
+ "epoch": 9.73,
2184
+ "learning_rate": 4.297241379310345e-06,
2185
+ "loss": 0.001,
2186
+ "step": 8775
2187
+ },
2188
+ {
2189
+ "epoch": 9.76,
2190
+ "learning_rate": 4.2800000000000005e-06,
2191
+ "loss": 0.0016,
2192
+ "step": 8800
2193
+ },
2194
+ {
2195
+ "epoch": 9.78,
2196
+ "learning_rate": 4.262758620689655e-06,
2197
+ "loss": 0.0005,
2198
+ "step": 8825
2199
+ },
2200
+ {
2201
+ "epoch": 9.81,
2202
+ "learning_rate": 4.24551724137931e-06,
2203
+ "loss": 0.0006,
2204
+ "step": 8850
2205
+ },
2206
+ {
2207
+ "epoch": 9.84,
2208
+ "learning_rate": 4.228275862068966e-06,
2209
+ "loss": 0.0031,
2210
+ "step": 8875
2211
+ },
2212
+ {
2213
+ "epoch": 9.87,
2214
+ "learning_rate": 4.211034482758621e-06,
2215
+ "loss": 0.0005,
2216
+ "step": 8900
2217
+ },
2218
+ {
2219
+ "epoch": 9.89,
2220
+ "learning_rate": 4.193793103448276e-06,
2221
+ "loss": 0.0003,
2222
+ "step": 8925
2223
+ },
2224
+ {
2225
+ "epoch": 9.92,
2226
+ "learning_rate": 4.1765517241379314e-06,
2227
+ "loss": 0.0034,
2228
+ "step": 8950
2229
+ },
2230
+ {
2231
+ "epoch": 9.95,
2232
+ "learning_rate": 4.159310344827587e-06,
2233
+ "loss": 0.0011,
2234
+ "step": 8975
2235
+ },
2236
+ {
2237
+ "epoch": 9.98,
2238
+ "learning_rate": 4.142068965517242e-06,
2239
+ "loss": 0.0004,
2240
+ "step": 9000
2241
+ },
2242
+ {
2243
+ "epoch": 9.98,
2244
+ "eval_loss": 0.44217491149902344,
2245
+ "eval_runtime": 3046.7205,
2246
+ "eval_samples_per_second": 0.559,
2247
+ "eval_steps_per_second": 0.07,
2248
+ "eval_wer": 83.05021467239126,
2249
+ "step": 9000
2250
+ },
2251
+ {
2252
+ "epoch": 10.01,
2253
+ "learning_rate": 4.124827586206897e-06,
2254
+ "loss": 0.0005,
2255
+ "step": 9025
2256
+ },
2257
+ {
2258
+ "epoch": 10.03,
2259
+ "learning_rate": 4.107586206896552e-06,
2260
+ "loss": 0.0003,
2261
+ "step": 9050
2262
+ },
2263
+ {
2264
+ "epoch": 10.06,
2265
+ "learning_rate": 4.0903448275862075e-06,
2266
+ "loss": 0.0023,
2267
+ "step": 9075
2268
+ },
2269
+ {
2270
+ "epoch": 10.09,
2271
+ "learning_rate": 4.073103448275862e-06,
2272
+ "loss": 0.0002,
2273
+ "step": 9100
2274
+ },
2275
+ {
2276
+ "epoch": 10.12,
2277
+ "learning_rate": 4.055862068965517e-06,
2278
+ "loss": 0.0002,
2279
+ "step": 9125
2280
+ },
2281
+ {
2282
+ "epoch": 10.14,
2283
+ "learning_rate": 4.038620689655173e-06,
2284
+ "loss": 0.0002,
2285
+ "step": 9150
2286
+ },
2287
+ {
2288
+ "epoch": 10.17,
2289
+ "learning_rate": 4.021379310344828e-06,
2290
+ "loss": 0.0003,
2291
+ "step": 9175
2292
+ },
2293
+ {
2294
+ "epoch": 10.2,
2295
+ "learning_rate": 4.004137931034483e-06,
2296
+ "loss": 0.0002,
2297
+ "step": 9200
2298
+ },
2299
+ {
2300
+ "epoch": 10.23,
2301
+ "learning_rate": 3.986896551724138e-06,
2302
+ "loss": 0.0046,
2303
+ "step": 9225
2304
+ },
2305
+ {
2306
+ "epoch": 10.25,
2307
+ "learning_rate": 3.969655172413793e-06,
2308
+ "loss": 0.0003,
2309
+ "step": 9250
2310
+ },
2311
+ {
2312
+ "epoch": 10.28,
2313
+ "learning_rate": 3.952413793103448e-06,
2314
+ "loss": 0.0002,
2315
+ "step": 9275
2316
+ },
2317
+ {
2318
+ "epoch": 10.31,
2319
+ "learning_rate": 3.935172413793104e-06,
2320
+ "loss": 0.0001,
2321
+ "step": 9300
2322
+ },
2323
+ {
2324
+ "epoch": 10.34,
2325
+ "learning_rate": 3.917931034482759e-06,
2326
+ "loss": 0.0002,
2327
+ "step": 9325
2328
+ },
2329
+ {
2330
+ "epoch": 10.37,
2331
+ "learning_rate": 3.9006896551724144e-06,
2332
+ "loss": 0.0002,
2333
+ "step": 9350
2334
+ },
2335
+ {
2336
+ "epoch": 10.39,
2337
+ "learning_rate": 3.883448275862069e-06,
2338
+ "loss": 0.0049,
2339
+ "step": 9375
2340
+ },
2341
+ {
2342
+ "epoch": 10.42,
2343
+ "learning_rate": 3.866206896551725e-06,
2344
+ "loss": 0.0005,
2345
+ "step": 9400
2346
+ },
2347
+ {
2348
+ "epoch": 10.45,
2349
+ "learning_rate": 3.84896551724138e-06,
2350
+ "loss": 0.0011,
2351
+ "step": 9425
2352
+ },
2353
+ {
2354
+ "epoch": 10.48,
2355
+ "learning_rate": 3.831724137931035e-06,
2356
+ "loss": 0.0007,
2357
+ "step": 9450
2358
+ },
2359
+ {
2360
+ "epoch": 10.5,
2361
+ "learning_rate": 3.81448275862069e-06,
2362
+ "loss": 0.0002,
2363
+ "step": 9475
2364
+ },
2365
+ {
2366
+ "epoch": 10.53,
2367
+ "learning_rate": 3.7972413793103454e-06,
2368
+ "loss": 0.0002,
2369
+ "step": 9500
2370
+ },
2371
+ {
2372
+ "epoch": 10.56,
2373
+ "learning_rate": 3.7800000000000002e-06,
2374
+ "loss": 0.0016,
2375
+ "step": 9525
2376
+ },
2377
+ {
2378
+ "epoch": 10.59,
2379
+ "learning_rate": 3.7627586206896555e-06,
2380
+ "loss": 0.0012,
2381
+ "step": 9550
2382
+ },
2383
+ {
2384
+ "epoch": 10.62,
2385
+ "learning_rate": 3.745517241379311e-06,
2386
+ "loss": 0.0028,
2387
+ "step": 9575
2388
+ },
2389
+ {
2390
+ "epoch": 10.64,
2391
+ "learning_rate": 3.7282758620689657e-06,
2392
+ "loss": 0.0002,
2393
+ "step": 9600
2394
+ },
2395
+ {
2396
+ "epoch": 10.67,
2397
+ "learning_rate": 3.711034482758621e-06,
2398
+ "loss": 0.0004,
2399
+ "step": 9625
2400
+ },
2401
+ {
2402
+ "epoch": 10.7,
2403
+ "learning_rate": 3.693793103448276e-06,
2404
+ "loss": 0.0002,
2405
+ "step": 9650
2406
+ },
2407
+ {
2408
+ "epoch": 10.73,
2409
+ "learning_rate": 3.676551724137931e-06,
2410
+ "loss": 0.0002,
2411
+ "step": 9675
2412
+ },
2413
+ {
2414
+ "epoch": 10.75,
2415
+ "learning_rate": 3.6593103448275864e-06,
2416
+ "loss": 0.0002,
2417
+ "step": 9700
2418
+ },
2419
+ {
2420
+ "epoch": 10.78,
2421
+ "learning_rate": 3.6420689655172413e-06,
2422
+ "loss": 0.0015,
2423
+ "step": 9725
2424
+ },
2425
+ {
2426
+ "epoch": 10.81,
2427
+ "learning_rate": 3.624827586206897e-06,
2428
+ "loss": 0.0002,
2429
+ "step": 9750
2430
+ },
2431
+ {
2432
+ "epoch": 10.84,
2433
+ "learning_rate": 3.6075862068965523e-06,
2434
+ "loss": 0.0005,
2435
+ "step": 9775
2436
+ },
2437
+ {
2438
+ "epoch": 10.86,
2439
+ "learning_rate": 3.590344827586207e-06,
2440
+ "loss": 0.0003,
2441
+ "step": 9800
2442
+ },
2443
+ {
2444
+ "epoch": 10.89,
2445
+ "learning_rate": 3.5731034482758625e-06,
2446
+ "loss": 0.0004,
2447
+ "step": 9825
2448
+ },
2449
+ {
2450
+ "epoch": 10.92,
2451
+ "learning_rate": 3.5558620689655178e-06,
2452
+ "loss": 0.002,
2453
+ "step": 9850
2454
+ },
2455
+ {
2456
+ "epoch": 10.95,
2457
+ "learning_rate": 3.5386206896551726e-06,
2458
+ "loss": 0.0002,
2459
+ "step": 9875
2460
+ },
2461
+ {
2462
+ "epoch": 10.98,
2463
+ "learning_rate": 3.521379310344828e-06,
2464
+ "loss": 0.0002,
2465
+ "step": 9900
2466
+ },
2467
+ {
2468
+ "epoch": 11.0,
2469
+ "learning_rate": 3.5041379310344832e-06,
2470
+ "loss": 0.0004,
2471
+ "step": 9925
2472
+ },
2473
+ {
2474
+ "epoch": 11.03,
2475
+ "learning_rate": 3.486896551724138e-06,
2476
+ "loss": 0.0013,
2477
+ "step": 9950
2478
+ },
2479
+ {
2480
+ "epoch": 11.06,
2481
+ "learning_rate": 3.4696551724137934e-06,
2482
+ "loss": 0.0008,
2483
+ "step": 9975
2484
+ },
2485
+ {
2486
+ "epoch": 11.09,
2487
+ "learning_rate": 3.4524137931034487e-06,
2488
+ "loss": 0.0002,
2489
+ "step": 10000
2490
+ },
2491
+ {
2492
+ "epoch": 11.09,
2493
+ "eval_loss": 0.45697087049484253,
2494
+ "eval_runtime": 3012.0295,
2495
+ "eval_samples_per_second": 0.566,
2496
+ "eval_steps_per_second": 0.071,
2497
+ "eval_wer": 78.59809594922531,
2498
+ "step": 10000
2499
+ },
2500
+ {
2501
+ "epoch": 11.11,
2502
+ "learning_rate": 3.4351724137931036e-06,
2503
+ "loss": 0.0006,
2504
+ "step": 10025
2505
+ },
2506
+ {
2507
+ "epoch": 11.14,
2508
+ "learning_rate": 3.417931034482759e-06,
2509
+ "loss": 0.0002,
2510
+ "step": 10050
2511
+ },
2512
+ {
2513
+ "epoch": 11.17,
2514
+ "learning_rate": 3.4006896551724137e-06,
2515
+ "loss": 0.0003,
2516
+ "step": 10075
2517
+ },
2518
+ {
2519
+ "epoch": 11.2,
2520
+ "learning_rate": 3.383448275862069e-06,
2521
+ "loss": 0.0001,
2522
+ "step": 10100
2523
+ },
2524
+ {
2525
+ "epoch": 11.23,
2526
+ "learning_rate": 3.3662068965517243e-06,
2527
+ "loss": 0.0001,
2528
+ "step": 10125
2529
+ },
2530
+ {
2531
+ "epoch": 11.25,
2532
+ "learning_rate": 3.34896551724138e-06,
2533
+ "loss": 0.0001,
2534
+ "step": 10150
2535
+ },
2536
+ {
2537
+ "epoch": 11.28,
2538
+ "learning_rate": 3.331724137931035e-06,
2539
+ "loss": 0.0007,
2540
+ "step": 10175
2541
+ },
2542
+ {
2543
+ "epoch": 11.31,
2544
+ "learning_rate": 3.31448275862069e-06,
2545
+ "loss": 0.0001,
2546
+ "step": 10200
2547
+ },
2548
+ {
2549
+ "epoch": 11.34,
2550
+ "learning_rate": 3.297241379310345e-06,
2551
+ "loss": 0.0015,
2552
+ "step": 10225
2553
+ },
2554
+ {
2555
+ "epoch": 11.36,
2556
+ "learning_rate": 3.2800000000000004e-06,
2557
+ "loss": 0.0001,
2558
+ "step": 10250
2559
+ },
2560
+ {
2561
+ "epoch": 11.39,
2562
+ "learning_rate": 3.2627586206896557e-06,
2563
+ "loss": 0.0003,
2564
+ "step": 10275
2565
+ },
2566
+ {
2567
+ "epoch": 11.42,
2568
+ "learning_rate": 3.2455172413793105e-06,
2569
+ "loss": 0.0004,
2570
+ "step": 10300
2571
+ },
2572
+ {
2573
+ "epoch": 11.45,
2574
+ "learning_rate": 3.228275862068966e-06,
2575
+ "loss": 0.0024,
2576
+ "step": 10325
2577
+ },
2578
+ {
2579
+ "epoch": 11.47,
2580
+ "learning_rate": 3.211034482758621e-06,
2581
+ "loss": 0.0018,
2582
+ "step": 10350
2583
+ },
2584
+ {
2585
+ "epoch": 11.5,
2586
+ "learning_rate": 3.193793103448276e-06,
2587
+ "loss": 0.0001,
2588
+ "step": 10375
2589
+ },
2590
+ {
2591
+ "epoch": 11.53,
2592
+ "learning_rate": 3.1765517241379313e-06,
2593
+ "loss": 0.0002,
2594
+ "step": 10400
2595
+ },
2596
+ {
2597
+ "epoch": 11.56,
2598
+ "learning_rate": 3.1593103448275866e-06,
2599
+ "loss": 0.0004,
2600
+ "step": 10425
2601
+ },
2602
+ {
2603
+ "epoch": 11.59,
2604
+ "learning_rate": 3.1420689655172414e-06,
2605
+ "loss": 0.0004,
2606
+ "step": 10450
2607
+ },
2608
+ {
2609
+ "epoch": 11.61,
2610
+ "learning_rate": 3.1248275862068967e-06,
2611
+ "loss": 0.0019,
2612
+ "step": 10475
2613
+ },
2614
+ {
2615
+ "epoch": 11.64,
2616
+ "learning_rate": 3.1075862068965516e-06,
2617
+ "loss": 0.0002,
2618
+ "step": 10500
2619
+ },
2620
+ {
2621
+ "epoch": 11.67,
2622
+ "learning_rate": 3.091034482758621e-06,
2623
+ "loss": 0.0012,
2624
+ "step": 10525
2625
+ },
2626
+ {
2627
+ "epoch": 11.7,
2628
+ "learning_rate": 3.073793103448276e-06,
2629
+ "loss": 0.0001,
2630
+ "step": 10550
2631
+ },
2632
+ {
2633
+ "epoch": 11.72,
2634
+ "learning_rate": 3.056551724137931e-06,
2635
+ "loss": 0.0002,
2636
+ "step": 10575
2637
+ },
2638
+ {
2639
+ "epoch": 11.75,
2640
+ "learning_rate": 3.0393103448275864e-06,
2641
+ "loss": 0.0002,
2642
+ "step": 10600
2643
+ },
2644
+ {
2645
+ "epoch": 11.78,
2646
+ "learning_rate": 3.0220689655172413e-06,
2647
+ "loss": 0.0034,
2648
+ "step": 10625
2649
+ },
2650
+ {
2651
+ "epoch": 11.81,
2652
+ "learning_rate": 3.004827586206897e-06,
2653
+ "loss": 0.0002,
2654
+ "step": 10650
2655
+ },
2656
+ {
2657
+ "epoch": 11.83,
2658
+ "learning_rate": 2.9875862068965523e-06,
2659
+ "loss": 0.0001,
2660
+ "step": 10675
2661
+ },
2662
+ {
2663
+ "epoch": 11.86,
2664
+ "learning_rate": 2.970344827586207e-06,
2665
+ "loss": 0.0032,
2666
+ "step": 10700
2667
+ },
2668
+ {
2669
+ "epoch": 11.89,
2670
+ "learning_rate": 2.9531034482758625e-06,
2671
+ "loss": 0.0011,
2672
+ "step": 10725
2673
+ },
2674
+ {
2675
+ "epoch": 11.92,
2676
+ "learning_rate": 2.9358620689655178e-06,
2677
+ "loss": 0.0002,
2678
+ "step": 10750
2679
+ },
2680
+ {
2681
+ "epoch": 11.95,
2682
+ "learning_rate": 2.9186206896551727e-06,
2683
+ "loss": 0.0014,
2684
+ "step": 10775
2685
+ },
2686
+ {
2687
+ "epoch": 11.97,
2688
+ "learning_rate": 2.901379310344828e-06,
2689
+ "loss": 0.0024,
2690
+ "step": 10800
2691
+ },
2692
+ {
2693
+ "epoch": 12.0,
2694
+ "learning_rate": 2.884137931034483e-06,
2695
+ "loss": 0.0002,
2696
+ "step": 10825
2697
+ },
2698
+ {
2699
+ "epoch": 12.03,
2700
+ "learning_rate": 2.866896551724138e-06,
2701
+ "loss": 0.0008,
2702
+ "step": 10850
2703
+ },
2704
+ {
2705
+ "epoch": 12.06,
2706
+ "learning_rate": 2.8496551724137934e-06,
2707
+ "loss": 0.0002,
2708
+ "step": 10875
2709
+ },
2710
+ {
2711
+ "epoch": 12.08,
2712
+ "learning_rate": 2.8324137931034483e-06,
2713
+ "loss": 0.0001,
2714
+ "step": 10900
2715
+ },
2716
+ {
2717
+ "epoch": 12.11,
2718
+ "learning_rate": 2.8151724137931036e-06,
2719
+ "loss": 0.0003,
2720
+ "step": 10925
2721
+ },
2722
+ {
2723
+ "epoch": 12.14,
2724
+ "learning_rate": 2.797931034482759e-06,
2725
+ "loss": 0.0043,
2726
+ "step": 10950
2727
+ },
2728
+ {
2729
+ "epoch": 12.17,
2730
+ "learning_rate": 2.7806896551724137e-06,
2731
+ "loss": 0.0011,
2732
+ "step": 10975
2733
+ },
2734
+ {
2735
+ "epoch": 12.2,
2736
+ "learning_rate": 2.763448275862069e-06,
2737
+ "loss": 0.0001,
2738
+ "step": 11000
2739
+ },
2740
+ {
2741
+ "epoch": 12.2,
2742
+ "eval_loss": 0.47608184814453125,
2743
+ "eval_runtime": 3039.0267,
2744
+ "eval_samples_per_second": 0.561,
2745
+ "eval_steps_per_second": 0.07,
2746
+ "eval_wer": 91.78644763860369,
2747
+ "step": 11000
2748
+ },
2749
+ {
2750
+ "epoch": 12.22,
2751
+ "learning_rate": 2.7462068965517243e-06,
2752
+ "loss": 0.0019,
2753
+ "step": 11025
2754
+ },
2755
+ {
2756
+ "epoch": 12.25,
2757
+ "learning_rate": 2.728965517241379e-06,
2758
+ "loss": 0.0002,
2759
+ "step": 11050
2760
+ },
2761
+ {
2762
+ "epoch": 12.28,
2763
+ "learning_rate": 2.711724137931035e-06,
2764
+ "loss": 0.003,
2765
+ "step": 11075
2766
+ },
2767
+ {
2768
+ "epoch": 12.31,
2769
+ "learning_rate": 2.69448275862069e-06,
2770
+ "loss": 0.0002,
2771
+ "step": 11100
2772
+ },
2773
+ {
2774
+ "epoch": 12.33,
2775
+ "learning_rate": 2.677241379310345e-06,
2776
+ "loss": 0.0018,
2777
+ "step": 11125
2778
+ },
2779
+ {
2780
+ "epoch": 12.36,
2781
+ "learning_rate": 2.6600000000000004e-06,
2782
+ "loss": 0.0079,
2783
+ "step": 11150
2784
+ },
2785
+ {
2786
+ "epoch": 12.39,
2787
+ "learning_rate": 2.6427586206896557e-06,
2788
+ "loss": 0.0001,
2789
+ "step": 11175
2790
+ },
2791
+ {
2792
+ "epoch": 12.42,
2793
+ "learning_rate": 2.6255172413793105e-06,
2794
+ "loss": 0.0046,
2795
+ "step": 11200
2796
+ },
2797
+ {
2798
+ "epoch": 12.44,
2799
+ "learning_rate": 2.608275862068966e-06,
2800
+ "loss": 0.0008,
2801
+ "step": 11225
2802
+ },
2803
+ {
2804
+ "epoch": 12.47,
2805
+ "learning_rate": 2.5910344827586207e-06,
2806
+ "loss": 0.001,
2807
+ "step": 11250
2808
+ },
2809
+ {
2810
+ "epoch": 12.5,
2811
+ "learning_rate": 2.573793103448276e-06,
2812
+ "loss": 0.001,
2813
+ "step": 11275
2814
+ },
2815
+ {
2816
+ "epoch": 12.53,
2817
+ "learning_rate": 2.5565517241379313e-06,
2818
+ "loss": 0.0005,
2819
+ "step": 11300
2820
+ },
2821
+ {
2822
+ "epoch": 12.56,
2823
+ "learning_rate": 2.539310344827586e-06,
2824
+ "loss": 0.0008,
2825
+ "step": 11325
2826
+ },
2827
+ {
2828
+ "epoch": 12.58,
2829
+ "learning_rate": 2.5220689655172414e-06,
2830
+ "loss": 0.0002,
2831
+ "step": 11350
2832
+ },
2833
+ {
2834
+ "epoch": 12.61,
2835
+ "learning_rate": 2.5048275862068967e-06,
2836
+ "loss": 0.0054,
2837
+ "step": 11375
2838
+ },
2839
+ {
2840
+ "epoch": 12.64,
2841
+ "learning_rate": 2.487586206896552e-06,
2842
+ "loss": 0.0004,
2843
+ "step": 11400
2844
+ },
2845
+ {
2846
+ "epoch": 12.67,
2847
+ "learning_rate": 2.4703448275862073e-06,
2848
+ "loss": 0.0001,
2849
+ "step": 11425
2850
+ },
2851
+ {
2852
+ "epoch": 12.69,
2853
+ "learning_rate": 2.453103448275862e-06,
2854
+ "loss": 0.0001,
2855
+ "step": 11450
2856
+ },
2857
+ {
2858
+ "epoch": 12.72,
2859
+ "learning_rate": 2.4358620689655175e-06,
2860
+ "loss": 0.0001,
2861
+ "step": 11475
2862
+ },
2863
+ {
2864
+ "epoch": 12.75,
2865
+ "learning_rate": 2.4186206896551724e-06,
2866
+ "loss": 0.0001,
2867
+ "step": 11500
2868
+ },
2869
+ {
2870
+ "epoch": 12.78,
2871
+ "learning_rate": 2.4013793103448277e-06,
2872
+ "loss": 0.0001,
2873
+ "step": 11525
2874
+ },
2875
+ {
2876
+ "epoch": 12.8,
2877
+ "learning_rate": 2.384137931034483e-06,
2878
+ "loss": 0.0001,
2879
+ "step": 11550
2880
+ },
2881
+ {
2882
+ "epoch": 12.83,
2883
+ "learning_rate": 2.3668965517241382e-06,
2884
+ "loss": 0.0014,
2885
+ "step": 11575
2886
+ },
2887
+ {
2888
+ "epoch": 12.86,
2889
+ "learning_rate": 2.3496551724137935e-06,
2890
+ "loss": 0.0012,
2891
+ "step": 11600
2892
+ },
2893
+ {
2894
+ "epoch": 12.89,
2895
+ "learning_rate": 2.3324137931034484e-06,
2896
+ "loss": 0.0001,
2897
+ "step": 11625
2898
+ },
2899
+ {
2900
+ "epoch": 12.92,
2901
+ "learning_rate": 2.3151724137931037e-06,
2902
+ "loss": 0.0001,
2903
+ "step": 11650
2904
+ },
2905
+ {
2906
+ "epoch": 12.94,
2907
+ "learning_rate": 2.297931034482759e-06,
2908
+ "loss": 0.0004,
2909
+ "step": 11675
2910
+ },
2911
+ {
2912
+ "epoch": 12.97,
2913
+ "learning_rate": 2.280689655172414e-06,
2914
+ "loss": 0.0001,
2915
+ "step": 11700
2916
+ },
2917
+ {
2918
+ "epoch": 13.0,
2919
+ "learning_rate": 2.263448275862069e-06,
2920
+ "loss": 0.0001,
2921
+ "step": 11725
2922
+ },
2923
+ {
2924
+ "epoch": 13.03,
2925
+ "learning_rate": 2.246206896551724e-06,
2926
+ "loss": 0.0001,
2927
+ "step": 11750
2928
+ },
2929
+ {
2930
+ "epoch": 13.05,
2931
+ "learning_rate": 2.2289655172413797e-06,
2932
+ "loss": 0.0001,
2933
+ "step": 11775
2934
+ },
2935
+ {
2936
+ "epoch": 13.08,
2937
+ "learning_rate": 2.2117241379310346e-06,
2938
+ "loss": 0.0001,
2939
+ "step": 11800
2940
+ },
2941
+ {
2942
+ "epoch": 13.11,
2943
+ "learning_rate": 2.19448275862069e-06,
2944
+ "loss": 0.0001,
2945
+ "step": 11825
2946
+ },
2947
+ {
2948
+ "epoch": 13.14,
2949
+ "learning_rate": 2.177241379310345e-06,
2950
+ "loss": 0.0001,
2951
+ "step": 11850
2952
+ },
2953
+ {
2954
+ "epoch": 13.17,
2955
+ "learning_rate": 2.16e-06,
2956
+ "loss": 0.0001,
2957
+ "step": 11875
2958
+ },
2959
+ {
2960
+ "epoch": 13.19,
2961
+ "learning_rate": 2.1427586206896554e-06,
2962
+ "loss": 0.0001,
2963
+ "step": 11900
2964
+ },
2965
+ {
2966
+ "epoch": 13.22,
2967
+ "learning_rate": 2.1255172413793102e-06,
2968
+ "loss": 0.0001,
2969
+ "step": 11925
2970
+ },
2971
+ {
2972
+ "epoch": 13.25,
2973
+ "learning_rate": 2.1082758620689655e-06,
2974
+ "loss": 0.0001,
2975
+ "step": 11950
2976
+ },
2977
+ {
2978
+ "epoch": 13.28,
2979
+ "learning_rate": 2.091034482758621e-06,
2980
+ "loss": 0.0001,
2981
+ "step": 11975
2982
+ },
2983
+ {
2984
+ "epoch": 13.3,
2985
+ "learning_rate": 2.073793103448276e-06,
2986
+ "loss": 0.0002,
2987
+ "step": 12000
2988
+ },
2989
+ {
2990
+ "epoch": 13.3,
2991
+ "eval_loss": 0.4738454222679138,
2992
+ "eval_runtime": 2933.9692,
2993
+ "eval_samples_per_second": 0.581,
2994
+ "eval_steps_per_second": 0.073,
2995
+ "eval_wer": 62.61900317341795,
2996
+ "step": 12000
2997
+ },
2998
+ {
2999
+ "epoch": 13.33,
3000
+ "learning_rate": 2.0565517241379314e-06,
3001
+ "loss": 0.0001,
3002
+ "step": 12025
3003
+ },
3004
+ {
3005
+ "epoch": 13.36,
3006
+ "learning_rate": 2.0393103448275863e-06,
3007
+ "loss": 0.0004,
3008
+ "step": 12050
3009
+ },
3010
+ {
3011
+ "epoch": 13.39,
3012
+ "learning_rate": 2.0220689655172416e-06,
3013
+ "loss": 0.0001,
3014
+ "step": 12075
3015
+ },
3016
+ {
3017
+ "epoch": 13.41,
3018
+ "learning_rate": 2.004827586206897e-06,
3019
+ "loss": 0.0018,
3020
+ "step": 12100
3021
+ },
3022
+ {
3023
+ "epoch": 13.44,
3024
+ "learning_rate": 1.9875862068965517e-06,
3025
+ "loss": 0.0001,
3026
+ "step": 12125
3027
+ },
3028
+ {
3029
+ "epoch": 13.47,
3030
+ "learning_rate": 1.970344827586207e-06,
3031
+ "loss": 0.0001,
3032
+ "step": 12150
3033
+ },
3034
+ {
3035
+ "epoch": 13.5,
3036
+ "learning_rate": 1.9531034482758623e-06,
3037
+ "loss": 0.0001,
3038
+ "step": 12175
3039
+ },
3040
+ {
3041
+ "epoch": 13.53,
3042
+ "learning_rate": 1.9358620689655176e-06,
3043
+ "loss": 0.002,
3044
+ "step": 12200
3045
+ },
3046
+ {
3047
+ "epoch": 13.55,
3048
+ "learning_rate": 1.9186206896551725e-06,
3049
+ "loss": 0.0002,
3050
+ "step": 12225
3051
+ },
3052
+ {
3053
+ "epoch": 13.58,
3054
+ "learning_rate": 1.9013793103448278e-06,
3055
+ "loss": 0.0047,
3056
+ "step": 12250
3057
+ },
3058
+ {
3059
+ "epoch": 13.61,
3060
+ "learning_rate": 1.8841379310344829e-06,
3061
+ "loss": 0.0001,
3062
+ "step": 12275
3063
+ },
3064
+ {
3065
+ "epoch": 13.64,
3066
+ "learning_rate": 1.866896551724138e-06,
3067
+ "loss": 0.0001,
3068
+ "step": 12300
3069
+ },
3070
+ {
3071
+ "epoch": 13.66,
3072
+ "learning_rate": 1.8496551724137932e-06,
3073
+ "loss": 0.0001,
3074
+ "step": 12325
3075
+ },
3076
+ {
3077
+ "epoch": 13.69,
3078
+ "learning_rate": 1.8324137931034483e-06,
3079
+ "loss": 0.0001,
3080
+ "step": 12350
3081
+ },
3082
+ {
3083
+ "epoch": 13.72,
3084
+ "learning_rate": 1.8151724137931036e-06,
3085
+ "loss": 0.0012,
3086
+ "step": 12375
3087
+ },
3088
+ {
3089
+ "epoch": 13.75,
3090
+ "learning_rate": 1.797931034482759e-06,
3091
+ "loss": 0.0006,
3092
+ "step": 12400
3093
+ },
3094
+ {
3095
+ "epoch": 13.77,
3096
+ "learning_rate": 1.780689655172414e-06,
3097
+ "loss": 0.0001,
3098
+ "step": 12425
3099
+ },
3100
+ {
3101
+ "epoch": 13.8,
3102
+ "learning_rate": 1.763448275862069e-06,
3103
+ "loss": 0.0004,
3104
+ "step": 12450
3105
+ },
3106
+ {
3107
+ "epoch": 13.83,
3108
+ "learning_rate": 1.7462068965517244e-06,
3109
+ "loss": 0.0005,
3110
+ "step": 12475
3111
+ },
3112
+ {
3113
+ "epoch": 13.86,
3114
+ "learning_rate": 1.7289655172413794e-06,
3115
+ "loss": 0.0002,
3116
+ "step": 12500
3117
+ },
3118
+ {
3119
+ "epoch": 13.89,
3120
+ "learning_rate": 1.7117241379310345e-06,
3121
+ "loss": 0.0001,
3122
+ "step": 12525
3123
+ },
3124
+ {
3125
+ "epoch": 13.91,
3126
+ "learning_rate": 1.6944827586206896e-06,
3127
+ "loss": 0.0001,
3128
+ "step": 12550
3129
+ },
3130
+ {
3131
+ "epoch": 13.94,
3132
+ "learning_rate": 1.6772413793103451e-06,
3133
+ "loss": 0.0001,
3134
+ "step": 12575
3135
+ },
3136
+ {
3137
+ "epoch": 13.97,
3138
+ "learning_rate": 1.6600000000000002e-06,
3139
+ "loss": 0.0001,
3140
+ "step": 12600
3141
+ },
3142
+ {
3143
+ "epoch": 14.0,
3144
+ "learning_rate": 1.6427586206896553e-06,
3145
+ "loss": 0.0001,
3146
+ "step": 12625
3147
+ },
3148
+ {
3149
+ "epoch": 14.02,
3150
+ "learning_rate": 1.6255172413793106e-06,
3151
+ "loss": 0.0001,
3152
+ "step": 12650
3153
+ },
3154
+ {
3155
+ "epoch": 14.05,
3156
+ "learning_rate": 1.6082758620689657e-06,
3157
+ "loss": 0.0001,
3158
+ "step": 12675
3159
+ },
3160
+ {
3161
+ "epoch": 14.08,
3162
+ "learning_rate": 1.5910344827586207e-06,
3163
+ "loss": 0.0001,
3164
+ "step": 12700
3165
+ },
3166
+ {
3167
+ "epoch": 14.11,
3168
+ "learning_rate": 1.5737931034482758e-06,
3169
+ "loss": 0.0001,
3170
+ "step": 12725
3171
+ },
3172
+ {
3173
+ "epoch": 14.14,
3174
+ "learning_rate": 1.5565517241379311e-06,
3175
+ "loss": 0.0001,
3176
+ "step": 12750
3177
+ },
3178
+ {
3179
+ "epoch": 14.16,
3180
+ "learning_rate": 1.5393103448275864e-06,
3181
+ "loss": 0.0001,
3182
+ "step": 12775
3183
+ },
3184
+ {
3185
+ "epoch": 14.19,
3186
+ "learning_rate": 1.5220689655172415e-06,
3187
+ "loss": 0.0001,
3188
+ "step": 12800
3189
+ },
3190
+ {
3191
+ "epoch": 14.22,
3192
+ "learning_rate": 1.5048275862068968e-06,
3193
+ "loss": 0.0001,
3194
+ "step": 12825
3195
+ },
3196
+ {
3197
+ "epoch": 14.25,
3198
+ "learning_rate": 1.4875862068965519e-06,
3199
+ "loss": 0.0001,
3200
+ "step": 12850
3201
+ },
3202
+ {
3203
+ "epoch": 14.27,
3204
+ "learning_rate": 1.470344827586207e-06,
3205
+ "loss": 0.0001,
3206
+ "step": 12875
3207
+ },
3208
+ {
3209
+ "epoch": 14.3,
3210
+ "learning_rate": 1.4531034482758622e-06,
3211
+ "loss": 0.0001,
3212
+ "step": 12900
3213
+ },
3214
+ {
3215
+ "epoch": 14.33,
3216
+ "learning_rate": 1.4358620689655173e-06,
3217
+ "loss": 0.0001,
3218
+ "step": 12925
3219
+ },
3220
+ {
3221
+ "epoch": 14.36,
3222
+ "learning_rate": 1.4186206896551724e-06,
3223
+ "loss": 0.0001,
3224
+ "step": 12950
3225
+ },
3226
+ {
3227
+ "epoch": 14.38,
3228
+ "learning_rate": 1.401379310344828e-06,
3229
+ "loss": 0.0001,
3230
+ "step": 12975
3231
+ },
3232
+ {
3233
+ "epoch": 14.41,
3234
+ "learning_rate": 1.384137931034483e-06,
3235
+ "loss": 0.0001,
3236
+ "step": 13000
3237
+ },
3238
+ {
3239
+ "epoch": 14.41,
3240
+ "eval_loss": 0.478272408246994,
3241
+ "eval_runtime": 2940.524,
3242
+ "eval_samples_per_second": 0.579,
3243
+ "eval_steps_per_second": 0.072,
3244
+ "eval_wer": 56.720179204778795,
3245
+ "step": 13000
3246
+ }
3247
+ ],
3248
+ "max_steps": 15000,
3249
+ "num_train_epochs": 17,
3250
+ "total_flos": 1.500240062103552e+19,
3251
+ "trial_name": null,
3252
+ "trial_params": null
3253
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c41907287a2decc2880f903f081ad68eed15c658e4c877576dc336e772444bb3
3
+ size 3771