Hugging Face's logo

AsemBadr
/
final-whisper-for-initial-publish-v2

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
whisper
Model card Files
xet
Metrics Community
AsemBadr commited on
Commit
a56ed34
·
verified ·
1 Parent(s): 0d28f50

Training in progress, step 11000, checkpoint

Browse files
Files changed (9) hide show
  1. checkpoint-11000/config.json +65 -0
  2. checkpoint-11000/generation_config.json +265 -0
  3. checkpoint-11000/model.safetensors +3 -0
  4. checkpoint-11000/optimizer.pt +3 -0
  5. checkpoint-11000/preprocessor_config.json +14 -0
  6. checkpoint-11000/rng_state.pth +3 -0
  7. checkpoint-11000/scheduler.pt +3 -0
  8. checkpoint-11000/trainer_state.json +3311 -0
  9. checkpoint-11000/training_args.bin +3 -0
checkpoint-11000/config.json ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "AsemBadr/final-whisper-for-initial-publish-v2",
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": [
29
+ [
30
+ 1,
31
+ 50272
32
+ ],
33
+ [
34
+ 2,
35
+ 50359
36
+ ],
37
+ [
38
+ 3,
39
+ 50363
40
+ ]
41
+ ],
42
+ "init_std": 0.02,
43
+ "is_encoder_decoder": true,
44
+ "mask_feature_length": 10,
45
+ "mask_feature_min_masks": 0,
46
+ "mask_feature_prob": 0.0,
47
+ "mask_time_length": 10,
48
+ "mask_time_min_masks": 2,
49
+ "mask_time_prob": 0.05,
50
+ "max_length": 448,
51
+ "max_source_positions": 1500,
52
+ "max_target_positions": 448,
53
+ "median_filter_width": 7,
54
+ "model_type": "whisper",
55
+ "num_hidden_layers": 12,
56
+ "num_mel_bins": 80,
57
+ "pad_token_id": 50257,
58
+ "scale_embedding": false,
59
+ "suppress_tokens": [],
60
+ "torch_dtype": "float32",
61
+ "transformers_version": "4.41.0.dev0",
62
+ "use_cache": true,
63
+ "use_weighted_layer_sum": false,
64
+ "vocab_size": 51865
65
+ }
checkpoint-11000/generation_config.json ADDED
@@ -0,0 +1,265 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alignment_heads": [
3
+ [
4
+ 5,
5
+ 3
6
+ ],
7
+ [
8
+ 5,
9
+ 9
10
+ ],
11
+ [
12
+ 8,
13
+ 0
14
+ ],
15
+ [
16
+ 8,
17
+ 4
18
+ ],
19
+ [
20
+ 8,
21
+ 7
22
+ ],
23
+ [
24
+ 8,
25
+ 8
26
+ ],
27
+ [
28
+ 9,
29
+ 0
30
+ ],
31
+ [
32
+ 9,
33
+ 7
34
+ ],
35
+ [
36
+ 9,
37
+ 9
38
+ ],
39
+ [
40
+ 10,
41
+ 5
42
+ ]
43
+ ],
44
+ "begin_suppress_tokens": [
45
+ 220,
46
+ 50257
47
+ ],
48
+ "bos_token_id": 50257,
49
+ "decoder_start_token_id": 50258,
50
+ "eos_token_id": 50257,
51
+ "forced_decoder_ids": [
52
+ [
53
+ 1,
54
+ null
55
+ ],
56
+ [
57
+ 2,
58
+ 50359
59
+ ]
60
+ ],
61
+ "is_multilingual": true,
62
+ "lang_to_id": {
63
+ "<|af|>": 50327,
64
+ "<|am|>": 50334,
65
+ "<|ar|>": 50272,
66
+ "<|as|>": 50350,
67
+ "<|az|>": 50304,
68
+ "<|ba|>": 50355,
69
+ "<|be|>": 50330,
70
+ "<|bg|>": 50292,
71
+ "<|bn|>": 50302,
72
+ "<|bo|>": 50347,
73
+ "<|br|>": 50309,
74
+ "<|bs|>": 50315,
75
+ "<|ca|>": 50270,
76
+ "<|cs|>": 50283,
77
+ "<|cy|>": 50297,
78
+ "<|da|>": 50285,
79
+ "<|de|>": 50261,
80
+ "<|el|>": 50281,
81
+ "<|en|>": 50259,
82
+ "<|es|>": 50262,
83
+ "<|et|>": 50307,
84
+ "<|eu|>": 50310,
85
+ "<|fa|>": 50300,
86
+ "<|fi|>": 50277,
87
+ "<|fo|>": 50338,
88
+ "<|fr|>": 50265,
89
+ "<|gl|>": 50319,
90
+ "<|gu|>": 50333,
91
+ "<|haw|>": 50352,
92
+ "<|ha|>": 50354,
93
+ "<|he|>": 50279,
94
+ "<|hi|>": 50276,
95
+ "<|hr|>": 50291,
96
+ "<|ht|>": 50339,
97
+ "<|hu|>": 50286,
98
+ "<|hy|>": 50312,
99
+ "<|id|>": 50275,
100
+ "<|is|>": 50311,
101
+ "<|it|>": 50274,
102
+ "<|ja|>": 50266,
103
+ "<|jw|>": 50356,
104
+ "<|ka|>": 50329,
105
+ "<|kk|>": 50316,
106
+ "<|km|>": 50323,
107
+ "<|kn|>": 50306,
108
+ "<|ko|>": 50264,
109
+ "<|la|>": 50294,
110
+ "<|lb|>": 50345,
111
+ "<|ln|>": 50353,
112
+ "<|lo|>": 50336,
113
+ "<|lt|>": 50293,
114
+ "<|lv|>": 50301,
115
+ "<|mg|>": 50349,
116
+ "<|mi|>": 50295,
117
+ "<|mk|>": 50308,
118
+ "<|ml|>": 50296,
119
+ "<|mn|>": 50314,
120
+ "<|mr|>": 50320,
121
+ "<|ms|>": 50282,
122
+ "<|mt|>": 50343,
123
+ "<|my|>": 50346,
124
+ "<|ne|>": 50313,
125
+ "<|nl|>": 50271,
126
+ "<|nn|>": 50342,
127
+ "<|no|>": 50288,
128
+ "<|oc|>": 50328,
129
+ "<|pa|>": 50321,
130
+ "<|pl|>": 50269,
131
+ "<|ps|>": 50340,
132
+ "<|pt|>": 50267,
133
+ "<|ro|>": 50284,
134
+ "<|ru|>": 50263,
135
+ "<|sa|>": 50344,
136
+ "<|sd|>": 50332,
137
+ "<|si|>": 50322,
138
+ "<|sk|>": 50298,
139
+ "<|sl|>": 50305,
140
+ "<|sn|>": 50324,
141
+ "<|so|>": 50326,
142
+ "<|sq|>": 50317,
143
+ "<|sr|>": 50303,
144
+ "<|su|>": 50357,
145
+ "<|sv|>": 50273,
146
+ "<|sw|>": 50318,
147
+ "<|ta|>": 50287,
148
+ "<|te|>": 50299,
149
+ "<|tg|>": 50331,
150
+ "<|th|>": 50289,
151
+ "<|tk|>": 50341,
152
+ "<|tl|>": 50348,
153
+ "<|tr|>": 50268,
154
+ "<|tt|>": 50351,
155
+ "<|uk|>": 50280,
156
+ "<|ur|>": 50290,
157
+ "<|uz|>": 50337,
158
+ "<|vi|>": 50278,
159
+ "<|yi|>": 50335,
160
+ "<|yo|>": 50325,
161
+ "<|zh|>": 50260
162
+ },
163
+ "language": "ar",
164
+ "max_initial_timestamp_index": 50,
165
+ "max_length": 448,
166
+ "no_timestamps_token_id": 50363,
167
+ "pad_token_id": 50257,
168
+ "prev_sot_token_id": 50361,
169
+ "return_timestamps": false,
170
+ "suppress_tokens": [
171
+ 1,
172
+ 2,
173
+ 7,
174
+ 8,
175
+ 9,
176
+ 10,
177
+ 14,
178
+ 25,
179
+ 26,
180
+ 27,
181
+ 28,
182
+ 29,
183
+ 31,
184
+ 58,
185
+ 59,
186
+ 60,
187
+ 61,
188
+ 62,
189
+ 63,
190
+ 90,
191
+ 91,
192
+ 92,
193
+ 93,
194
+ 359,
195
+ 503,
196
+ 522,
197
+ 542,
198
+ 873,
199
+ 893,
200
+ 902,
201
+ 918,
202
+ 922,
203
+ 931,
204
+ 1350,
205
+ 1853,
206
+ 1982,
207
+ 2460,
208
+ 2627,
209
+ 3246,
210
+ 3253,
211
+ 3268,
212
+ 3536,
213
+ 3846,
214
+ 3961,
215
+ 4183,
216
+ 4667,
217
+ 6585,
218
+ 6647,
219
+ 7273,
220
+ 9061,
221
+ 9383,
222
+ 10428,
223
+ 10929,
224
+ 11938,
225
+ 12033,
226
+ 12331,
227
+ 12562,
228
+ 13793,
229
+ 14157,
230
+ 14635,
231
+ 15265,
232
+ 15618,
233
+ 16553,
234
+ 16604,
235
+ 18362,
236
+ 18956,
237
+ 20075,
238
+ 21675,
239
+ 22520,
240
+ 26130,
241
+ 26161,
242
+ 26435,
243
+ 28279,
244
+ 29464,
245
+ 31650,
246
+ 32302,
247
+ 32470,
248
+ 36865,
249
+ 42863,
250
+ 47425,
251
+ 49870,
252
+ 50254,
253
+ 50258,
254
+ 50358,
255
+ 50359,
256
+ 50360,
257
+ 50361,
258
+ 50362
259
+ ],
260
+ "task_to_id": {
261
+ "transcribe": 50359,
262
+ "translate": 50358
263
+ },
264
+ "transformers_version": "4.41.0.dev0"
265
+ }
checkpoint-11000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8735e5e28bcc547ffae9fba439acc87fc78a1d9f826fe91e5c8fc540b645572a
3
+ size 966995080
checkpoint-11000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62cca8508338c4ad3783efcd07ca7bf77ef22de32fe1ca96645da5bff93003d4
3
+ size 1925064044
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/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e8ecb3947a79291540be57c28bdfd1c7c1e5f2b89ad206edfceb2d76a7df5e23
3
+ size 14244
checkpoint-11000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:783efff83665f8186a9d8244750eb012cfc2201637bd0ab8afc1d26752546185
3
+ size 1064
checkpoint-11000/trainer_state.json ADDED
@@ -0,0 +1,3311 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 2.4780548434708303,
3
+ "best_model_checkpoint": "./final-whisper-for-initial-publish-v2/checkpoint-8000",
4
+ "epoch": 18.12191103789127,
5
+ "eval_steps": 500,
6
+ "global_step": 11000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.04118616144975288,
13
+ "grad_norm": 67.08155822753906,
14
+ "learning_rate": 4.2000000000000006e-07,
15
+ "loss": 2.8368,
16
+ "step": 25
17
+ },
18
+ {
19
+ "epoch": 0.08237232289950576,
20
+ "grad_norm": 17.44784164428711,
21
+ "learning_rate": 9.200000000000001e-07,
22
+ "loss": 2.0282,
23
+ "step": 50
24
+ },
25
+ {
26
+ "epoch": 0.12355848434925865,
27
+ "grad_norm": 12.481098175048828,
28
+ "learning_rate": 1.42e-06,
29
+ "loss": 1.3152,
30
+ "step": 75
31
+ },
32
+ {
33
+ "epoch": 0.16474464579901152,
34
+ "grad_norm": 8.900260925292969,
35
+ "learning_rate": 1.9200000000000003e-06,
36
+ "loss": 0.7248,
37
+ "step": 100
38
+ },
39
+ {
40
+ "epoch": 0.20593080724876442,
41
+ "grad_norm": 7.8956146240234375,
42
+ "learning_rate": 2.42e-06,
43
+ "loss": 0.5731,
44
+ "step": 125
45
+ },
46
+ {
47
+ "epoch": 0.2471169686985173,
48
+ "grad_norm": 9.542287826538086,
49
+ "learning_rate": 2.92e-06,
50
+ "loss": 0.474,
51
+ "step": 150
52
+ },
53
+ {
54
+ "epoch": 0.2883031301482702,
55
+ "grad_norm": 6.937229633331299,
56
+ "learning_rate": 3.4200000000000007e-06,
57
+ "loss": 0.4301,
58
+ "step": 175
59
+ },
60
+ {
61
+ "epoch": 0.32948929159802304,
62
+ "grad_norm": 6.919783115386963,
63
+ "learning_rate": 3.920000000000001e-06,
64
+ "loss": 0.351,
65
+ "step": 200
66
+ },
67
+ {
68
+ "epoch": 0.37067545304777594,
69
+ "grad_norm": 6.427859783172607,
70
+ "learning_rate": 4.42e-06,
71
+ "loss": 0.2726,
72
+ "step": 225
73
+ },
74
+ {
75
+ "epoch": 0.41186161449752884,
76
+ "grad_norm": 5.83529806137085,
77
+ "learning_rate": 4.92e-06,
78
+ "loss": 0.1663,
79
+ "step": 250
80
+ },
81
+ {
82
+ "epoch": 0.45304777594728174,
83
+ "grad_norm": 2.4859721660614014,
84
+ "learning_rate": 5.420000000000001e-06,
85
+ "loss": 0.0893,
86
+ "step": 275
87
+ },
88
+ {
89
+ "epoch": 0.4942339373970346,
90
+ "grad_norm": 3.2526772022247314,
91
+ "learning_rate": 5.92e-06,
92
+ "loss": 0.076,
93
+ "step": 300
94
+ },
95
+ {
96
+ "epoch": 0.5354200988467874,
97
+ "grad_norm": 4.417870044708252,
98
+ "learning_rate": 6.42e-06,
99
+ "loss": 0.0764,
100
+ "step": 325
101
+ },
102
+ {
103
+ "epoch": 0.5766062602965404,
104
+ "grad_norm": 3.149733781814575,
105
+ "learning_rate": 6.92e-06,
106
+ "loss": 0.0679,
107
+ "step": 350
108
+ },
109
+ {
110
+ "epoch": 0.6177924217462932,
111
+ "grad_norm": 3.8823468685150146,
112
+ "learning_rate": 7.420000000000001e-06,
113
+ "loss": 0.0525,
114
+ "step": 375
115
+ },
116
+ {
117
+ "epoch": 0.6589785831960461,
118
+ "grad_norm": 6.1009745597839355,
119
+ "learning_rate": 7.92e-06,
120
+ "loss": 0.0552,
121
+ "step": 400
122
+ },
123
+ {
124
+ "epoch": 0.700164744645799,
125
+ "grad_norm": 5.714720726013184,
126
+ "learning_rate": 8.42e-06,
127
+ "loss": 0.0529,
128
+ "step": 425
129
+ },
130
+ {
131
+ "epoch": 0.7413509060955519,
132
+ "grad_norm": 4.403444290161133,
133
+ "learning_rate": 8.920000000000001e-06,
134
+ "loss": 0.0472,
135
+ "step": 450
136
+ },
137
+ {
138
+ "epoch": 0.7825370675453048,
139
+ "grad_norm": 3.7249884605407715,
140
+ "learning_rate": 9.42e-06,
141
+ "loss": 0.0424,
142
+ "step": 475
143
+ },
144
+ {
145
+ "epoch": 0.8237232289950577,
146
+ "grad_norm": 3.8490548133850098,
147
+ "learning_rate": 9.920000000000002e-06,
148
+ "loss": 0.0461,
149
+ "step": 500
150
+ },
151
+ {
152
+ "epoch": 0.8237232289950577,
153
+ "eval_loss": 0.04240331053733826,
154
+ "eval_runtime": 1653.3644,
155
+ "eval_samples_per_second": 1.468,
156
+ "eval_steps_per_second": 0.184,
157
+ "eval_wer": 11.04637613609881,
158
+ "step": 500
159
+ },
160
+ {
161
+ "epoch": 0.8649093904448105,
162
+ "grad_norm": 2.538515329360962,
163
+ "learning_rate": 9.961818181818183e-06,
164
+ "loss": 0.0384,
165
+ "step": 525
166
+ },
167
+ {
168
+ "epoch": 0.9060955518945635,
169
+ "grad_norm": 1.6598188877105713,
170
+ "learning_rate": 9.916363636363637e-06,
171
+ "loss": 0.0375,
172
+ "step": 550
173
+ },
174
+ {
175
+ "epoch": 0.9472817133443163,
176
+ "grad_norm": 3.256746292114258,
177
+ "learning_rate": 9.870909090909092e-06,
178
+ "loss": 0.0346,
179
+ "step": 575
180
+ },
181
+ {
182
+ "epoch": 0.9884678747940692,
183
+ "grad_norm": 3.04457426071167,
184
+ "learning_rate": 9.825454545454546e-06,
185
+ "loss": 0.034,
186
+ "step": 600
187
+ },
188
+ {
189
+ "epoch": 1.029654036243822,
190
+ "grad_norm": 0.6082770228385925,
191
+ "learning_rate": 9.780000000000001e-06,
192
+ "loss": 0.0272,
193
+ "step": 625
194
+ },
195
+ {
196
+ "epoch": 1.0708401976935749,
197
+ "grad_norm": 1.6969329118728638,
198
+ "learning_rate": 9.734545454545455e-06,
199
+ "loss": 0.0206,
200
+ "step": 650
201
+ },
202
+ {
203
+ "epoch": 1.112026359143328,
204
+ "grad_norm": 1.9652377367019653,
205
+ "learning_rate": 9.68909090909091e-06,
206
+ "loss": 0.0197,
207
+ "step": 675
208
+ },
209
+ {
210
+ "epoch": 1.1532125205930808,
211
+ "grad_norm": 2.402329206466675,
212
+ "learning_rate": 9.643636363636364e-06,
213
+ "loss": 0.0221,
214
+ "step": 700
215
+ },
216
+ {
217
+ "epoch": 1.1943986820428336,
218
+ "grad_norm": 3.167133092880249,
219
+ "learning_rate": 9.598181818181818e-06,
220
+ "loss": 0.0176,
221
+ "step": 725
222
+ },
223
+ {
224
+ "epoch": 1.2355848434925865,
225
+ "grad_norm": 0.4075870215892792,
226
+ "learning_rate": 9.552727272727273e-06,
227
+ "loss": 0.0165,
228
+ "step": 750
229
+ },
230
+ {
231
+ "epoch": 1.2767710049423393,
232
+ "grad_norm": 2.739176034927368,
233
+ "learning_rate": 9.507272727272729e-06,
234
+ "loss": 0.0156,
235
+ "step": 775
236
+ },
237
+ {
238
+ "epoch": 1.3179571663920924,
239
+ "grad_norm": 2.4662702083587646,
240
+ "learning_rate": 9.461818181818183e-06,
241
+ "loss": 0.0133,
242
+ "step": 800
243
+ },
244
+ {
245
+ "epoch": 1.3591433278418452,
246
+ "grad_norm": 2.955909252166748,
247
+ "learning_rate": 9.416363636363636e-06,
248
+ "loss": 0.0184,
249
+ "step": 825
250
+ },
251
+ {
252
+ "epoch": 1.400329489291598,
253
+ "grad_norm": 2.0838463306427,
254
+ "learning_rate": 9.370909090909092e-06,
255
+ "loss": 0.0138,
256
+ "step": 850
257
+ },
258
+ {
259
+ "epoch": 1.441515650741351,
260
+ "grad_norm": 2.2245779037475586,
261
+ "learning_rate": 9.325454545454547e-06,
262
+ "loss": 0.0132,
263
+ "step": 875
264
+ },
265
+ {
266
+ "epoch": 1.4827018121911038,
267
+ "grad_norm": 2.146961212158203,
268
+ "learning_rate": 9.280000000000001e-06,
269
+ "loss": 0.0126,
270
+ "step": 900
271
+ },
272
+ {
273
+ "epoch": 1.5238879736408566,
274
+ "grad_norm": 1.955930471420288,
275
+ "learning_rate": 9.234545454545455e-06,
276
+ "loss": 0.0111,
277
+ "step": 925
278
+ },
279
+ {
280
+ "epoch": 1.5650741350906094,
281
+ "grad_norm": 2.04664945602417,
282
+ "learning_rate": 9.18909090909091e-06,
283
+ "loss": 0.0118,
284
+ "step": 950
285
+ },
286
+ {
287
+ "epoch": 1.6062602965403623,
288
+ "grad_norm": 2.1349849700927734,
289
+ "learning_rate": 9.143636363636364e-06,
290
+ "loss": 0.0116,
291
+ "step": 975
292
+ },
293
+ {
294
+ "epoch": 1.6474464579901154,
295
+ "grad_norm": 2.336024761199951,
296
+ "learning_rate": 9.09818181818182e-06,
297
+ "loss": 0.0124,
298
+ "step": 1000
299
+ },
300
+ {
301
+ "epoch": 1.6474464579901154,
302
+ "eval_loss": 0.02211674302816391,
303
+ "eval_runtime": 1633.7086,
304
+ "eval_samples_per_second": 1.486,
305
+ "eval_steps_per_second": 0.186,
306
+ "eval_wer": 5.11147362697118,
307
+ "step": 1000
308
+ },
309
+ {
310
+ "epoch": 1.6886326194398682,
311
+ "grad_norm": 0.4564249813556671,
312
+ "learning_rate": 9.052727272727273e-06,
313
+ "loss": 0.0119,
314
+ "step": 1025
315
+ },
316
+ {
317
+ "epoch": 1.729818780889621,
318
+ "grad_norm": 1.9494842290878296,
319
+ "learning_rate": 9.007272727272729e-06,
320
+ "loss": 0.0114,
321
+ "step": 1050
322
+ },
323
+ {
324
+ "epoch": 1.771004942339374,
325
+ "grad_norm": 1.7129623889923096,
326
+ "learning_rate": 8.961818181818182e-06,
327
+ "loss": 0.0116,
328
+ "step": 1075
329
+ },
330
+ {
331
+ "epoch": 1.812191103789127,
332
+ "grad_norm": 0.8837174773216248,
333
+ "learning_rate": 8.916363636363638e-06,
334
+ "loss": 0.01,
335
+ "step": 1100
336
+ },
337
+ {
338
+ "epoch": 1.8533772652388798,
339
+ "grad_norm": 3.1779348850250244,
340
+ "learning_rate": 8.870909090909091e-06,
341
+ "loss": 0.0087,
342
+ "step": 1125
343
+ },
344
+ {
345
+ "epoch": 1.8945634266886326,
346
+ "grad_norm": 1.1833815574645996,
347
+ "learning_rate": 8.825454545454545e-06,
348
+ "loss": 0.0122,
349
+ "step": 1150
350
+ },
351
+ {
352
+ "epoch": 1.9357495881383855,
353
+ "grad_norm": 2.3456814289093018,
354
+ "learning_rate": 8.78e-06,
355
+ "loss": 0.0089,
356
+ "step": 1175
357
+ },
358
+ {
359
+ "epoch": 1.9769357495881383,
360
+ "grad_norm": 1.5834710597991943,
361
+ "learning_rate": 8.734545454545456e-06,
362
+ "loss": 0.0111,
363
+ "step": 1200
364
+ },
365
+ {
366
+ "epoch": 2.018121911037891,
367
+ "grad_norm": 1.1926366090774536,
368
+ "learning_rate": 8.68909090909091e-06,
369
+ "loss": 0.0069,
370
+ "step": 1225
371
+ },
372
+ {
373
+ "epoch": 2.059308072487644,
374
+ "grad_norm": 0.5671884417533875,
375
+ "learning_rate": 8.643636363636364e-06,
376
+ "loss": 0.0043,
377
+ "step": 1250
378
+ },
379
+ {
380
+ "epoch": 2.100494233937397,
381
+ "grad_norm": 0.8897690176963806,
382
+ "learning_rate": 8.598181818181819e-06,
383
+ "loss": 0.0034,
384
+ "step": 1275
385
+ },
386
+ {
387
+ "epoch": 2.1416803953871497,
388
+ "grad_norm": 1.6989620923995972,
389
+ "learning_rate": 8.552727272727274e-06,
390
+ "loss": 0.003,
391
+ "step": 1300
392
+ },
393
+ {
394
+ "epoch": 2.182866556836903,
395
+ "grad_norm": 2.0327303409576416,
396
+ "learning_rate": 8.507272727272728e-06,
397
+ "loss": 0.0037,
398
+ "step": 1325
399
+ },
400
+ {
401
+ "epoch": 2.224052718286656,
402
+ "grad_norm": 0.6554267406463623,
403
+ "learning_rate": 8.461818181818182e-06,
404
+ "loss": 0.0034,
405
+ "step": 1350
406
+ },
407
+ {
408
+ "epoch": 2.2652388797364087,
409
+ "grad_norm": 0.21176663041114807,
410
+ "learning_rate": 8.416363636363637e-06,
411
+ "loss": 0.0046,
412
+ "step": 1375
413
+ },
414
+ {
415
+ "epoch": 2.3064250411861615,
416
+ "grad_norm": 1.2915725708007812,
417
+ "learning_rate": 8.370909090909091e-06,
418
+ "loss": 0.0026,
419
+ "step": 1400
420
+ },
421
+ {
422
+ "epoch": 2.3476112026359144,
423
+ "grad_norm": 0.15533868968486786,
424
+ "learning_rate": 8.325454545454547e-06,
425
+ "loss": 0.0043,
426
+ "step": 1425
427
+ },
428
+ {
429
+ "epoch": 2.3887973640856672,
430
+ "grad_norm": 0.19825848937034607,
431
+ "learning_rate": 8.28e-06,
432
+ "loss": 0.0042,
433
+ "step": 1450
434
+ },
435
+ {
436
+ "epoch": 2.42998352553542,
437
+ "grad_norm": 0.14249640703201294,
438
+ "learning_rate": 8.234545454545456e-06,
439
+ "loss": 0.0023,
440
+ "step": 1475
441
+ },
442
+ {
443
+ "epoch": 2.471169686985173,
444
+ "grad_norm": 0.12349322438240051,
445
+ "learning_rate": 8.18909090909091e-06,
446
+ "loss": 0.0026,
447
+ "step": 1500
448
+ },
449
+ {
450
+ "epoch": 2.471169686985173,
451
+ "eval_loss": 0.015178721398115158,
452
+ "eval_runtime": 1647.8725,
453
+ "eval_samples_per_second": 1.473,
454
+ "eval_steps_per_second": 0.184,
455
+ "eval_wer": 3.5966752116833685,
456
+ "step": 1500
457
+ },
458
+ {
459
+ "epoch": 2.5123558484349258,
460
+ "grad_norm": 0.24572087824344635,
461
+ "learning_rate": 8.143636363636365e-06,
462
+ "loss": 0.0032,
463
+ "step": 1525
464
+ },
465
+ {
466
+ "epoch": 2.5535420098846786,
467
+ "grad_norm": 0.2075294703245163,
468
+ "learning_rate": 8.098181818181819e-06,
469
+ "loss": 0.0023,
470
+ "step": 1550
471
+ },
472
+ {
473
+ "epoch": 2.594728171334432,
474
+ "grad_norm": 0.264713317155838,
475
+ "learning_rate": 8.052727272727272e-06,
476
+ "loss": 0.0044,
477
+ "step": 1575
478
+ },
479
+ {
480
+ "epoch": 2.6359143327841847,
481
+ "grad_norm": 1.9720451831817627,
482
+ "learning_rate": 8.007272727272728e-06,
483
+ "loss": 0.0048,
484
+ "step": 1600
485
+ },
486
+ {
487
+ "epoch": 2.6771004942339376,
488
+ "grad_norm": 0.7824451327323914,
489
+ "learning_rate": 7.961818181818183e-06,
490
+ "loss": 0.0048,
491
+ "step": 1625
492
+ },
493
+ {
494
+ "epoch": 2.7182866556836904,
495
+ "grad_norm": 0.681707501411438,
496
+ "learning_rate": 7.916363636363637e-06,
497
+ "loss": 0.0036,
498
+ "step": 1650
499
+ },
500
+ {
501
+ "epoch": 2.7594728171334433,
502
+ "grad_norm": 0.39950481057167053,
503
+ "learning_rate": 7.870909090909091e-06,
504
+ "loss": 0.0047,
505
+ "step": 1675
506
+ },
507
+ {
508
+ "epoch": 2.800658978583196,
509
+ "grad_norm": 0.7777596712112427,
510
+ "learning_rate": 7.825454545454546e-06,
511
+ "loss": 0.0029,
512
+ "step": 1700
513
+ },
514
+ {
515
+ "epoch": 2.841845140032949,
516
+ "grad_norm": 0.06510231643915176,
517
+ "learning_rate": 7.78e-06,
518
+ "loss": 0.0067,
519
+ "step": 1725
520
+ },
521
+ {
522
+ "epoch": 2.883031301482702,
523
+ "grad_norm": 2.2729947566986084,
524
+ "learning_rate": 7.734545454545455e-06,
525
+ "loss": 0.0024,
526
+ "step": 1750
527
+ },
528
+ {
529
+ "epoch": 2.9242174629324547,
530
+ "grad_norm": 1.6635749340057373,
531
+ "learning_rate": 7.68909090909091e-06,
532
+ "loss": 0.0035,
533
+ "step": 1775
534
+ },
535
+ {
536
+ "epoch": 2.9654036243822075,
537
+ "grad_norm": 0.06815456598997116,
538
+ "learning_rate": 7.643636363636365e-06,
539
+ "loss": 0.004,
540
+ "step": 1800
541
+ },
542
+ {
543
+ "epoch": 3.0065897858319603,
544
+ "grad_norm": 1.5607892274856567,
545
+ "learning_rate": 7.598181818181819e-06,
546
+ "loss": 0.0046,
547
+ "step": 1825
548
+ },
549
+ {
550
+ "epoch": 3.047775947281713,
551
+ "grad_norm": 0.8395469784736633,
552
+ "learning_rate": 7.552727272727274e-06,
553
+ "loss": 0.0018,
554
+ "step": 1850
555
+ },
556
+ {
557
+ "epoch": 3.088962108731466,
558
+ "grad_norm": 0.04032081738114357,
559
+ "learning_rate": 7.507272727272728e-06,
560
+ "loss": 0.0011,
561
+ "step": 1875
562
+ },
563
+ {
564
+ "epoch": 3.130148270181219,
565
+ "grad_norm": 0.3644678294658661,
566
+ "learning_rate": 7.461818181818182e-06,
567
+ "loss": 0.0014,
568
+ "step": 1900
569
+ },
570
+ {
571
+ "epoch": 3.171334431630972,
572
+ "grad_norm": 0.35729554295539856,
573
+ "learning_rate": 7.416363636363637e-06,
574
+ "loss": 0.0009,
575
+ "step": 1925
576
+ },
577
+ {
578
+ "epoch": 3.212520593080725,
579
+ "grad_norm": 0.05964287370443344,
580
+ "learning_rate": 7.370909090909092e-06,
581
+ "loss": 0.0014,
582
+ "step": 1950
583
+ },
584
+ {
585
+ "epoch": 3.253706754530478,
586
+ "grad_norm": 0.799541711807251,
587
+ "learning_rate": 7.325454545454546e-06,
588
+ "loss": 0.0015,
589
+ "step": 1975
590
+ },
591
+ {
592
+ "epoch": 3.2948929159802307,
593
+ "grad_norm": 0.058581918478012085,
594
+ "learning_rate": 7.280000000000001e-06,
595
+ "loss": 0.0008,
596
+ "step": 2000
597
+ },
598
+ {
599
+ "epoch": 3.2948929159802307,
600
+ "eval_loss": 0.015332825481891632,
601
+ "eval_runtime": 1629.67,
602
+ "eval_samples_per_second": 1.489,
603
+ "eval_steps_per_second": 0.187,
604
+ "eval_wer": 3.24710634661695,
605
+ "step": 2000
606
+ },
607
+ {
608
+ "epoch": 3.3360790774299836,
609
+ "grad_norm": 0.20925483107566833,
610
+ "learning_rate": 7.234545454545455e-06,
611
+ "loss": 0.001,
612
+ "step": 2025
613
+ },
614
+ {
615
+ "epoch": 3.3772652388797364,
616
+ "grad_norm": 0.5220603346824646,
617
+ "learning_rate": 7.18909090909091e-06,
618
+ "loss": 0.0006,
619
+ "step": 2050
620
+ },
621
+ {
622
+ "epoch": 3.4184514003294892,
623
+ "grad_norm": 0.33813387155532837,
624
+ "learning_rate": 7.1436363636363635e-06,
625
+ "loss": 0.0008,
626
+ "step": 2075
627
+ },
628
+ {
629
+ "epoch": 3.459637561779242,
630
+ "grad_norm": 0.5609467625617981,
631
+ "learning_rate": 7.098181818181818e-06,
632
+ "loss": 0.0006,
633
+ "step": 2100
634
+ },
635
+ {
636
+ "epoch": 3.500823723228995,
637
+ "grad_norm": 0.028300393372774124,
638
+ "learning_rate": 7.052727272727274e-06,
639
+ "loss": 0.0008,
640
+ "step": 2125
641
+ },
642
+ {
643
+ "epoch": 3.5420098846787478,
644
+ "grad_norm": 0.056196194142103195,
645
+ "learning_rate": 7.007272727272728e-06,
646
+ "loss": 0.0009,
647
+ "step": 2150
648
+ },
649
+ {
650
+ "epoch": 3.583196046128501,
651
+ "grad_norm": 0.202606201171875,
652
+ "learning_rate": 6.961818181818183e-06,
653
+ "loss": 0.0019,
654
+ "step": 2175
655
+ },
656
+ {
657
+ "epoch": 3.624382207578254,
658
+ "grad_norm": 0.014030909165740013,
659
+ "learning_rate": 6.9163636363636365e-06,
660
+ "loss": 0.001,
661
+ "step": 2200
662
+ },
663
+ {
664
+ "epoch": 3.6655683690280068,
665
+ "grad_norm": 0.08081363141536713,
666
+ "learning_rate": 6.870909090909091e-06,
667
+ "loss": 0.002,
668
+ "step": 2225
669
+ },
670
+ {
671
+ "epoch": 3.7067545304777596,
672
+ "grad_norm": 0.6110771894454956,
673
+ "learning_rate": 6.8254545454545466e-06,
674
+ "loss": 0.0005,
675
+ "step": 2250
676
+ },
677
+ {
678
+ "epoch": 3.7479406919275124,
679
+ "grad_norm": 0.30081290006637573,
680
+ "learning_rate": 6.780000000000001e-06,
681
+ "loss": 0.0021,
682
+ "step": 2275
683
+ },
684
+ {
685
+ "epoch": 3.7891268533772653,
686
+ "grad_norm": 0.042301613837480545,
687
+ "learning_rate": 6.734545454545455e-06,
688
+ "loss": 0.0012,
689
+ "step": 2300
690
+ },
691
+ {
692
+ "epoch": 3.830313014827018,
693
+ "grad_norm": 1.5658293962478638,
694
+ "learning_rate": 6.6890909090909095e-06,
695
+ "loss": 0.0018,
696
+ "step": 2325
697
+ },
698
+ {
699
+ "epoch": 3.871499176276771,
700
+ "grad_norm": 0.20856991410255432,
701
+ "learning_rate": 6.643636363636364e-06,
702
+ "loss": 0.0017,
703
+ "step": 2350
704
+ },
705
+ {
706
+ "epoch": 3.912685337726524,
707
+ "grad_norm": 0.19910617172718048,
708
+ "learning_rate": 6.5981818181818195e-06,
709
+ "loss": 0.0011,
710
+ "step": 2375
711
+ },
712
+ {
713
+ "epoch": 3.9538714991762767,
714
+ "grad_norm": 0.1468815952539444,
715
+ "learning_rate": 6.5527272727272724e-06,
716
+ "loss": 0.0011,
717
+ "step": 2400
718
+ },
719
+ {
720
+ "epoch": 3.9950576606260295,
721
+ "grad_norm": 0.06176592409610748,
722
+ "learning_rate": 6.507272727272728e-06,
723
+ "loss": 0.0009,
724
+ "step": 2425
725
+ },
726
+ {
727
+ "epoch": 4.036243822075782,
728
+ "grad_norm": 0.013582455925643444,
729
+ "learning_rate": 6.4618181818181825e-06,
730
+ "loss": 0.0004,
731
+ "step": 2450
732
+ },
733
+ {
734
+ "epoch": 4.077429983525535,
735
+ "grad_norm": 0.013000385835766792,
736
+ "learning_rate": 6.416363636363637e-06,
737
+ "loss": 0.0003,
738
+ "step": 2475
739
+ },
740
+ {
741
+ "epoch": 4.118616144975288,
742
+ "grad_norm": 0.04700480401515961,
743
+ "learning_rate": 6.370909090909091e-06,
744
+ "loss": 0.0004,
745
+ "step": 2500
746
+ },
747
+ {
748
+ "epoch": 4.118616144975288,
749
+ "eval_loss": 0.013498911634087563,
750
+ "eval_runtime": 1637.8085,
751
+ "eval_samples_per_second": 1.482,
752
+ "eval_steps_per_second": 0.186,
753
+ "eval_wer": 2.718868950516585,
754
+ "step": 2500
755
+ },
756
+ {
757
+ "epoch": 4.159802306425041,
758
+ "grad_norm": 0.049776773899793625,
759
+ "learning_rate": 6.325454545454545e-06,
760
+ "loss": 0.0007,
761
+ "step": 2525
762
+ },
763
+ {
764
+ "epoch": 4.200988467874794,
765
+ "grad_norm": 0.047507863491773605,
766
+ "learning_rate": 6.280000000000001e-06,
767
+ "loss": 0.001,
768
+ "step": 2550
769
+ },
770
+ {
771
+ "epoch": 4.242174629324547,
772
+ "grad_norm": 0.058455199003219604,
773
+ "learning_rate": 6.2345454545454555e-06,
774
+ "loss": 0.0004,
775
+ "step": 2575
776
+ },
777
+ {
778
+ "epoch": 4.283360790774299,
779
+ "grad_norm": 0.03676324710249901,
780
+ "learning_rate": 6.18909090909091e-06,
781
+ "loss": 0.0008,
782
+ "step": 2600
783
+ },
784
+ {
785
+ "epoch": 4.324546952224052,
786
+ "grad_norm": 0.08864962309598923,
787
+ "learning_rate": 6.143636363636364e-06,
788
+ "loss": 0.0008,
789
+ "step": 2625
790
+ },
791
+ {
792
+ "epoch": 4.365733113673806,
793
+ "grad_norm": 0.25547558069229126,
794
+ "learning_rate": 6.098181818181818e-06,
795
+ "loss": 0.0014,
796
+ "step": 2650
797
+ },
798
+ {
799
+ "epoch": 4.406919275123559,
800
+ "grad_norm": 0.014656963758170605,
801
+ "learning_rate": 6.052727272727274e-06,
802
+ "loss": 0.0004,
803
+ "step": 2675
804
+ },
805
+ {
806
+ "epoch": 4.448105436573312,
807
+ "grad_norm": 0.01910262741148472,
808
+ "learning_rate": 6.0072727272727284e-06,
809
+ "loss": 0.0004,
810
+ "step": 2700
811
+ },
812
+ {
813
+ "epoch": 4.4892915980230645,
814
+ "grad_norm": 0.021646469831466675,
815
+ "learning_rate": 5.961818181818182e-06,
816
+ "loss": 0.0005,
817
+ "step": 2725
818
+ },
819
+ {
820
+ "epoch": 4.530477759472817,
821
+ "grad_norm": 0.024704741314053535,
822
+ "learning_rate": 5.916363636363637e-06,
823
+ "loss": 0.0005,
824
+ "step": 2750
825
+ },
826
+ {
827
+ "epoch": 4.57166392092257,
828
+ "grad_norm": 0.024691058322787285,
829
+ "learning_rate": 5.870909090909091e-06,
830
+ "loss": 0.0004,
831
+ "step": 2775
832
+ },
833
+ {
834
+ "epoch": 4.612850082372323,
835
+ "grad_norm": 0.023056305944919586,
836
+ "learning_rate": 5.825454545454546e-06,
837
+ "loss": 0.0002,
838
+ "step": 2800
839
+ },
840
+ {
841
+ "epoch": 4.654036243822076,
842
+ "grad_norm": 0.02144230715930462,
843
+ "learning_rate": 5.78e-06,
844
+ "loss": 0.0006,
845
+ "step": 2825
846
+ },
847
+ {
848
+ "epoch": 4.695222405271829,
849
+ "grad_norm": 0.056895911693573,
850
+ "learning_rate": 5.734545454545455e-06,
851
+ "loss": 0.0006,
852
+ "step": 2850
853
+ },
854
+ {
855
+ "epoch": 4.736408566721582,
856
+ "grad_norm": 0.026411665603518486,
857
+ "learning_rate": 5.68909090909091e-06,
858
+ "loss": 0.0002,
859
+ "step": 2875
860
+ },
861
+ {
862
+ "epoch": 4.7775947281713345,
863
+ "grad_norm": 0.6490919589996338,
864
+ "learning_rate": 5.643636363636364e-06,
865
+ "loss": 0.0005,
866
+ "step": 2900
867
+ },
868
+ {
869
+ "epoch": 4.818780889621087,
870
+ "grad_norm": 0.01870059408247471,
871
+ "learning_rate": 5.598181818181818e-06,
872
+ "loss": 0.0004,
873
+ "step": 2925
874
+ },
875
+ {
876
+ "epoch": 4.85996705107084,
877
+ "grad_norm": 0.36998653411865234,
878
+ "learning_rate": 5.552727272727273e-06,
879
+ "loss": 0.001,
880
+ "step": 2950
881
+ },
882
+ {
883
+ "epoch": 4.901153212520593,
884
+ "grad_norm": 0.09682920575141907,
885
+ "learning_rate": 5.507272727272728e-06,
886
+ "loss": 0.0008,
887
+ "step": 2975
888
+ },
889
+ {
890
+ "epoch": 4.942339373970346,
891
+ "grad_norm": 0.12028121948242188,
892
+ "learning_rate": 5.461818181818183e-06,
893
+ "loss": 0.0005,
894
+ "step": 3000
895
+ },
896
+ {
897
+ "epoch": 4.942339373970346,
898
+ "eval_loss": 0.013804786838591099,
899
+ "eval_runtime": 1636.262,
900
+ "eval_samples_per_second": 1.483,
901
+ "eval_steps_per_second": 0.186,
902
+ "eval_wer": 2.8043191175328204,
903
+ "step": 3000
904
+ },
905
+ {
906
+ "epoch": 4.983525535420099,
907
+ "grad_norm": 0.36403629183769226,
908
+ "learning_rate": 5.4163636363636365e-06,
909
+ "loss": 0.0002,
910
+ "step": 3025
911
+ },
912
+ {
913
+ "epoch": 5.0247116968698515,
914
+ "grad_norm": 0.020813830196857452,
915
+ "learning_rate": 5.370909090909091e-06,
916
+ "loss": 0.0007,
917
+ "step": 3050
918
+ },
919
+ {
920
+ "epoch": 5.065897858319604,
921
+ "grad_norm": 0.0156087102368474,
922
+ "learning_rate": 5.325454545454546e-06,
923
+ "loss": 0.0002,
924
+ "step": 3075
925
+ },
926
+ {
927
+ "epoch": 5.107084019769357,
928
+ "grad_norm": 0.08841024339199066,
929
+ "learning_rate": 5.28e-06,
930
+ "loss": 0.0003,
931
+ "step": 3100
932
+ },
933
+ {
934
+ "epoch": 5.14827018121911,
935
+ "grad_norm": 0.09709523618221283,
936
+ "learning_rate": 5.234545454545456e-06,
937
+ "loss": 0.0002,
938
+ "step": 3125
939
+ },
940
+ {
941
+ "epoch": 5.189456342668863,
942
+ "grad_norm": 0.006491546519100666,
943
+ "learning_rate": 5.1890909090909095e-06,
944
+ "loss": 0.0004,
945
+ "step": 3150
946
+ },
947
+ {
948
+ "epoch": 5.230642504118616,
949
+ "grad_norm": 0.03339486941695213,
950
+ "learning_rate": 5.143636363636364e-06,
951
+ "loss": 0.0001,
952
+ "step": 3175
953
+ },
954
+ {
955
+ "epoch": 5.2718286655683695,
956
+ "grad_norm": 0.019623372703790665,
957
+ "learning_rate": 5.098181818181819e-06,
958
+ "loss": 0.0002,
959
+ "step": 3200
960
+ },
961
+ {
962
+ "epoch": 5.313014827018122,
963
+ "grad_norm": 0.012536031194031239,
964
+ "learning_rate": 5.052727272727273e-06,
965
+ "loss": 0.0004,
966
+ "step": 3225
967
+ },
968
+ {
969
+ "epoch": 5.354200988467875,
970
+ "grad_norm": 0.00885727908462286,
971
+ "learning_rate": 5.007272727272727e-06,
972
+ "loss": 0.0001,
973
+ "step": 3250
974
+ },
975
+ {
976
+ "epoch": 5.395387149917628,
977
+ "grad_norm": 0.012449279427528381,
978
+ "learning_rate": 4.9618181818181824e-06,
979
+ "loss": 0.0003,
980
+ "step": 3275
981
+ },
982
+ {
983
+ "epoch": 5.436573311367381,
984
+ "grad_norm": 0.04439382627606392,
985
+ "learning_rate": 4.916363636363637e-06,
986
+ "loss": 0.0003,
987
+ "step": 3300
988
+ },
989
+ {
990
+ "epoch": 5.477759472817134,
991
+ "grad_norm": 0.08744444698095322,
992
+ "learning_rate": 4.870909090909091e-06,
993
+ "loss": 0.0006,
994
+ "step": 3325
995
+ },
996
+ {
997
+ "epoch": 5.518945634266887,
998
+ "grad_norm": 0.015908684581518173,
999
+ "learning_rate": 4.825454545454546e-06,
1000
+ "loss": 0.0003,
1001
+ "step": 3350
1002
+ },
1003
+ {
1004
+ "epoch": 5.560131795716639,
1005
+ "grad_norm": 0.04740598052740097,
1006
+ "learning_rate": 4.78e-06,
1007
+ "loss": 0.0002,
1008
+ "step": 3375
1009
+ },
1010
+ {
1011
+ "epoch": 5.601317957166392,
1012
+ "grad_norm": 0.011928090825676918,
1013
+ "learning_rate": 4.734545454545455e-06,
1014
+ "loss": 0.0001,
1015
+ "step": 3400
1016
+ },
1017
+ {
1018
+ "epoch": 5.642504118616145,
1019
+ "grad_norm": 0.044589392840862274,
1020
+ "learning_rate": 4.689090909090909e-06,
1021
+ "loss": 0.0002,
1022
+ "step": 3425
1023
+ },
1024
+ {
1025
+ "epoch": 5.683690280065898,
1026
+ "grad_norm": 0.010786905884742737,
1027
+ "learning_rate": 4.643636363636364e-06,
1028
+ "loss": 0.0001,
1029
+ "step": 3450
1030
+ },
1031
+ {
1032
+ "epoch": 5.724876441515651,
1033
+ "grad_norm": 0.006343629211187363,
1034
+ "learning_rate": 4.598181818181818e-06,
1035
+ "loss": 0.0003,
1036
+ "step": 3475
1037
+ },
1038
+ {
1039
+ "epoch": 5.766062602965404,
1040
+ "grad_norm": 0.006056615151464939,
1041
+ "learning_rate": 4.552727272727273e-06,
1042
+ "loss": 0.0003,
1043
+ "step": 3500
1044
+ },
1045
+ {
1046
+ "epoch": 5.766062602965404,
1047
+ "eval_loss": 0.013986092992126942,
1048
+ "eval_runtime": 1723.588,
1049
+ "eval_samples_per_second": 1.408,
1050
+ "eval_steps_per_second": 0.176,
1051
+ "eval_wer": 2.734405344519537,
1052
+ "step": 3500
1053
+ },
1054
+ {
1055
+ "epoch": 5.8072487644151565,
1056
+ "grad_norm": 0.043298039585351944,
1057
+ "learning_rate": 4.5072727272727275e-06,
1058
+ "loss": 0.0002,
1059
+ "step": 3525
1060
+ },
1061
+ {
1062
+ "epoch": 5.848434925864909,
1063
+ "grad_norm": 0.007331861648708582,
1064
+ "learning_rate": 4.461818181818182e-06,
1065
+ "loss": 0.0001,
1066
+ "step": 3550
1067
+ },
1068
+ {
1069
+ "epoch": 5.889621087314662,
1070
+ "grad_norm": 0.012765350751578808,
1071
+ "learning_rate": 4.416363636363637e-06,
1072
+ "loss": 0.0002,
1073
+ "step": 3575
1074
+ },
1075
+ {
1076
+ "epoch": 5.930807248764415,
1077
+ "grad_norm": 0.033175356686115265,
1078
+ "learning_rate": 4.370909090909091e-06,
1079
+ "loss": 0.0004,
1080
+ "step": 3600
1081
+ },
1082
+ {
1083
+ "epoch": 5.971993410214168,
1084
+ "grad_norm": 0.005575309973210096,
1085
+ "learning_rate": 4.325454545454546e-06,
1086
+ "loss": 0.0001,
1087
+ "step": 3625
1088
+ },
1089
+ {
1090
+ "epoch": 6.013179571663921,
1091
+ "grad_norm": 0.005467335693538189,
1092
+ "learning_rate": 4.2800000000000005e-06,
1093
+ "loss": 0.0003,
1094
+ "step": 3650
1095
+ },
1096
+ {
1097
+ "epoch": 6.0543657331136735,
1098
+ "grad_norm": 0.0050591155886650085,
1099
+ "learning_rate": 4.234545454545455e-06,
1100
+ "loss": 0.0001,
1101
+ "step": 3675
1102
+ },
1103
+ {
1104
+ "epoch": 6.095551894563426,
1105
+ "grad_norm": 0.004481570329517126,
1106
+ "learning_rate": 4.18909090909091e-06,
1107
+ "loss": 0.0001,
1108
+ "step": 3700
1109
+ },
1110
+ {
1111
+ "epoch": 6.136738056013179,
1112
+ "grad_norm": 0.002914627082645893,
1113
+ "learning_rate": 4.143636363636364e-06,
1114
+ "loss": 0.0001,
1115
+ "step": 3725
1116
+ },
1117
+ {
1118
+ "epoch": 6.177924217462932,
1119
+ "grad_norm": 0.004947616718709469,
1120
+ "learning_rate": 4.098181818181818e-06,
1121
+ "loss": 0.0001,
1122
+ "step": 3750
1123
+ },
1124
+ {
1125
+ "epoch": 6.219110378912685,
1126
+ "grad_norm": 0.003795625874772668,
1127
+ "learning_rate": 4.0527272727272735e-06,
1128
+ "loss": 0.0001,
1129
+ "step": 3775
1130
+ },
1131
+ {
1132
+ "epoch": 6.260296540362438,
1133
+ "grad_norm": 0.025668440386652946,
1134
+ "learning_rate": 4.007272727272727e-06,
1135
+ "loss": 0.0001,
1136
+ "step": 3800
1137
+ },
1138
+ {
1139
+ "epoch": 6.3014827018121915,
1140
+ "grad_norm": 0.01409939769655466,
1141
+ "learning_rate": 3.961818181818182e-06,
1142
+ "loss": 0.0001,
1143
+ "step": 3825
1144
+ },
1145
+ {
1146
+ "epoch": 6.342668863261944,
1147
+ "grad_norm": 0.0054998998530209064,
1148
+ "learning_rate": 3.9163636363636364e-06,
1149
+ "loss": 0.0001,
1150
+ "step": 3850
1151
+ },
1152
+ {
1153
+ "epoch": 6.383855024711697,
1154
+ "grad_norm": 0.0049223750829696655,
1155
+ "learning_rate": 3.870909090909091e-06,
1156
+ "loss": 0.0001,
1157
+ "step": 3875
1158
+ },
1159
+ {
1160
+ "epoch": 6.42504118616145,
1161
+ "grad_norm": 0.0028136519249528646,
1162
+ "learning_rate": 3.825454545454546e-06,
1163
+ "loss": 0.0001,
1164
+ "step": 3900
1165
+ },
1166
+ {
1167
+ "epoch": 6.466227347611203,
1168
+ "grad_norm": 0.005505116190761328,
1169
+ "learning_rate": 3.7800000000000002e-06,
1170
+ "loss": 0.0001,
1171
+ "step": 3925
1172
+ },
1173
+ {
1174
+ "epoch": 6.507413509060956,
1175
+ "grad_norm": 0.007019472308456898,
1176
+ "learning_rate": 3.7345454545454544e-06,
1177
+ "loss": 0.0001,
1178
+ "step": 3950
1179
+ },
1180
+ {
1181
+ "epoch": 6.548599670510709,
1182
+ "grad_norm": 0.006632586009800434,
1183
+ "learning_rate": 3.6890909090909094e-06,
1184
+ "loss": 0.0001,
1185
+ "step": 3975
1186
+ },
1187
+ {
1188
+ "epoch": 6.589785831960461,
1189
+ "grad_norm": 0.002231697551906109,
1190
+ "learning_rate": 3.643636363636364e-06,
1191
+ "loss": 0.0001,
1192
+ "step": 4000
1193
+ },
1194
+ {
1195
+ "epoch": 6.589785831960461,
1196
+ "eval_loss": 0.013741781935095787,
1197
+ "eval_runtime": 1710.3123,
1198
+ "eval_samples_per_second": 1.419,
1199
+ "eval_steps_per_second": 0.178,
1200
+ "eval_wer": 2.532432222481162,
1201
+ "step": 4000
1202
+ },
1203
+ {
1204
+ "epoch": 6.630971993410214,
1205
+ "grad_norm": 0.008344273082911968,
1206
+ "learning_rate": 3.5981818181818186e-06,
1207
+ "loss": 0.0,
1208
+ "step": 4025
1209
+ },
1210
+ {
1211
+ "epoch": 6.672158154859967,
1212
+ "grad_norm": 0.0035063326358795166,
1213
+ "learning_rate": 3.552727272727273e-06,
1214
+ "loss": 0.0,
1215
+ "step": 4050
1216
+ },
1217
+ {
1218
+ "epoch": 6.71334431630972,
1219
+ "grad_norm": 0.007493205368518829,
1220
+ "learning_rate": 3.5072727272727274e-06,
1221
+ "loss": 0.0002,
1222
+ "step": 4075
1223
+ },
1224
+ {
1225
+ "epoch": 6.754530477759473,
1226
+ "grad_norm": 0.013130983337759972,
1227
+ "learning_rate": 3.4618181818181824e-06,
1228
+ "loss": 0.0001,
1229
+ "step": 4100
1230
+ },
1231
+ {
1232
+ "epoch": 6.795716639209226,
1233
+ "grad_norm": 0.022349372506141663,
1234
+ "learning_rate": 3.4163636363636366e-06,
1235
+ "loss": 0.0001,
1236
+ "step": 4125
1237
+ },
1238
+ {
1239
+ "epoch": 6.8369028006589785,
1240
+ "grad_norm": 0.004155050031840801,
1241
+ "learning_rate": 3.370909090909091e-06,
1242
+ "loss": 0.0001,
1243
+ "step": 4150
1244
+ },
1245
+ {
1246
+ "epoch": 6.878088962108731,
1247
+ "grad_norm": 0.004423918202519417,
1248
+ "learning_rate": 3.3254545454545458e-06,
1249
+ "loss": 0.0001,
1250
+ "step": 4175
1251
+ },
1252
+ {
1253
+ "epoch": 6.919275123558484,
1254
+ "grad_norm": 0.010945485904812813,
1255
+ "learning_rate": 3.2800000000000004e-06,
1256
+ "loss": 0.0001,
1257
+ "step": 4200
1258
+ },
1259
+ {
1260
+ "epoch": 6.960461285008237,
1261
+ "grad_norm": 0.00635559344664216,
1262
+ "learning_rate": 3.2345454545454545e-06,
1263
+ "loss": 0.0001,
1264
+ "step": 4225
1265
+ },
1266
+ {
1267
+ "epoch": 7.00164744645799,
1268
+ "grad_norm": 0.0034311157651245594,
1269
+ "learning_rate": 3.1890909090909096e-06,
1270
+ "loss": 0.0003,
1271
+ "step": 4250
1272
+ },
1273
+ {
1274
+ "epoch": 7.042833607907743,
1275
+ "grad_norm": 0.01130327396094799,
1276
+ "learning_rate": 3.1436363636363637e-06,
1277
+ "loss": 0.0,
1278
+ "step": 4275
1279
+ },
1280
+ {
1281
+ "epoch": 7.0840197693574956,
1282
+ "grad_norm": 0.004722132347524166,
1283
+ "learning_rate": 3.0981818181818183e-06,
1284
+ "loss": 0.0,
1285
+ "step": 4300
1286
+ },
1287
+ {
1288
+ "epoch": 7.125205930807248,
1289
+ "grad_norm": 0.004732927773147821,
1290
+ "learning_rate": 3.052727272727273e-06,
1291
+ "loss": 0.0002,
1292
+ "step": 4325
1293
+ },
1294
+ {
1295
+ "epoch": 7.166392092257001,
1296
+ "grad_norm": 0.0030134401749819517,
1297
+ "learning_rate": 3.0072727272727275e-06,
1298
+ "loss": 0.0,
1299
+ "step": 4350
1300
+ },
1301
+ {
1302
+ "epoch": 7.207578253706755,
1303
+ "grad_norm": 0.002724016085267067,
1304
+ "learning_rate": 2.9618181818181817e-06,
1305
+ "loss": 0.0001,
1306
+ "step": 4375
1307
+ },
1308
+ {
1309
+ "epoch": 7.248764415156508,
1310
+ "grad_norm": 0.004742628429085016,
1311
+ "learning_rate": 2.9163636363636367e-06,
1312
+ "loss": 0.0001,
1313
+ "step": 4400
1314
+ },
1315
+ {
1316
+ "epoch": 7.289950576606261,
1317
+ "grad_norm": 0.009233110584318638,
1318
+ "learning_rate": 2.870909090909091e-06,
1319
+ "loss": 0.0001,
1320
+ "step": 4425
1321
+ },
1322
+ {
1323
+ "epoch": 7.3311367380560135,
1324
+ "grad_norm": 0.004101255908608437,
1325
+ "learning_rate": 2.8254545454545455e-06,
1326
+ "loss": 0.0,
1327
+ "step": 4450
1328
+ },
1329
+ {
1330
+ "epoch": 7.372322899505766,
1331
+ "grad_norm": 0.0037873326800763607,
1332
+ "learning_rate": 2.7800000000000005e-06,
1333
+ "loss": 0.0001,
1334
+ "step": 4475
1335
+ },
1336
+ {
1337
+ "epoch": 7.413509060955519,
1338
+ "grad_norm": 0.015564778819680214,
1339
+ "learning_rate": 2.7345454545454547e-06,
1340
+ "loss": 0.0001,
1341
+ "step": 4500
1342
+ },
1343
+ {
1344
+ "epoch": 7.413509060955519,
1345
+ "eval_loss": 0.013747855089604855,
1346
+ "eval_runtime": 1717.6068,
1347
+ "eval_samples_per_second": 1.413,
1348
+ "eval_steps_per_second": 0.177,
1349
+ "eval_wer": 2.5013594344752583,
1350
+ "step": 4500
1351
+ },
1352
+ {
1353
+ "epoch": 7.454695222405272,
1354
+ "grad_norm": 0.0023361605126410723,
1355
+ "learning_rate": 2.6890909090909097e-06,
1356
+ "loss": 0.0002,
1357
+ "step": 4525
1358
+ },
1359
+ {
1360
+ "epoch": 7.495881383855025,
1361
+ "grad_norm": 0.0026130119804292917,
1362
+ "learning_rate": 2.643636363636364e-06,
1363
+ "loss": 0.0001,
1364
+ "step": 4550
1365
+ },
1366
+ {
1367
+ "epoch": 7.537067545304778,
1368
+ "grad_norm": 0.005539346020668745,
1369
+ "learning_rate": 2.5981818181818184e-06,
1370
+ "loss": 0.0001,
1371
+ "step": 4575
1372
+ },
1373
+ {
1374
+ "epoch": 7.578253706754531,
1375
+ "grad_norm": 0.00382342585362494,
1376
+ "learning_rate": 2.552727272727273e-06,
1377
+ "loss": 0.0001,
1378
+ "step": 4600
1379
+ },
1380
+ {
1381
+ "epoch": 7.619439868204283,
1382
+ "grad_norm": 0.0032592774368822575,
1383
+ "learning_rate": 2.5072727272727276e-06,
1384
+ "loss": 0.0001,
1385
+ "step": 4625
1386
+ },
1387
+ {
1388
+ "epoch": 7.660626029654036,
1389
+ "grad_norm": 0.003578683128580451,
1390
+ "learning_rate": 2.461818181818182e-06,
1391
+ "loss": 0.0001,
1392
+ "step": 4650
1393
+ },
1394
+ {
1395
+ "epoch": 7.701812191103789,
1396
+ "grad_norm": 0.006414057221263647,
1397
+ "learning_rate": 2.416363636363637e-06,
1398
+ "loss": 0.0,
1399
+ "step": 4675
1400
+ },
1401
+ {
1402
+ "epoch": 7.742998352553542,
1403
+ "grad_norm": 0.0032208007760345936,
1404
+ "learning_rate": 2.3709090909090914e-06,
1405
+ "loss": 0.0,
1406
+ "step": 4700
1407
+ },
1408
+ {
1409
+ "epoch": 7.784184514003295,
1410
+ "grad_norm": 0.0016529737040400505,
1411
+ "learning_rate": 2.3254545454545456e-06,
1412
+ "loss": 0.0,
1413
+ "step": 4725
1414
+ },
1415
+ {
1416
+ "epoch": 7.825370675453048,
1417
+ "grad_norm": 0.002902130363509059,
1418
+ "learning_rate": 2.28e-06,
1419
+ "loss": 0.0001,
1420
+ "step": 4750
1421
+ },
1422
+ {
1423
+ "epoch": 7.8665568369028005,
1424
+ "grad_norm": 0.002918258076533675,
1425
+ "learning_rate": 2.234545454545455e-06,
1426
+ "loss": 0.0,
1427
+ "step": 4775
1428
+ },
1429
+ {
1430
+ "epoch": 7.907742998352553,
1431
+ "grad_norm": 0.0021400710102170706,
1432
+ "learning_rate": 2.1890909090909094e-06,
1433
+ "loss": 0.0,
1434
+ "step": 4800
1435
+ },
1436
+ {
1437
+ "epoch": 7.948929159802306,
1438
+ "grad_norm": 0.0016904472140595317,
1439
+ "learning_rate": 2.143636363636364e-06,
1440
+ "loss": 0.0,
1441
+ "step": 4825
1442
+ },
1443
+ {
1444
+ "epoch": 7.990115321252059,
1445
+ "grad_norm": 0.006286232266575098,
1446
+ "learning_rate": 2.0981818181818186e-06,
1447
+ "loss": 0.0002,
1448
+ "step": 4850
1449
+ },
1450
+ {
1451
+ "epoch": 8.031301482701812,
1452
+ "grad_norm": 0.004598398692905903,
1453
+ "learning_rate": 2.0527272727272727e-06,
1454
+ "loss": 0.0,
1455
+ "step": 4875
1456
+ },
1457
+ {
1458
+ "epoch": 8.072487644151565,
1459
+ "grad_norm": 0.0023243261966854334,
1460
+ "learning_rate": 2.0072727272727273e-06,
1461
+ "loss": 0.0,
1462
+ "step": 4900
1463
+ },
1464
+ {
1465
+ "epoch": 8.113673805601318,
1466
+ "grad_norm": 0.0025131104048341513,
1467
+ "learning_rate": 1.961818181818182e-06,
1468
+ "loss": 0.0,
1469
+ "step": 4925
1470
+ },
1471
+ {
1472
+ "epoch": 8.15485996705107,
1473
+ "grad_norm": 0.005834953393787146,
1474
+ "learning_rate": 1.9163636363636365e-06,
1475
+ "loss": 0.0003,
1476
+ "step": 4950
1477
+ },
1478
+ {
1479
+ "epoch": 8.196046128500823,
1480
+ "grad_norm": 0.005762421526014805,
1481
+ "learning_rate": 1.870909090909091e-06,
1482
+ "loss": 0.0001,
1483
+ "step": 4975
1484
+ },
1485
+ {
1486
+ "epoch": 8.237232289950576,
1487
+ "grad_norm": 0.002938208868727088,
1488
+ "learning_rate": 1.8254545454545455e-06,
1489
+ "loss": 0.0,
1490
+ "step": 5000
1491
+ },
1492
+ {
1493
+ "epoch": 8.237232289950576,
1494
+ "eval_loss": 0.013800903223454952,
1495
+ "eval_runtime": 1718.2773,
1496
+ "eval_samples_per_second": 1.412,
1497
+ "eval_steps_per_second": 0.177,
1498
+ "eval_wer": 2.602345995494446,
1499
+ "step": 5000
1500
+ },
1501
+ {
1502
+ "epoch": 8.278418451400329,
1503
+ "grad_norm": 0.0034450343810021877,
1504
+ "learning_rate": 1.7800000000000001e-06,
1505
+ "loss": 0.0001,
1506
+ "step": 5025
1507
+ },
1508
+ {
1509
+ "epoch": 8.319604612850082,
1510
+ "grad_norm": 0.002541518537327647,
1511
+ "learning_rate": 1.7345454545454545e-06,
1512
+ "loss": 0.0,
1513
+ "step": 5050
1514
+ },
1515
+ {
1516
+ "epoch": 8.360790774299835,
1517
+ "grad_norm": 0.002959892386570573,
1518
+ "learning_rate": 1.689090909090909e-06,
1519
+ "loss": 0.0,
1520
+ "step": 5075
1521
+ },
1522
+ {
1523
+ "epoch": 8.401976935749587,
1524
+ "grad_norm": 0.004116767086088657,
1525
+ "learning_rate": 1.6436363636363637e-06,
1526
+ "loss": 0.0001,
1527
+ "step": 5100
1528
+ },
1529
+ {
1530
+ "epoch": 8.44316309719934,
1531
+ "grad_norm": 0.01786639727652073,
1532
+ "learning_rate": 1.5981818181818185e-06,
1533
+ "loss": 0.0,
1534
+ "step": 5125
1535
+ },
1536
+ {
1537
+ "epoch": 8.484349258649093,
1538
+ "grad_norm": 0.008570323698222637,
1539
+ "learning_rate": 1.5527272727272729e-06,
1540
+ "loss": 0.0,
1541
+ "step": 5150
1542
+ },
1543
+ {
1544
+ "epoch": 8.525535420098846,
1545
+ "grad_norm": 0.0028420346789062023,
1546
+ "learning_rate": 1.5072727272727275e-06,
1547
+ "loss": 0.0001,
1548
+ "step": 5175
1549
+ },
1550
+ {
1551
+ "epoch": 8.566721581548599,
1552
+ "grad_norm": 0.002692845184355974,
1553
+ "learning_rate": 1.461818181818182e-06,
1554
+ "loss": 0.0001,
1555
+ "step": 5200
1556
+ },
1557
+ {
1558
+ "epoch": 8.607907742998353,
1559
+ "grad_norm": 0.0014716355362907052,
1560
+ "learning_rate": 1.4163636363636365e-06,
1561
+ "loss": 0.0,
1562
+ "step": 5225
1563
+ },
1564
+ {
1565
+ "epoch": 8.649093904448105,
1566
+ "grad_norm": 0.004192767199128866,
1567
+ "learning_rate": 1.370909090909091e-06,
1568
+ "loss": 0.0,
1569
+ "step": 5250
1570
+ },
1571
+ {
1572
+ "epoch": 8.69028006589786,
1573
+ "grad_norm": 0.0037771447096019983,
1574
+ "learning_rate": 1.3254545454545456e-06,
1575
+ "loss": 0.0,
1576
+ "step": 5275
1577
+ },
1578
+ {
1579
+ "epoch": 8.731466227347612,
1580
+ "grad_norm": 0.0033924980089068413,
1581
+ "learning_rate": 1.28e-06,
1582
+ "loss": 0.0,
1583
+ "step": 5300
1584
+ },
1585
+ {
1586
+ "epoch": 8.772652388797365,
1587
+ "grad_norm": 0.007148527540266514,
1588
+ "learning_rate": 1.2345454545454546e-06,
1589
+ "loss": 0.0,
1590
+ "step": 5325
1591
+ },
1592
+ {
1593
+ "epoch": 8.813838550247118,
1594
+ "grad_norm": 0.0036525556351989508,
1595
+ "learning_rate": 1.1890909090909092e-06,
1596
+ "loss": 0.0001,
1597
+ "step": 5350
1598
+ },
1599
+ {
1600
+ "epoch": 8.85502471169687,
1601
+ "grad_norm": 0.004420670215040445,
1602
+ "learning_rate": 1.1436363636363638e-06,
1603
+ "loss": 0.0,
1604
+ "step": 5375
1605
+ },
1606
+ {
1607
+ "epoch": 8.896210873146623,
1608
+ "grad_norm": 0.007836336269974709,
1609
+ "learning_rate": 1.0981818181818182e-06,
1610
+ "loss": 0.0,
1611
+ "step": 5400
1612
+ },
1613
+ {
1614
+ "epoch": 8.937397034596376,
1615
+ "grad_norm": 0.0029930309392511845,
1616
+ "learning_rate": 1.0527272727272728e-06,
1617
+ "loss": 0.0,
1618
+ "step": 5425
1619
+ },
1620
+ {
1621
+ "epoch": 8.978583196046129,
1622
+ "grad_norm": 0.004859840031713247,
1623
+ "learning_rate": 1.0072727272727274e-06,
1624
+ "loss": 0.0001,
1625
+ "step": 5450
1626
+ },
1627
+ {
1628
+ "epoch": 9.019769357495882,
1629
+ "grad_norm": 0.0022968496195971966,
1630
+ "learning_rate": 9.61818181818182e-07,
1631
+ "loss": 0.0,
1632
+ "step": 5475
1633
+ },
1634
+ {
1635
+ "epoch": 9.060955518945635,
1636
+ "grad_norm": 0.0016871189000084996,
1637
+ "learning_rate": 9.163636363636365e-07,
1638
+ "loss": 0.0001,
1639
+ "step": 5500
1640
+ },
1641
+ {
1642
+ "epoch": 9.060955518945635,
1643
+ "eval_loss": 0.013956129550933838,
1644
+ "eval_runtime": 1736.8144,
1645
+ "eval_samples_per_second": 1.397,
1646
+ "eval_steps_per_second": 0.175,
1647
+ "eval_wer": 2.5091276314767343,
1648
+ "step": 5500
1649
+ },
1650
+ {
1651
+ "epoch": 9.102141680395388,
1652
+ "grad_norm": 0.003995977807790041,
1653
+ "learning_rate": 8.70909090909091e-07,
1654
+ "loss": 0.0001,
1655
+ "step": 5525
1656
+ },
1657
+ {
1658
+ "epoch": 9.14332784184514,
1659
+ "grad_norm": 0.0030109367799013853,
1660
+ "learning_rate": 8.254545454545456e-07,
1661
+ "loss": 0.0,
1662
+ "step": 5550
1663
+ },
1664
+ {
1665
+ "epoch": 9.184514003294893,
1666
+ "grad_norm": 0.004546226933598518,
1667
+ "learning_rate": 7.8e-07,
1668
+ "loss": 0.0,
1669
+ "step": 5575
1670
+ },
1671
+ {
1672
+ "epoch": 9.225700164744646,
1673
+ "grad_norm": 0.003448527306318283,
1674
+ "learning_rate": 7.345454545454545e-07,
1675
+ "loss": 0.0,
1676
+ "step": 5600
1677
+ },
1678
+ {
1679
+ "epoch": 9.266886326194399,
1680
+ "grad_norm": 0.003833542810752988,
1681
+ "learning_rate": 6.890909090909091e-07,
1682
+ "loss": 0.0,
1683
+ "step": 5625
1684
+ },
1685
+ {
1686
+ "epoch": 9.308072487644152,
1687
+ "grad_norm": 0.0035726006608456373,
1688
+ "learning_rate": 6.436363636363636e-07,
1689
+ "loss": 0.0001,
1690
+ "step": 5650
1691
+ },
1692
+ {
1693
+ "epoch": 9.349258649093905,
1694
+ "grad_norm": 0.0038840766064822674,
1695
+ "learning_rate": 5.981818181818182e-07,
1696
+ "loss": 0.0,
1697
+ "step": 5675
1698
+ },
1699
+ {
1700
+ "epoch": 9.390444810543658,
1701
+ "grad_norm": 0.001967709045857191,
1702
+ "learning_rate": 5.527272727272728e-07,
1703
+ "loss": 0.0,
1704
+ "step": 5700
1705
+ },
1706
+ {
1707
+ "epoch": 9.43163097199341,
1708
+ "grad_norm": 0.0022351674269884825,
1709
+ "learning_rate": 5.072727272727273e-07,
1710
+ "loss": 0.0,
1711
+ "step": 5725
1712
+ },
1713
+ {
1714
+ "epoch": 9.472817133443163,
1715
+ "grad_norm": 0.004692086949944496,
1716
+ "learning_rate": 4.618181818181819e-07,
1717
+ "loss": 0.0,
1718
+ "step": 5750
1719
+ },
1720
+ {
1721
+ "epoch": 9.514003294892916,
1722
+ "grad_norm": 0.0022310977801680565,
1723
+ "learning_rate": 4.163636363636364e-07,
1724
+ "loss": 0.0,
1725
+ "step": 5775
1726
+ },
1727
+ {
1728
+ "epoch": 9.555189456342669,
1729
+ "grad_norm": 0.0022611399181187153,
1730
+ "learning_rate": 3.7090909090909093e-07,
1731
+ "loss": 0.0,
1732
+ "step": 5800
1733
+ },
1734
+ {
1735
+ "epoch": 9.596375617792422,
1736
+ "grad_norm": 0.0022751970682293177,
1737
+ "learning_rate": 3.254545454545455e-07,
1738
+ "loss": 0.0,
1739
+ "step": 5825
1740
+ },
1741
+ {
1742
+ "epoch": 9.637561779242175,
1743
+ "grad_norm": 0.003461811225861311,
1744
+ "learning_rate": 2.8e-07,
1745
+ "loss": 0.0,
1746
+ "step": 5850
1747
+ },
1748
+ {
1749
+ "epoch": 9.678747940691927,
1750
+ "grad_norm": 0.002553915837779641,
1751
+ "learning_rate": 2.3454545454545456e-07,
1752
+ "loss": 0.0,
1753
+ "step": 5875
1754
+ },
1755
+ {
1756
+ "epoch": 9.71993410214168,
1757
+ "grad_norm": 0.002379871206358075,
1758
+ "learning_rate": 1.890909090909091e-07,
1759
+ "loss": 0.0,
1760
+ "step": 5900
1761
+ },
1762
+ {
1763
+ "epoch": 9.761120263591433,
1764
+ "grad_norm": 0.0019515318563207984,
1765
+ "learning_rate": 1.4363636363636365e-07,
1766
+ "loss": 0.0,
1767
+ "step": 5925
1768
+ },
1769
+ {
1770
+ "epoch": 9.802306425041186,
1771
+ "grad_norm": 0.0017870984738692641,
1772
+ "learning_rate": 9.818181818181819e-08,
1773
+ "loss": 0.0,
1774
+ "step": 5950
1775
+ },
1776
+ {
1777
+ "epoch": 9.843492586490939,
1778
+ "grad_norm": 0.0032659450080245733,
1779
+ "learning_rate": 5.272727272727273e-08,
1780
+ "loss": 0.0,
1781
+ "step": 5975
1782
+ },
1783
+ {
1784
+ "epoch": 9.884678747940692,
1785
+ "grad_norm": 0.01834402047097683,
1786
+ "learning_rate": 7.272727272727273e-09,
1787
+ "loss": 0.0,
1788
+ "step": 6000
1789
+ },
1790
+ {
1791
+ "epoch": 9.884678747940692,
1792
+ "eval_loss": 0.013986977748572826,
1793
+ "eval_runtime": 1723.1255,
1794
+ "eval_samples_per_second": 1.408,
1795
+ "eval_steps_per_second": 0.176,
1796
+ "eval_wer": 2.5013594344752583,
1797
+ "step": 6000
1798
+ },
1799
+ {
1800
+ "epoch": 9.925864909390445,
1801
+ "grad_norm": 0.0242526326328516,
1802
+ "learning_rate": 3.5047058823529417e-06,
1803
+ "loss": 0.0,
1804
+ "step": 6025
1805
+ },
1806
+ {
1807
+ "epoch": 9.967051070840197,
1808
+ "grad_norm": 0.002037163358181715,
1809
+ "learning_rate": 3.4752941176470588e-06,
1810
+ "loss": 0.0,
1811
+ "step": 6050
1812
+ },
1813
+ {
1814
+ "epoch": 10.00823723228995,
1815
+ "grad_norm": 0.0032065610866993666,
1816
+ "learning_rate": 3.4458823529411767e-06,
1817
+ "loss": 0.0,
1818
+ "step": 6075
1819
+ },
1820
+ {
1821
+ "epoch": 10.049423393739703,
1822
+ "grad_norm": 0.0026907401625066996,
1823
+ "learning_rate": 3.4164705882352946e-06,
1824
+ "loss": 0.0,
1825
+ "step": 6100
1826
+ },
1827
+ {
1828
+ "epoch": 10.090609555189456,
1829
+ "grad_norm": 0.0022176543716341257,
1830
+ "learning_rate": 3.387058823529412e-06,
1831
+ "loss": 0.0,
1832
+ "step": 6125
1833
+ },
1834
+ {
1835
+ "epoch": 10.131795716639209,
1836
+ "grad_norm": 0.0032068961299955845,
1837
+ "learning_rate": 3.35764705882353e-06,
1838
+ "loss": 0.0,
1839
+ "step": 6150
1840
+ },
1841
+ {
1842
+ "epoch": 10.172981878088962,
1843
+ "grad_norm": 0.0039983587339520454,
1844
+ "learning_rate": 3.328235294117647e-06,
1845
+ "loss": 0.0,
1846
+ "step": 6175
1847
+ },
1848
+ {
1849
+ "epoch": 10.214168039538714,
1850
+ "grad_norm": 0.0018802180420607328,
1851
+ "learning_rate": 3.298823529411765e-06,
1852
+ "loss": 0.0,
1853
+ "step": 6200
1854
+ },
1855
+ {
1856
+ "epoch": 10.255354200988467,
1857
+ "grad_norm": 0.0026977595407515764,
1858
+ "learning_rate": 3.2694117647058824e-06,
1859
+ "loss": 0.0,
1860
+ "step": 6225
1861
+ },
1862
+ {
1863
+ "epoch": 10.29654036243822,
1864
+ "grad_norm": 0.0030683802906423807,
1865
+ "learning_rate": 3.2400000000000003e-06,
1866
+ "loss": 0.0,
1867
+ "step": 6250
1868
+ },
1869
+ {
1870
+ "epoch": 10.337726523887973,
1871
+ "grad_norm": 0.002406976418569684,
1872
+ "learning_rate": 3.2105882352941182e-06,
1873
+ "loss": 0.0,
1874
+ "step": 6275
1875
+ },
1876
+ {
1877
+ "epoch": 10.378912685337726,
1878
+ "grad_norm": 0.004232148639857769,
1879
+ "learning_rate": 3.1811764705882353e-06,
1880
+ "loss": 0.0,
1881
+ "step": 6300
1882
+ },
1883
+ {
1884
+ "epoch": 10.420098846787479,
1885
+ "grad_norm": 0.0028111624997109175,
1886
+ "learning_rate": 3.151764705882353e-06,
1887
+ "loss": 0.0,
1888
+ "step": 6325
1889
+ },
1890
+ {
1891
+ "epoch": 10.461285008237232,
1892
+ "grad_norm": 0.0027185629587620497,
1893
+ "learning_rate": 3.1223529411764707e-06,
1894
+ "loss": 0.0,
1895
+ "step": 6350
1896
+ },
1897
+ {
1898
+ "epoch": 10.502471169686984,
1899
+ "grad_norm": 0.002079539466649294,
1900
+ "learning_rate": 3.0929411764705886e-06,
1901
+ "loss": 0.0,
1902
+ "step": 6375
1903
+ },
1904
+ {
1905
+ "epoch": 10.543657331136739,
1906
+ "grad_norm": 0.0021090833470225334,
1907
+ "learning_rate": 3.0635294117647065e-06,
1908
+ "loss": 0.0,
1909
+ "step": 6400
1910
+ },
1911
+ {
1912
+ "epoch": 10.58484349258649,
1913
+ "grad_norm": 0.0029999890830367804,
1914
+ "learning_rate": 3.0341176470588235e-06,
1915
+ "loss": 0.0,
1916
+ "step": 6425
1917
+ },
1918
+ {
1919
+ "epoch": 10.626029654036245,
1920
+ "grad_norm": 0.0023287280928343534,
1921
+ "learning_rate": 3.0047058823529414e-06,
1922
+ "loss": 0.0,
1923
+ "step": 6450
1924
+ },
1925
+ {
1926
+ "epoch": 10.667215815485998,
1927
+ "grad_norm": 0.0014742282219231129,
1928
+ "learning_rate": 2.975294117647059e-06,
1929
+ "loss": 0.0,
1930
+ "step": 6475
1931
+ },
1932
+ {
1933
+ "epoch": 10.70840197693575,
1934
+ "grad_norm": 0.002694872207939625,
1935
+ "learning_rate": 2.945882352941177e-06,
1936
+ "loss": 0.0001,
1937
+ "step": 6500
1938
+ },
1939
+ {
1940
+ "epoch": 10.70840197693575,
1941
+ "eval_loss": 0.014553220942616463,
1942
+ "eval_runtime": 1723.0905,
1943
+ "eval_samples_per_second": 1.409,
1944
+ "eval_steps_per_second": 0.176,
1945
+ "eval_wer": 2.5945777984929697,
1946
+ "step": 6500
1947
+ },
1948
+ {
1949
+ "epoch": 10.749588138385503,
1950
+ "grad_norm": 0.001027889782562852,
1951
+ "learning_rate": 2.9164705882352943e-06,
1952
+ "loss": 0.0,
1953
+ "step": 6525
1954
+ },
1955
+ {
1956
+ "epoch": 10.790774299835256,
1957
+ "grad_norm": 0.0017091067275032401,
1958
+ "learning_rate": 2.887058823529412e-06,
1959
+ "loss": 0.0,
1960
+ "step": 6550
1961
+ },
1962
+ {
1963
+ "epoch": 10.831960461285009,
1964
+ "grad_norm": 0.001180932391434908,
1965
+ "learning_rate": 2.8576470588235293e-06,
1966
+ "loss": 0.0,
1967
+ "step": 6575
1968
+ },
1969
+ {
1970
+ "epoch": 10.873146622734762,
1971
+ "grad_norm": 0.0035964970011264086,
1972
+ "learning_rate": 2.828235294117647e-06,
1973
+ "loss": 0.0005,
1974
+ "step": 6600
1975
+ },
1976
+ {
1977
+ "epoch": 10.914332784184515,
1978
+ "grad_norm": 0.0023179524578154087,
1979
+ "learning_rate": 2.798823529411765e-06,
1980
+ "loss": 0.0,
1981
+ "step": 6625
1982
+ },
1983
+ {
1984
+ "epoch": 10.955518945634267,
1985
+ "grad_norm": 0.10292612761259079,
1986
+ "learning_rate": 2.7694117647058825e-06,
1987
+ "loss": 0.0001,
1988
+ "step": 6650
1989
+ },
1990
+ {
1991
+ "epoch": 10.99670510708402,
1992
+ "grad_norm": 1.5822137594223022,
1993
+ "learning_rate": 2.7400000000000004e-06,
1994
+ "loss": 0.0007,
1995
+ "step": 6675
1996
+ },
1997
+ {
1998
+ "epoch": 11.037891268533773,
1999
+ "grad_norm": 0.09981144219636917,
2000
+ "learning_rate": 2.7105882352941175e-06,
2001
+ "loss": 0.0028,
2002
+ "step": 6700
2003
+ },
2004
+ {
2005
+ "epoch": 11.079077429983526,
2006
+ "grad_norm": 1.2205625772476196,
2007
+ "learning_rate": 2.6811764705882354e-06,
2008
+ "loss": 0.0018,
2009
+ "step": 6725
2010
+ },
2011
+ {
2012
+ "epoch": 11.120263591433279,
2013
+ "grad_norm": 2.4946441650390625,
2014
+ "learning_rate": 2.6517647058823533e-06,
2015
+ "loss": 0.0037,
2016
+ "step": 6750
2017
+ },
2018
+ {
2019
+ "epoch": 11.161449752883032,
2020
+ "grad_norm": 0.01300705224275589,
2021
+ "learning_rate": 2.622352941176471e-06,
2022
+ "loss": 0.0007,
2023
+ "step": 6775
2024
+ },
2025
+ {
2026
+ "epoch": 11.202635914332784,
2027
+ "grad_norm": 0.050090231001377106,
2028
+ "learning_rate": 2.5929411764705887e-06,
2029
+ "loss": 0.0004,
2030
+ "step": 6800
2031
+ },
2032
+ {
2033
+ "epoch": 11.243822075782537,
2034
+ "grad_norm": 0.1064329594373703,
2035
+ "learning_rate": 2.5635294117647058e-06,
2036
+ "loss": 0.0003,
2037
+ "step": 6825
2038
+ },
2039
+ {
2040
+ "epoch": 11.28500823723229,
2041
+ "grad_norm": 0.00640516122803092,
2042
+ "learning_rate": 2.5341176470588237e-06,
2043
+ "loss": 0.0002,
2044
+ "step": 6850
2045
+ },
2046
+ {
2047
+ "epoch": 11.326194398682043,
2048
+ "grad_norm": 0.011775782331824303,
2049
+ "learning_rate": 2.504705882352941e-06,
2050
+ "loss": 0.0001,
2051
+ "step": 6875
2052
+ },
2053
+ {
2054
+ "epoch": 11.367380560131796,
2055
+ "grad_norm": 0.00485497061163187,
2056
+ "learning_rate": 2.475294117647059e-06,
2057
+ "loss": 0.0011,
2058
+ "step": 6900
2059
+ },
2060
+ {
2061
+ "epoch": 11.408566721581549,
2062
+ "grad_norm": 1.4145103693008423,
2063
+ "learning_rate": 2.4458823529411765e-06,
2064
+ "loss": 0.0005,
2065
+ "step": 6925
2066
+ },
2067
+ {
2068
+ "epoch": 11.449752883031302,
2069
+ "grad_norm": 0.01526624895632267,
2070
+ "learning_rate": 2.4164705882352944e-06,
2071
+ "loss": 0.0002,
2072
+ "step": 6950
2073
+ },
2074
+ {
2075
+ "epoch": 11.490939044481054,
2076
+ "grad_norm": 0.02947794646024704,
2077
+ "learning_rate": 2.387058823529412e-06,
2078
+ "loss": 0.0001,
2079
+ "step": 6975
2080
+ },
2081
+ {
2082
+ "epoch": 11.532125205930807,
2083
+ "grad_norm": 0.008897418156266212,
2084
+ "learning_rate": 2.3576470588235294e-06,
2085
+ "loss": 0.0002,
2086
+ "step": 7000
2087
+ },
2088
+ {
2089
+ "epoch": 11.532125205930807,
2090
+ "eval_loss": 0.016228729858994484,
2091
+ "eval_runtime": 1688.3118,
2092
+ "eval_samples_per_second": 1.438,
2093
+ "eval_steps_per_second": 0.18,
2094
+ "eval_wer": 3.029596830575623,
2095
+ "step": 7000
2096
+ },
2097
+ {
2098
+ "epoch": 11.57331136738056,
2099
+ "grad_norm": 0.1610272079706192,
2100
+ "learning_rate": 2.3282352941176473e-06,
2101
+ "loss": 0.0004,
2102
+ "step": 7025
2103
+ },
2104
+ {
2105
+ "epoch": 11.614497528830313,
2106
+ "grad_norm": 1.4256123304367065,
2107
+ "learning_rate": 2.298823529411765e-06,
2108
+ "loss": 0.0002,
2109
+ "step": 7050
2110
+ },
2111
+ {
2112
+ "epoch": 11.655683690280066,
2113
+ "grad_norm": 0.004848022013902664,
2114
+ "learning_rate": 2.2694117647058827e-06,
2115
+ "loss": 0.0003,
2116
+ "step": 7075
2117
+ },
2118
+ {
2119
+ "epoch": 11.696869851729819,
2120
+ "grad_norm": 0.007711484562605619,
2121
+ "learning_rate": 2.24e-06,
2122
+ "loss": 0.0006,
2123
+ "step": 7100
2124
+ },
2125
+ {
2126
+ "epoch": 11.738056013179571,
2127
+ "grad_norm": 0.11259625852108002,
2128
+ "learning_rate": 2.2105882352941176e-06,
2129
+ "loss": 0.0002,
2130
+ "step": 7125
2131
+ },
2132
+ {
2133
+ "epoch": 11.779242174629324,
2134
+ "grad_norm": 0.004204327240586281,
2135
+ "learning_rate": 2.1811764705882355e-06,
2136
+ "loss": 0.0001,
2137
+ "step": 7150
2138
+ },
2139
+ {
2140
+ "epoch": 11.820428336079077,
2141
+ "grad_norm": 0.007289988454431295,
2142
+ "learning_rate": 2.151764705882353e-06,
2143
+ "loss": 0.0001,
2144
+ "step": 7175
2145
+ },
2146
+ {
2147
+ "epoch": 11.86161449752883,
2148
+ "grad_norm": 0.02570285275578499,
2149
+ "learning_rate": 2.1223529411764705e-06,
2150
+ "loss": 0.0003,
2151
+ "step": 7200
2152
+ },
2153
+ {
2154
+ "epoch": 11.902800658978583,
2155
+ "grad_norm": 0.009201893582940102,
2156
+ "learning_rate": 2.0929411764705884e-06,
2157
+ "loss": 0.0002,
2158
+ "step": 7225
2159
+ },
2160
+ {
2161
+ "epoch": 11.943986820428336,
2162
+ "grad_norm": 0.003304847748950124,
2163
+ "learning_rate": 2.0635294117647063e-06,
2164
+ "loss": 0.0002,
2165
+ "step": 7250
2166
+ },
2167
+ {
2168
+ "epoch": 11.985172981878089,
2169
+ "grad_norm": 0.07221189141273499,
2170
+ "learning_rate": 2.0341176470588238e-06,
2171
+ "loss": 0.0001,
2172
+ "step": 7275
2173
+ },
2174
+ {
2175
+ "epoch": 12.026359143327841,
2176
+ "grad_norm": 0.01542985625565052,
2177
+ "learning_rate": 2.0058823529411765e-06,
2178
+ "loss": 0.0002,
2179
+ "step": 7300
2180
+ },
2181
+ {
2182
+ "epoch": 12.067545304777594,
2183
+ "grad_norm": 0.009480349719524384,
2184
+ "learning_rate": 1.976470588235294e-06,
2185
+ "loss": 0.0,
2186
+ "step": 7325
2187
+ },
2188
+ {
2189
+ "epoch": 12.108731466227347,
2190
+ "grad_norm": 0.0024922864977270365,
2191
+ "learning_rate": 1.947058823529412e-06,
2192
+ "loss": 0.0002,
2193
+ "step": 7350
2194
+ },
2195
+ {
2196
+ "epoch": 12.1499176276771,
2197
+ "grad_norm": 0.0071571036241948605,
2198
+ "learning_rate": 1.9176470588235298e-06,
2199
+ "loss": 0.0001,
2200
+ "step": 7375
2201
+ },
2202
+ {
2203
+ "epoch": 12.191103789126853,
2204
+ "grad_norm": 0.008463573642075062,
2205
+ "learning_rate": 1.888235294117647e-06,
2206
+ "loss": 0.0001,
2207
+ "step": 7400
2208
+ },
2209
+ {
2210
+ "epoch": 12.232289950576606,
2211
+ "grad_norm": 0.03403167426586151,
2212
+ "learning_rate": 1.858823529411765e-06,
2213
+ "loss": 0.0001,
2214
+ "step": 7425
2215
+ },
2216
+ {
2217
+ "epoch": 12.273476112026358,
2218
+ "grad_norm": 0.0034997081384062767,
2219
+ "learning_rate": 1.8294117647058827e-06,
2220
+ "loss": 0.0,
2221
+ "step": 7450
2222
+ },
2223
+ {
2224
+ "epoch": 12.314662273476111,
2225
+ "grad_norm": 0.0037921613547950983,
2226
+ "learning_rate": 1.8011764705882354e-06,
2227
+ "loss": 0.0004,
2228
+ "step": 7475
2229
+ },
2230
+ {
2231
+ "epoch": 12.355848434925864,
2232
+ "grad_norm": 0.0021253039594739676,
2233
+ "learning_rate": 1.771764705882353e-06,
2234
+ "loss": 0.0001,
2235
+ "step": 7500
2236
+ },
2237
+ {
2238
+ "epoch": 12.355848434925864,
2239
+ "eval_loss": 0.012992938980460167,
2240
+ "eval_runtime": 1716.627,
2241
+ "eval_samples_per_second": 1.414,
2242
+ "eval_steps_per_second": 0.177,
2243
+ "eval_wer": 2.3382272974442633,
2244
+ "step": 7500
2245
+ },
2246
+ {
2247
+ "epoch": 12.397034596375617,
2248
+ "grad_norm": 0.023194992914795876,
2249
+ "learning_rate": 1.742352941176471e-06,
2250
+ "loss": 0.0,
2251
+ "step": 7525
2252
+ },
2253
+ {
2254
+ "epoch": 12.43822075782537,
2255
+ "grad_norm": 0.15046222507953644,
2256
+ "learning_rate": 1.7129411764705882e-06,
2257
+ "loss": 0.0001,
2258
+ "step": 7550
2259
+ },
2260
+ {
2261
+ "epoch": 12.479406919275124,
2262
+ "grad_norm": 0.003174582729116082,
2263
+ "learning_rate": 1.6835294117647061e-06,
2264
+ "loss": 0.0001,
2265
+ "step": 7575
2266
+ },
2267
+ {
2268
+ "epoch": 12.520593080724876,
2269
+ "grad_norm": 0.0067563788034021854,
2270
+ "learning_rate": 1.6541176470588236e-06,
2271
+ "loss": 0.0,
2272
+ "step": 7600
2273
+ },
2274
+ {
2275
+ "epoch": 12.56177924217463,
2276
+ "grad_norm": 0.002005406189709902,
2277
+ "learning_rate": 1.6247058823529413e-06,
2278
+ "loss": 0.0,
2279
+ "step": 7625
2280
+ },
2281
+ {
2282
+ "epoch": 12.602965403624383,
2283
+ "grad_norm": 0.002191119594499469,
2284
+ "learning_rate": 1.595294117647059e-06,
2285
+ "loss": 0.0001,
2286
+ "step": 7650
2287
+ },
2288
+ {
2289
+ "epoch": 12.644151565074136,
2290
+ "grad_norm": 0.0046321190893650055,
2291
+ "learning_rate": 1.5658823529411765e-06,
2292
+ "loss": 0.0001,
2293
+ "step": 7675
2294
+ },
2295
+ {
2296
+ "epoch": 12.685337726523889,
2297
+ "grad_norm": 0.006401429884135723,
2298
+ "learning_rate": 1.5364705882352942e-06,
2299
+ "loss": 0.0001,
2300
+ "step": 7700
2301
+ },
2302
+ {
2303
+ "epoch": 12.726523887973642,
2304
+ "grad_norm": 0.0027000552508980036,
2305
+ "learning_rate": 1.507058823529412e-06,
2306
+ "loss": 0.0002,
2307
+ "step": 7725
2308
+ },
2309
+ {
2310
+ "epoch": 12.767710049423394,
2311
+ "grad_norm": 0.19640517234802246,
2312
+ "learning_rate": 1.4776470588235296e-06,
2313
+ "loss": 0.0001,
2314
+ "step": 7750
2315
+ },
2316
+ {
2317
+ "epoch": 12.808896210873147,
2318
+ "grad_norm": 0.0019183914409950376,
2319
+ "learning_rate": 1.4482352941176472e-06,
2320
+ "loss": 0.0001,
2321
+ "step": 7775
2322
+ },
2323
+ {
2324
+ "epoch": 12.8500823723229,
2325
+ "grad_norm": 0.012867034412920475,
2326
+ "learning_rate": 1.4188235294117647e-06,
2327
+ "loss": 0.0001,
2328
+ "step": 7800
2329
+ },
2330
+ {
2331
+ "epoch": 12.891268533772653,
2332
+ "grad_norm": 0.0031524612568318844,
2333
+ "learning_rate": 1.3894117647058824e-06,
2334
+ "loss": 0.0001,
2335
+ "step": 7825
2336
+ },
2337
+ {
2338
+ "epoch": 12.932454695222406,
2339
+ "grad_norm": 0.004168114624917507,
2340
+ "learning_rate": 1.3600000000000001e-06,
2341
+ "loss": 0.0,
2342
+ "step": 7850
2343
+ },
2344
+ {
2345
+ "epoch": 12.973640856672159,
2346
+ "grad_norm": 0.007319289725273848,
2347
+ "learning_rate": 1.3305882352941176e-06,
2348
+ "loss": 0.0001,
2349
+ "step": 7875
2350
+ },
2351
+ {
2352
+ "epoch": 13.014827018121911,
2353
+ "grad_norm": 0.007646168116480112,
2354
+ "learning_rate": 1.3011764705882355e-06,
2355
+ "loss": 0.0,
2356
+ "step": 7900
2357
+ },
2358
+ {
2359
+ "epoch": 13.056013179571664,
2360
+ "grad_norm": 0.0034050345420837402,
2361
+ "learning_rate": 1.2717647058823532e-06,
2362
+ "loss": 0.0,
2363
+ "step": 7925
2364
+ },
2365
+ {
2366
+ "epoch": 13.097199341021417,
2367
+ "grad_norm": 0.0018609551480039954,
2368
+ "learning_rate": 1.2423529411764707e-06,
2369
+ "loss": 0.0,
2370
+ "step": 7950
2371
+ },
2372
+ {
2373
+ "epoch": 13.13838550247117,
2374
+ "grad_norm": 0.005138488952070475,
2375
+ "learning_rate": 1.2129411764705884e-06,
2376
+ "loss": 0.0,
2377
+ "step": 7975
2378
+ },
2379
+ {
2380
+ "epoch": 13.179571663920923,
2381
+ "grad_norm": 0.003146812552586198,
2382
+ "learning_rate": 1.183529411764706e-06,
2383
+ "loss": 0.0,
2384
+ "step": 8000
2385
+ },
2386
+ {
2387
+ "epoch": 13.179571663920923,
2388
+ "eval_loss": 0.013023543171584606,
2389
+ "eval_runtime": 1630.2411,
2390
+ "eval_samples_per_second": 1.489,
2391
+ "eval_steps_per_second": 0.186,
2392
+ "eval_wer": 2.4780548434708303,
2393
+ "step": 8000
2394
+ },
2395
+ {
2396
+ "epoch": 13.220757825370676,
2397
+ "grad_norm": 0.0029420964419841766,
2398
+ "learning_rate": 1.1541176470588235e-06,
2399
+ "loss": 0.0,
2400
+ "step": 8025
2401
+ },
2402
+ {
2403
+ "epoch": 13.261943986820429,
2404
+ "grad_norm": 0.006158372852951288,
2405
+ "learning_rate": 1.1247058823529412e-06,
2406
+ "loss": 0.0001,
2407
+ "step": 8050
2408
+ },
2409
+ {
2410
+ "epoch": 13.303130148270181,
2411
+ "grad_norm": 0.00477724289521575,
2412
+ "learning_rate": 1.095294117647059e-06,
2413
+ "loss": 0.0,
2414
+ "step": 8075
2415
+ },
2416
+ {
2417
+ "epoch": 13.344316309719934,
2418
+ "grad_norm": 0.012528087943792343,
2419
+ "learning_rate": 1.0658823529411766e-06,
2420
+ "loss": 0.0,
2421
+ "step": 8100
2422
+ },
2423
+ {
2424
+ "epoch": 13.385502471169687,
2425
+ "grad_norm": 0.0019438378512859344,
2426
+ "learning_rate": 1.0364705882352943e-06,
2427
+ "loss": 0.0001,
2428
+ "step": 8125
2429
+ },
2430
+ {
2431
+ "epoch": 13.42668863261944,
2432
+ "grad_norm": 0.003603472374379635,
2433
+ "learning_rate": 1.0070588235294118e-06,
2434
+ "loss": 0.0,
2435
+ "step": 8150
2436
+ },
2437
+ {
2438
+ "epoch": 13.467874794069193,
2439
+ "grad_norm": 0.00542394258081913,
2440
+ "learning_rate": 9.776470588235295e-07,
2441
+ "loss": 0.0,
2442
+ "step": 8175
2443
+ },
2444
+ {
2445
+ "epoch": 13.509060955518946,
2446
+ "grad_norm": 0.001990295248106122,
2447
+ "learning_rate": 9.482352941176472e-07,
2448
+ "loss": 0.0,
2449
+ "step": 8200
2450
+ },
2451
+ {
2452
+ "epoch": 13.550247116968698,
2453
+ "grad_norm": 0.009004420600831509,
2454
+ "learning_rate": 9.188235294117647e-07,
2455
+ "loss": 0.0,
2456
+ "step": 8225
2457
+ },
2458
+ {
2459
+ "epoch": 13.591433278418451,
2460
+ "grad_norm": 0.0023559145629405975,
2461
+ "learning_rate": 8.894117647058824e-07,
2462
+ "loss": 0.0,
2463
+ "step": 8250
2464
+ },
2465
+ {
2466
+ "epoch": 13.632619439868204,
2467
+ "grad_norm": 0.006258898414671421,
2468
+ "learning_rate": 8.6e-07,
2469
+ "loss": 0.0,
2470
+ "step": 8275
2471
+ },
2472
+ {
2473
+ "epoch": 13.673805601317957,
2474
+ "grad_norm": 0.0016879294998943806,
2475
+ "learning_rate": 8.305882352941177e-07,
2476
+ "loss": 0.0,
2477
+ "step": 8300
2478
+ },
2479
+ {
2480
+ "epoch": 13.71499176276771,
2481
+ "grad_norm": 0.006166738457977772,
2482
+ "learning_rate": 8.011764705882354e-07,
2483
+ "loss": 0.0,
2484
+ "step": 8325
2485
+ },
2486
+ {
2487
+ "epoch": 13.756177924217463,
2488
+ "grad_norm": 0.002089226385578513,
2489
+ "learning_rate": 7.71764705882353e-07,
2490
+ "loss": 0.0,
2491
+ "step": 8350
2492
+ },
2493
+ {
2494
+ "epoch": 13.797364085667216,
2495
+ "grad_norm": 0.006312618497759104,
2496
+ "learning_rate": 7.423529411764706e-07,
2497
+ "loss": 0.0,
2498
+ "step": 8375
2499
+ },
2500
+ {
2501
+ "epoch": 13.838550247116968,
2502
+ "grad_norm": 0.002466730307787657,
2503
+ "learning_rate": 7.129411764705884e-07,
2504
+ "loss": 0.0,
2505
+ "step": 8400
2506
+ },
2507
+ {
2508
+ "epoch": 13.879736408566721,
2509
+ "grad_norm": 0.0038565557915717363,
2510
+ "learning_rate": 6.83529411764706e-07,
2511
+ "loss": 0.0,
2512
+ "step": 8425
2513
+ },
2514
+ {
2515
+ "epoch": 13.920922570016474,
2516
+ "grad_norm": 0.0016545254038646817,
2517
+ "learning_rate": 6.541176470588236e-07,
2518
+ "loss": 0.0,
2519
+ "step": 8450
2520
+ },
2521
+ {
2522
+ "epoch": 13.962108731466227,
2523
+ "grad_norm": 0.005063951946794987,
2524
+ "learning_rate": 6.247058823529412e-07,
2525
+ "loss": 0.0,
2526
+ "step": 8475
2527
+ },
2528
+ {
2529
+ "epoch": 14.00329489291598,
2530
+ "grad_norm": 0.006067285314202309,
2531
+ "learning_rate": 5.952941176470588e-07,
2532
+ "loss": 0.0,
2533
+ "step": 8500
2534
+ },
2535
+ {
2536
+ "epoch": 14.00329489291598,
2537
+ "eval_loss": 0.013567977584898472,
2538
+ "eval_runtime": 1612.6489,
2539
+ "eval_samples_per_second": 1.505,
2540
+ "eval_steps_per_second": 0.189,
2541
+ "eval_wer": 2.4469820554649266,
2542
+ "step": 8500
2543
+ },
2544
+ {
2545
+ "epoch": 14.044481054365733,
2546
+ "grad_norm": 0.0043179988861083984,
2547
+ "learning_rate": 5.658823529411765e-07,
2548
+ "loss": 0.0,
2549
+ "step": 8525
2550
+ },
2551
+ {
2552
+ "epoch": 14.085667215815485,
2553
+ "grad_norm": 0.0059325359761714935,
2554
+ "learning_rate": 5.364705882352941e-07,
2555
+ "loss": 0.0,
2556
+ "step": 8550
2557
+ },
2558
+ {
2559
+ "epoch": 14.126853377265238,
2560
+ "grad_norm": 0.002210423117503524,
2561
+ "learning_rate": 5.070588235294118e-07,
2562
+ "loss": 0.0,
2563
+ "step": 8575
2564
+ },
2565
+ {
2566
+ "epoch": 14.168039538714991,
2567
+ "grad_norm": 0.0023594272788614035,
2568
+ "learning_rate": 4.776470588235294e-07,
2569
+ "loss": 0.0,
2570
+ "step": 8600
2571
+ },
2572
+ {
2573
+ "epoch": 14.209225700164744,
2574
+ "grad_norm": 0.001723596709780395,
2575
+ "learning_rate": 4.482352941176471e-07,
2576
+ "loss": 0.0,
2577
+ "step": 8625
2578
+ },
2579
+ {
2580
+ "epoch": 14.250411861614497,
2581
+ "grad_norm": 0.0022638991940766573,
2582
+ "learning_rate": 4.188235294117647e-07,
2583
+ "loss": 0.0,
2584
+ "step": 8650
2585
+ },
2586
+ {
2587
+ "epoch": 14.29159802306425,
2588
+ "grad_norm": 0.0020570564083755016,
2589
+ "learning_rate": 3.894117647058824e-07,
2590
+ "loss": 0.0,
2591
+ "step": 8675
2592
+ },
2593
+ {
2594
+ "epoch": 14.332784184514002,
2595
+ "grad_norm": 0.017068659886717796,
2596
+ "learning_rate": 3.6e-07,
2597
+ "loss": 0.0,
2598
+ "step": 8700
2599
+ },
2600
+ {
2601
+ "epoch": 14.373970345963755,
2602
+ "grad_norm": 0.006104827858507633,
2603
+ "learning_rate": 3.305882352941177e-07,
2604
+ "loss": 0.0,
2605
+ "step": 8725
2606
+ },
2607
+ {
2608
+ "epoch": 14.41515650741351,
2609
+ "grad_norm": 0.004056777339428663,
2610
+ "learning_rate": 3.011764705882353e-07,
2611
+ "loss": 0.0,
2612
+ "step": 8750
2613
+ },
2614
+ {
2615
+ "epoch": 14.456342668863263,
2616
+ "grad_norm": 0.002557105151936412,
2617
+ "learning_rate": 2.7176470588235296e-07,
2618
+ "loss": 0.0,
2619
+ "step": 8775
2620
+ },
2621
+ {
2622
+ "epoch": 14.497528830313016,
2623
+ "grad_norm": 0.0033158042933791876,
2624
+ "learning_rate": 2.423529411764706e-07,
2625
+ "loss": 0.0,
2626
+ "step": 8800
2627
+ },
2628
+ {
2629
+ "epoch": 14.538714991762768,
2630
+ "grad_norm": 0.0037535373121500015,
2631
+ "learning_rate": 2.1294117647058827e-07,
2632
+ "loss": 0.0,
2633
+ "step": 8825
2634
+ },
2635
+ {
2636
+ "epoch": 14.579901153212521,
2637
+ "grad_norm": 0.0012597900349646807,
2638
+ "learning_rate": 1.835294117647059e-07,
2639
+ "loss": 0.0,
2640
+ "step": 8850
2641
+ },
2642
+ {
2643
+ "epoch": 14.621087314662274,
2644
+ "grad_norm": 0.006313154939562082,
2645
+ "learning_rate": 1.5411764705882355e-07,
2646
+ "loss": 0.0,
2647
+ "step": 8875
2648
+ },
2649
+ {
2650
+ "epoch": 14.662273476112027,
2651
+ "grad_norm": 0.0067338962107896805,
2652
+ "learning_rate": 1.2470588235294119e-07,
2653
+ "loss": 0.0,
2654
+ "step": 8900
2655
+ },
2656
+ {
2657
+ "epoch": 14.70345963756178,
2658
+ "grad_norm": 0.001957102445885539,
2659
+ "learning_rate": 9.529411764705883e-08,
2660
+ "loss": 0.0,
2661
+ "step": 8925
2662
+ },
2663
+ {
2664
+ "epoch": 14.744645799011533,
2665
+ "grad_norm": 0.0035015118774026632,
2666
+ "learning_rate": 6.588235294117648e-08,
2667
+ "loss": 0.0,
2668
+ "step": 8950
2669
+ },
2670
+ {
2671
+ "epoch": 14.785831960461286,
2672
+ "grad_norm": 0.0019214386120438576,
2673
+ "learning_rate": 3.6470588235294116e-08,
2674
+ "loss": 0.0,
2675
+ "step": 8975
2676
+ },
2677
+ {
2678
+ "epoch": 14.827018121911038,
2679
+ "grad_norm": 0.005048139486461878,
2680
+ "learning_rate": 7.0588235294117654e-09,
2681
+ "loss": 0.0,
2682
+ "step": 9000
2683
+ },
2684
+ {
2685
+ "epoch": 14.827018121911038,
2686
+ "eval_loss": 0.013581627048552036,
2687
+ "eval_runtime": 1622.8495,
2688
+ "eval_samples_per_second": 1.496,
2689
+ "eval_steps_per_second": 0.187,
2690
+ "eval_wer": 2.4858230404723063,
2691
+ "step": 9000
2692
+ },
2693
+ {
2694
+ "epoch": 14.868204283360791,
2695
+ "grad_norm": 0.002532125450670719,
2696
+ "learning_rate": 2.5921739130434787e-06,
2697
+ "loss": 0.0,
2698
+ "step": 9025
2699
+ },
2700
+ {
2701
+ "epoch": 14.909390444810544,
2702
+ "grad_norm": 0.0009551341645419598,
2703
+ "learning_rate": 2.5704347826086954e-06,
2704
+ "loss": 0.0,
2705
+ "step": 9050
2706
+ },
2707
+ {
2708
+ "epoch": 14.950576606260297,
2709
+ "grad_norm": 0.0017027213471010327,
2710
+ "learning_rate": 2.548695652173913e-06,
2711
+ "loss": 0.0,
2712
+ "step": 9075
2713
+ },
2714
+ {
2715
+ "epoch": 14.99176276771005,
2716
+ "grad_norm": 0.0021833304781466722,
2717
+ "learning_rate": 2.5269565217391305e-06,
2718
+ "loss": 0.0,
2719
+ "step": 9100
2720
+ },
2721
+ {
2722
+ "epoch": 15.032948929159803,
2723
+ "grad_norm": 0.0017440480878576636,
2724
+ "learning_rate": 2.505217391304348e-06,
2725
+ "loss": 0.0,
2726
+ "step": 9125
2727
+ },
2728
+ {
2729
+ "epoch": 15.074135090609555,
2730
+ "grad_norm": 0.005614432040601969,
2731
+ "learning_rate": 2.4834782608695652e-06,
2732
+ "loss": 0.0,
2733
+ "step": 9150
2734
+ },
2735
+ {
2736
+ "epoch": 15.115321252059308,
2737
+ "grad_norm": 0.0023146765306591988,
2738
+ "learning_rate": 2.461739130434783e-06,
2739
+ "loss": 0.0,
2740
+ "step": 9175
2741
+ },
2742
+ {
2743
+ "epoch": 15.156507413509061,
2744
+ "grad_norm": 0.00141927320510149,
2745
+ "learning_rate": 2.4400000000000004e-06,
2746
+ "loss": 0.0,
2747
+ "step": 9200
2748
+ },
2749
+ {
2750
+ "epoch": 15.197693574958814,
2751
+ "grad_norm": 0.0027971493545919657,
2752
+ "learning_rate": 2.4182608695652175e-06,
2753
+ "loss": 0.0,
2754
+ "step": 9225
2755
+ },
2756
+ {
2757
+ "epoch": 15.238879736408567,
2758
+ "grad_norm": 0.0009957117727026343,
2759
+ "learning_rate": 2.396521739130435e-06,
2760
+ "loss": 0.0,
2761
+ "step": 9250
2762
+ },
2763
+ {
2764
+ "epoch": 15.28006589785832,
2765
+ "grad_norm": 0.0014513045316562057,
2766
+ "learning_rate": 2.3747826086956526e-06,
2767
+ "loss": 0.0,
2768
+ "step": 9275
2769
+ },
2770
+ {
2771
+ "epoch": 15.321252059308073,
2772
+ "grad_norm": 0.0019499498885124922,
2773
+ "learning_rate": 2.3530434782608698e-06,
2774
+ "loss": 0.0,
2775
+ "step": 9300
2776
+ },
2777
+ {
2778
+ "epoch": 15.362438220757825,
2779
+ "grad_norm": 0.0013381906319409609,
2780
+ "learning_rate": 2.3313043478260873e-06,
2781
+ "loss": 0.0,
2782
+ "step": 9325
2783
+ },
2784
+ {
2785
+ "epoch": 15.403624382207578,
2786
+ "grad_norm": 0.0021802387200295925,
2787
+ "learning_rate": 2.3095652173913045e-06,
2788
+ "loss": 0.0,
2789
+ "step": 9350
2790
+ },
2791
+ {
2792
+ "epoch": 15.444810543657331,
2793
+ "grad_norm": 0.001607355079613626,
2794
+ "learning_rate": 2.287826086956522e-06,
2795
+ "loss": 0.0,
2796
+ "step": 9375
2797
+ },
2798
+ {
2799
+ "epoch": 15.485996705107084,
2800
+ "grad_norm": 0.0017373546725139022,
2801
+ "learning_rate": 2.2660869565217396e-06,
2802
+ "loss": 0.0,
2803
+ "step": 9400
2804
+ },
2805
+ {
2806
+ "epoch": 15.527182866556837,
2807
+ "grad_norm": 0.0009414468659088016,
2808
+ "learning_rate": 2.2443478260869567e-06,
2809
+ "loss": 0.0,
2810
+ "step": 9425
2811
+ },
2812
+ {
2813
+ "epoch": 15.56836902800659,
2814
+ "grad_norm": 0.0011722907656803727,
2815
+ "learning_rate": 2.2226086956521743e-06,
2816
+ "loss": 0.0,
2817
+ "step": 9450
2818
+ },
2819
+ {
2820
+ "epoch": 15.609555189456342,
2821
+ "grad_norm": 0.0012756260111927986,
2822
+ "learning_rate": 2.2008695652173914e-06,
2823
+ "loss": 0.0,
2824
+ "step": 9475
2825
+ },
2826
+ {
2827
+ "epoch": 15.650741350906095,
2828
+ "grad_norm": 0.003932903986424208,
2829
+ "learning_rate": 2.179130434782609e-06,
2830
+ "loss": 0.0,
2831
+ "step": 9500
2832
+ },
2833
+ {
2834
+ "epoch": 15.650741350906095,
2835
+ "eval_loss": 0.013979827053844929,
2836
+ "eval_runtime": 1641.4283,
2837
+ "eval_samples_per_second": 1.479,
2838
+ "eval_steps_per_second": 0.185,
2839
+ "eval_wer": 2.532432222481162,
2840
+ "step": 9500
2841
+ },
2842
+ {
2843
+ "epoch": 15.691927512355848,
2844
+ "grad_norm": 0.0014398075873032212,
2845
+ "learning_rate": 2.157391304347826e-06,
2846
+ "loss": 0.0,
2847
+ "step": 9525
2848
+ },
2849
+ {
2850
+ "epoch": 15.733113673805601,
2851
+ "grad_norm": 0.0019754846580326557,
2852
+ "learning_rate": 2.1356521739130437e-06,
2853
+ "loss": 0.0,
2854
+ "step": 9550
2855
+ },
2856
+ {
2857
+ "epoch": 15.774299835255354,
2858
+ "grad_norm": 0.001919199014082551,
2859
+ "learning_rate": 2.113913043478261e-06,
2860
+ "loss": 0.0,
2861
+ "step": 9575
2862
+ },
2863
+ {
2864
+ "epoch": 15.815485996705107,
2865
+ "grad_norm": 0.005105574149638414,
2866
+ "learning_rate": 2.0921739130434784e-06,
2867
+ "loss": 0.0,
2868
+ "step": 9600
2869
+ },
2870
+ {
2871
+ "epoch": 15.85667215815486,
2872
+ "grad_norm": 0.0087418919429183,
2873
+ "learning_rate": 2.070434782608696e-06,
2874
+ "loss": 0.0,
2875
+ "step": 9625
2876
+ },
2877
+ {
2878
+ "epoch": 15.897858319604612,
2879
+ "grad_norm": 0.0017288941890001297,
2880
+ "learning_rate": 2.048695652173913e-06,
2881
+ "loss": 0.0,
2882
+ "step": 9650
2883
+ },
2884
+ {
2885
+ "epoch": 15.939044481054365,
2886
+ "grad_norm": 0.0016419363673776388,
2887
+ "learning_rate": 2.0269565217391307e-06,
2888
+ "loss": 0.0,
2889
+ "step": 9675
2890
+ },
2891
+ {
2892
+ "epoch": 15.980230642504118,
2893
+ "grad_norm": 0.0014526849845424294,
2894
+ "learning_rate": 2.005217391304348e-06,
2895
+ "loss": 0.0,
2896
+ "step": 9700
2897
+ },
2898
+ {
2899
+ "epoch": 16.021416803953873,
2900
+ "grad_norm": 0.0014665353810414672,
2901
+ "learning_rate": 1.9834782608695654e-06,
2902
+ "loss": 0.0,
2903
+ "step": 9725
2904
+ },
2905
+ {
2906
+ "epoch": 16.062602965403624,
2907
+ "grad_norm": 0.003292278153821826,
2908
+ "learning_rate": 1.9617391304347825e-06,
2909
+ "loss": 0.0,
2910
+ "step": 9750
2911
+ },
2912
+ {
2913
+ "epoch": 16.10378912685338,
2914
+ "grad_norm": 0.0014827033737674356,
2915
+ "learning_rate": 1.94e-06,
2916
+ "loss": 0.0,
2917
+ "step": 9775
2918
+ },
2919
+ {
2920
+ "epoch": 16.14497528830313,
2921
+ "grad_norm": 0.001969423145055771,
2922
+ "learning_rate": 1.9182608695652176e-06,
2923
+ "loss": 0.0,
2924
+ "step": 9800
2925
+ },
2926
+ {
2927
+ "epoch": 16.186161449752884,
2928
+ "grad_norm": 0.0010216091759502888,
2929
+ "learning_rate": 1.896521739130435e-06,
2930
+ "loss": 0.0,
2931
+ "step": 9825
2932
+ },
2933
+ {
2934
+ "epoch": 16.227347611202635,
2935
+ "grad_norm": 0.002770212711766362,
2936
+ "learning_rate": 1.8747826086956521e-06,
2937
+ "loss": 0.0,
2938
+ "step": 9850
2939
+ },
2940
+ {
2941
+ "epoch": 16.26853377265239,
2942
+ "grad_norm": 0.002100354991853237,
2943
+ "learning_rate": 1.8530434782608697e-06,
2944
+ "loss": 0.0,
2945
+ "step": 9875
2946
+ },
2947
+ {
2948
+ "epoch": 16.30971993410214,
2949
+ "grad_norm": 0.001876773894764483,
2950
+ "learning_rate": 1.8313043478260872e-06,
2951
+ "loss": 0.0,
2952
+ "step": 9900
2953
+ },
2954
+ {
2955
+ "epoch": 16.350906095551895,
2956
+ "grad_norm": 0.001118077663704753,
2957
+ "learning_rate": 1.8095652173913044e-06,
2958
+ "loss": 0.0,
2959
+ "step": 9925
2960
+ },
2961
+ {
2962
+ "epoch": 16.392092257001647,
2963
+ "grad_norm": 0.0008371293661184609,
2964
+ "learning_rate": 1.787826086956522e-06,
2965
+ "loss": 0.0,
2966
+ "step": 9950
2967
+ },
2968
+ {
2969
+ "epoch": 16.4332784184514,
2970
+ "grad_norm": 0.0007290031644515693,
2971
+ "learning_rate": 1.7660869565217393e-06,
2972
+ "loss": 0.0,
2973
+ "step": 9975
2974
+ },
2975
+ {
2976
+ "epoch": 16.474464579901152,
2977
+ "grad_norm": 0.0012829427141696215,
2978
+ "learning_rate": 1.7443478260869566e-06,
2979
+ "loss": 0.0,
2980
+ "step": 10000
2981
+ },
2982
+ {
2983
+ "epoch": 16.474464579901152,
2984
+ "eval_loss": 0.014236521907150745,
2985
+ "eval_runtime": 1633.2029,
2986
+ "eval_samples_per_second": 1.486,
2987
+ "eval_steps_per_second": 0.186,
2988
+ "eval_wer": 2.5013594344752583,
2989
+ "step": 10000
2990
+ },
2991
+ {
2992
+ "epoch": 16.515650741350907,
2993
+ "grad_norm": 0.0031344795133918524,
2994
+ "learning_rate": 1.722608695652174e-06,
2995
+ "loss": 0.0,
2996
+ "step": 10025
2997
+ },
2998
+ {
2999
+ "epoch": 16.556836902800658,
3000
+ "grad_norm": 0.0019432102562859654,
3001
+ "learning_rate": 1.7008695652173916e-06,
3002
+ "loss": 0.0,
3003
+ "step": 10050
3004
+ },
3005
+ {
3006
+ "epoch": 16.598023064250413,
3007
+ "grad_norm": 0.001211246708407998,
3008
+ "learning_rate": 1.6791304347826087e-06,
3009
+ "loss": 0.0,
3010
+ "step": 10075
3011
+ },
3012
+ {
3013
+ "epoch": 16.639209225700164,
3014
+ "grad_norm": 0.001109111588448286,
3015
+ "learning_rate": 1.6573913043478263e-06,
3016
+ "loss": 0.0,
3017
+ "step": 10100
3018
+ },
3019
+ {
3020
+ "epoch": 16.680395387149918,
3021
+ "grad_norm": 0.0025855598505586386,
3022
+ "learning_rate": 1.6356521739130434e-06,
3023
+ "loss": 0.0,
3024
+ "step": 10125
3025
+ },
3026
+ {
3027
+ "epoch": 16.72158154859967,
3028
+ "grad_norm": 0.0008468018495477736,
3029
+ "learning_rate": 1.613913043478261e-06,
3030
+ "loss": 0.0,
3031
+ "step": 10150
3032
+ },
3033
+ {
3034
+ "epoch": 16.762767710049424,
3035
+ "grad_norm": 0.0015540047315880656,
3036
+ "learning_rate": 1.5921739130434785e-06,
3037
+ "loss": 0.0,
3038
+ "step": 10175
3039
+ },
3040
+ {
3041
+ "epoch": 16.803953871499175,
3042
+ "grad_norm": 0.0022469875402748585,
3043
+ "learning_rate": 1.5704347826086957e-06,
3044
+ "loss": 0.0,
3045
+ "step": 10200
3046
+ },
3047
+ {
3048
+ "epoch": 16.84514003294893,
3049
+ "grad_norm": 0.00045295292511582375,
3050
+ "learning_rate": 1.5486956521739132e-06,
3051
+ "loss": 0.0,
3052
+ "step": 10225
3053
+ },
3054
+ {
3055
+ "epoch": 16.88632619439868,
3056
+ "grad_norm": 0.0018261218210682273,
3057
+ "learning_rate": 1.5269565217391306e-06,
3058
+ "loss": 0.0,
3059
+ "step": 10250
3060
+ },
3061
+ {
3062
+ "epoch": 16.927512355848435,
3063
+ "grad_norm": 0.0018533602124080062,
3064
+ "learning_rate": 1.505217391304348e-06,
3065
+ "loss": 0.0,
3066
+ "step": 10275
3067
+ },
3068
+ {
3069
+ "epoch": 16.968698517298186,
3070
+ "grad_norm": 0.0009767924202606082,
3071
+ "learning_rate": 1.4834782608695653e-06,
3072
+ "loss": 0.0,
3073
+ "step": 10300
3074
+ },
3075
+ {
3076
+ "epoch": 17.00988467874794,
3077
+ "grad_norm": 0.0009673904860392213,
3078
+ "learning_rate": 1.4617391304347828e-06,
3079
+ "loss": 0.0,
3080
+ "step": 10325
3081
+ },
3082
+ {
3083
+ "epoch": 17.051070840197692,
3084
+ "grad_norm": 0.002202032832428813,
3085
+ "learning_rate": 1.44e-06,
3086
+ "loss": 0.0,
3087
+ "step": 10350
3088
+ },
3089
+ {
3090
+ "epoch": 17.092257001647447,
3091
+ "grad_norm": 0.001030725659802556,
3092
+ "learning_rate": 1.4182608695652175e-06,
3093
+ "loss": 0.0,
3094
+ "step": 10375
3095
+ },
3096
+ {
3097
+ "epoch": 17.133443163097198,
3098
+ "grad_norm": 0.0008200545562431216,
3099
+ "learning_rate": 1.3965217391304347e-06,
3100
+ "loss": 0.0,
3101
+ "step": 10400
3102
+ },
3103
+ {
3104
+ "epoch": 17.174629324546952,
3105
+ "grad_norm": 0.0010224173311144114,
3106
+ "learning_rate": 1.3747826086956522e-06,
3107
+ "loss": 0.0,
3108
+ "step": 10425
3109
+ },
3110
+ {
3111
+ "epoch": 17.215815485996703,
3112
+ "grad_norm": 0.000996377901174128,
3113
+ "learning_rate": 1.3530434782608698e-06,
3114
+ "loss": 0.0,
3115
+ "step": 10450
3116
+ },
3117
+ {
3118
+ "epoch": 17.257001647446458,
3119
+ "grad_norm": 0.0017713250126689672,
3120
+ "learning_rate": 1.331304347826087e-06,
3121
+ "loss": 0.0,
3122
+ "step": 10475
3123
+ },
3124
+ {
3125
+ "epoch": 17.29818780889621,
3126
+ "grad_norm": 0.0008379866485483944,
3127
+ "learning_rate": 1.3095652173913045e-06,
3128
+ "loss": 0.0,
3129
+ "step": 10500
3130
+ },
3131
+ {
3132
+ "epoch": 17.29818780889621,
3133
+ "eval_loss": 0.01441162545233965,
3134
+ "eval_runtime": 1632.1765,
3135
+ "eval_samples_per_second": 1.487,
3136
+ "eval_steps_per_second": 0.186,
3137
+ "eval_wer": 2.55573681348559,
3138
+ "step": 10500
3139
+ },
3140
+ {
3141
+ "epoch": 17.339373970345964,
3142
+ "grad_norm": 0.001256891991943121,
3143
+ "learning_rate": 1.2878260869565219e-06,
3144
+ "loss": 0.0,
3145
+ "step": 10525
3146
+ },
3147
+ {
3148
+ "epoch": 17.38056013179572,
3149
+ "grad_norm": 0.003084406489506364,
3150
+ "learning_rate": 1.2660869565217392e-06,
3151
+ "loss": 0.0,
3152
+ "step": 10550
3153
+ },
3154
+ {
3155
+ "epoch": 17.42174629324547,
3156
+ "grad_norm": 0.002908531576395035,
3157
+ "learning_rate": 1.2443478260869568e-06,
3158
+ "loss": 0.0,
3159
+ "step": 10575
3160
+ },
3161
+ {
3162
+ "epoch": 17.462932454695224,
3163
+ "grad_norm": 0.0008340766653418541,
3164
+ "learning_rate": 1.2226086956521741e-06,
3165
+ "loss": 0.0,
3166
+ "step": 10600
3167
+ },
3168
+ {
3169
+ "epoch": 17.504118616144975,
3170
+ "grad_norm": 0.0014115035301074386,
3171
+ "learning_rate": 1.2008695652173915e-06,
3172
+ "loss": 0.0,
3173
+ "step": 10625
3174
+ },
3175
+ {
3176
+ "epoch": 17.54530477759473,
3177
+ "grad_norm": 0.001977501902729273,
3178
+ "learning_rate": 1.1791304347826088e-06,
3179
+ "loss": 0.0,
3180
+ "step": 10650
3181
+ },
3182
+ {
3183
+ "epoch": 17.58649093904448,
3184
+ "grad_norm": 0.0015209122793748975,
3185
+ "learning_rate": 1.1573913043478262e-06,
3186
+ "loss": 0.0,
3187
+ "step": 10675
3188
+ },
3189
+ {
3190
+ "epoch": 17.627677100494235,
3191
+ "grad_norm": 0.0009397733956575394,
3192
+ "learning_rate": 1.1356521739130435e-06,
3193
+ "loss": 0.0,
3194
+ "step": 10700
3195
+ },
3196
+ {
3197
+ "epoch": 17.668863261943986,
3198
+ "grad_norm": 0.0006780756521038711,
3199
+ "learning_rate": 1.1139130434782609e-06,
3200
+ "loss": 0.0,
3201
+ "step": 10725
3202
+ },
3203
+ {
3204
+ "epoch": 17.71004942339374,
3205
+ "grad_norm": 0.001231124042533338,
3206
+ "learning_rate": 1.0921739130434782e-06,
3207
+ "loss": 0.0,
3208
+ "step": 10750
3209
+ },
3210
+ {
3211
+ "epoch": 17.751235584843492,
3212
+ "grad_norm": 0.001504486077465117,
3213
+ "learning_rate": 1.0704347826086956e-06,
3214
+ "loss": 0.0,
3215
+ "step": 10775
3216
+ },
3217
+ {
3218
+ "epoch": 17.792421746293247,
3219
+ "grad_norm": 0.0006280929665081203,
3220
+ "learning_rate": 1.0486956521739131e-06,
3221
+ "loss": 0.0,
3222
+ "step": 10800
3223
+ },
3224
+ {
3225
+ "epoch": 17.833607907742998,
3226
+ "grad_norm": 0.0007227340247482061,
3227
+ "learning_rate": 1.0269565217391305e-06,
3228
+ "loss": 0.0,
3229
+ "step": 10825
3230
+ },
3231
+ {
3232
+ "epoch": 17.874794069192752,
3233
+ "grad_norm": 0.0014692394761368632,
3234
+ "learning_rate": 1.005217391304348e-06,
3235
+ "loss": 0.0,
3236
+ "step": 10850
3237
+ },
3238
+ {
3239
+ "epoch": 17.915980230642504,
3240
+ "grad_norm": 0.0010748576605692506,
3241
+ "learning_rate": 9.834782608695654e-07,
3242
+ "loss": 0.0,
3243
+ "step": 10875
3244
+ },
3245
+ {
3246
+ "epoch": 17.957166392092258,
3247
+ "grad_norm": 0.0012865326134487987,
3248
+ "learning_rate": 9.617391304347828e-07,
3249
+ "loss": 0.0,
3250
+ "step": 10900
3251
+ },
3252
+ {
3253
+ "epoch": 17.99835255354201,
3254
+ "grad_norm": 0.0034256454091519117,
3255
+ "learning_rate": 9.400000000000001e-07,
3256
+ "loss": 0.0,
3257
+ "step": 10925
3258
+ },
3259
+ {
3260
+ "epoch": 18.039538714991764,
3261
+ "grad_norm": 0.001147790695540607,
3262
+ "learning_rate": 9.182608695652175e-07,
3263
+ "loss": 0.0,
3264
+ "step": 10950
3265
+ },
3266
+ {
3267
+ "epoch": 18.080724876441515,
3268
+ "grad_norm": 0.0010212146444246173,
3269
+ "learning_rate": 8.965217391304348e-07,
3270
+ "loss": 0.0,
3271
+ "step": 10975
3272
+ },
3273
+ {
3274
+ "epoch": 18.12191103789127,
3275
+ "grad_norm": 0.0022493095602840185,
3276
+ "learning_rate": 8.747826086956522e-07,
3277
+ "loss": 0.0,
3278
+ "step": 11000
3279
+ },
3280
+ {
3281
+ "epoch": 18.12191103789127,
3282
+ "eval_loss": 0.014553242363035679,
3283
+ "eval_runtime": 1629.5576,
3284
+ "eval_samples_per_second": 1.489,
3285
+ "eval_steps_per_second": 0.187,
3286
+ "eval_wer": 2.5168958284782104,
3287
+ "step": 11000
3288
+ }
3289
+ ],
3290
+ "logging_steps": 25,
3291
+ "max_steps": 12000,
3292
+ "num_input_tokens_seen": 0,
3293
+ "num_train_epochs": 20,
3294
+ "save_steps": 1000,
3295
+ "stateful_callbacks": {
3296
+ "TrainerControl": {
3297
+ "args": {
3298
+ "should_epoch_stop": false,
3299
+ "should_evaluate": false,
3300
+ "should_log": false,
3301
+ "should_save": true,
3302
+ "should_training_stop": false
3303
+ },
3304
+ "attributes": {}
3305
+ }
3306
+ },
3307
+ "total_flos": 5.077025230749696e+19,
3308
+ "train_batch_size": 16,
3309
+ "trial_name": null,
3310
+ "trial_params": null
3311
+ }
checkpoint-11000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d0b57b03f0ccd85eb2bbc846799ee3434ae5f028437174954b7053ae877da2c
3
+ size 5240