Hugging Face's logo

aomocelin
/
moonshine_tiny_pt_v05

Automatic Speech Recognition
Transformers
Safetensors
moonshine
Generated from Trainer
Model card Files
xet
 Community
aomocelin commited on
Commit
e1e00bd
·
verified ·
1 Parent(s): 12dc773

Model save

Browse files
Files changed (9) hide show
  1. README.md +213 -0
  2. final/config.json +37 -0
  3. final/generation_config.json +17 -0
  4. final/model.safetensors +3 -0
  5. final/processor_config.json +12 -0
  6. final/tokenizer.json +0 -0
  7. final/tokenizer_config.json +11 -0
  8. final/training_args.bin +3 -0
  9. model.safetensors +1 -1
README.md ADDED
@@ -0,0 +1,213 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: mit
4
+ base_model: aomocelin/moonshine_tiny_pt_v04
5
+ tags:
6
+ - generated_from_trainer
7
+ metrics:
8
+ - wer
9
+ model-index:
10
+ - name: moonshine_tiny_pt_v05
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # moonshine_tiny_pt_v05
18
+
19
+ This model is a fine-tuned version of [aomocelin/moonshine_tiny_pt_v04](https://huggingface.co/aomocelin/moonshine_tiny_pt_v04) on an unknown dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 10.5343
22
+ - Wer: 38.0208
23
+
24
+ ## Model description
25
+
26
+ More information needed
27
+
28
+ ## Intended uses & limitations
29
+
30
+ More information needed
31
+
32
+ ## Training and evaluation data
33
+
34
+ More information needed
35
+
36
+ ## Training procedure
37
+
38
+ ### Training hyperparameters
39
+
40
+ The following hyperparameters were used during training:
41
+ - learning_rate: 5e-06
42
+ - train_batch_size: 4
43
+ - eval_batch_size: 128
44
+ - seed: 42
45
+ - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
46
+ - lr_scheduler_type: linear
47
+ - lr_scheduler_warmup_steps: 0.03
48
+ - training_steps: 15000
49
+ - mixed_precision_training: Native AMP
50
+ - label_smoothing_factor: 0.1
51
+
52
+ ### Training results
53
+
54
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
55
+ |:-------------:|:--------:|:-----:|:---------------:|:-------:|
56
+ | 2.1248 | 1.6667 | 100 | 13.1150 | 33.3333 |
57
+ | 1.8041 | 3.3333 | 200 | 13.0609 | 10.9375 |
58
+ | 1.6729 | 5.0 | 300 | 13.0083 | 7.2917 |
59
+ | 1.5993 | 6.6667 | 400 | 12.9088 | 7.8125 |
60
+ | 1.5518 | 8.3333 | 500 | 12.7937 | 7.8125 |
61
+ | 1.5103 | 10.0 | 600 | 12.7596 | 9.375 |
62
+ | 1.4899 | 11.6667 | 700 | 12.6691 | 9.8958 |
63
+ | 1.4796 | 13.3333 | 800 | 12.5641 | 10.4167 |
64
+ | 1.4713 | 15.0 | 900 | 12.5095 | 10.4167 |
65
+ | 1.4647 | 16.6667 | 1000 | 12.4242 | 10.4167 |
66
+ | 1.4601 | 18.3333 | 1100 | 12.3369 | 10.4167 |
67
+ | 1.4557 | 20.0 | 1200 | 12.2838 | 10.4167 |
68
+ | 1.4527 | 21.6667 | 1300 | 12.2227 | 10.4167 |
69
+ | 1.4497 | 23.3333 | 1400 | 12.1341 | 11.9792 |
70
+ | 1.4468 | 25.0 | 1500 | 12.1018 | 11.4583 |
71
+ | 1.4443 | 26.6667 | 1600 | 12.0279 | 11.9792 |
72
+ | 1.4427 | 28.3333 | 1700 | 11.9715 | 10.4167 |
73
+ | 1.4403 | 30.0 | 1800 | 11.9419 | 10.4167 |
74
+ | 1.4387 | 31.6667 | 1900 | 11.8966 | 10.4167 |
75
+ | 1.4368 | 33.3333 | 2000 | 11.8448 | 9.8958 |
76
+ | 1.4350 | 35.0 | 2100 | 11.8112 | 10.4167 |
77
+ | 1.4335 | 36.6667 | 2200 | 11.7620 | 10.9375 |
78
+ | 1.4325 | 38.3333 | 2300 | 11.7204 | 10.4167 |
79
+ | 1.4301 | 40.0 | 2400 | 11.6758 | 11.4583 |
80
+ | 1.4293 | 41.6667 | 2500 | 11.6467 | 10.4167 |
81
+ | 1.4282 | 43.3333 | 2600 | 11.5954 | 10.4167 |
82
+ | 1.4264 | 45.0 | 2700 | 11.5686 | 12.5 |
83
+ | 1.4253 | 46.6667 | 2800 | 11.5368 | 14.5833 |
84
+ | 1.4245 | 48.3333 | 2900 | 11.4964 | 14.5833 |
85
+ | 1.4230 | 50.0 | 3000 | 11.4693 | 14.5833 |
86
+ | 1.4223 | 51.6667 | 3100 | 11.4469 | 14.5833 |
87
+ | 1.4214 | 53.3333 | 3200 | 11.4066 | 14.0625 |
88
+ | 1.4201 | 55.0 | 3300 | 11.3728 | 14.0625 |
89
+ | 1.4193 | 56.6667 | 3400 | 11.3398 | 14.0625 |
90
+ | 1.4186 | 58.3333 | 3500 | 11.3248 | 14.0625 |
91
+ | 1.4173 | 60.0 | 3600 | 11.3173 | 15.625 |
92
+ | 1.4166 | 61.6667 | 3700 | 11.2547 | 15.625 |
93
+ | 1.4160 | 63.3333 | 3800 | 11.2333 | 15.625 |
94
+ | 1.4148 | 65.0 | 3900 | 11.2121 | 16.6667 |
95
+ | 1.4142 | 66.6667 | 4000 | 11.1808 | 16.6667 |
96
+ | 1.4138 | 68.3333 | 4100 | 11.1715 | 18.2292 |
97
+ | 1.4127 | 70.0 | 4200 | 11.1491 | 18.75 |
98
+ | 1.4121 | 71.6667 | 4300 | 11.1304 | 18.2292 |
99
+ | 1.4124 | 73.3333 | 4400 | 11.0937 | 18.2292 |
100
+ | 1.4110 | 75.0 | 4500 | 11.0799 | 18.2292 |
101
+ | 1.4106 | 76.6667 | 4600 | 11.0685 | 21.3542 |
102
+ | 1.4101 | 78.3333 | 4700 | 11.0641 | 20.8333 |
103
+ | 1.4093 | 80.0 | 4800 | 11.0537 | 18.75 |
104
+ | 1.4091 | 81.6667 | 4900 | 11.0109 | 20.8333 |
105
+ | 1.4082 | 83.3333 | 5000 | 11.0075 | 21.875 |
106
+ | 1.4075 | 85.0 | 5100 | 11.0054 | 21.875 |
107
+ | 1.4070 | 86.6667 | 5200 | 10.9798 | 21.875 |
108
+ | 1.4071 | 88.3333 | 5300 | 10.9378 | 21.875 |
109
+ | 1.4062 | 90.0 | 5400 | 10.9437 | 21.3542 |
110
+ | 1.4057 | 91.6667 | 5500 | 10.9527 | 22.9167 |
111
+ | 1.4054 | 93.3333 | 5600 | 10.8885 | 25.0 |
112
+ | 1.4050 | 95.0 | 5700 | 10.9185 | 23.9583 |
113
+ | 1.4046 | 96.6667 | 5800 | 10.9053 | 24.4792 |
114
+ | 1.4042 | 98.3333 | 5900 | 10.8931 | 26.5625 |
115
+ | 1.4037 | 100.0 | 6000 | 10.8834 | 23.9583 |
116
+ | 1.4035 | 101.6667 | 6100 | 10.8508 | 25.0 |
117
+ | 1.4029 | 103.3333 | 6200 | 10.8649 | 25.5208 |
118
+ | 1.4026 | 105.0 | 6300 | 10.8728 | 26.0417 |
119
+ | 1.4026 | 106.6667 | 6400 | 10.8166 | 27.0833 |
120
+ | 1.4019 | 108.3333 | 6500 | 10.8345 | 27.0833 |
121
+ | 1.4018 | 110.0 | 6600 | 10.8192 | 27.6042 |
122
+ | 1.4012 | 111.6667 | 6700 | 10.8206 | 27.0833 |
123
+ | 1.4009 | 113.3333 | 6800 | 10.8049 | 27.0833 |
124
+ | 1.4007 | 115.0 | 6900 | 10.7937 | 27.6042 |
125
+ | 1.4007 | 116.6667 | 7000 | 10.7734 | 30.2083 |
126
+ | 1.4003 | 118.3333 | 7100 | 10.7916 | 29.6875 |
127
+ | 1.3999 | 120.0 | 7200 | 10.7654 | 27.6042 |
128
+ | 1.3997 | 121.6667 | 7300 | 10.7745 | 29.1667 |
129
+ | 1.3994 | 123.3333 | 7400 | 10.7421 | 28.6458 |
130
+ | 1.3991 | 125.0 | 7500 | 10.7555 | 29.1667 |
131
+ | 1.3990 | 126.6667 | 7600 | 10.7447 | 30.7292 |
132
+ | 1.3987 | 128.3333 | 7700 | 10.7374 | 30.7292 |
133
+ | 1.3985 | 130.0 | 7800 | 10.7313 | 30.7292 |
134
+ | 1.3982 | 131.6667 | 7900 | 10.7239 | 30.7292 |
135
+ | 1.3981 | 133.3333 | 8000 | 10.7134 | 31.25 |
136
+ | 1.3977 | 135.0 | 8100 | 10.7242 | 30.7292 |
137
+ | 1.3976 | 136.6667 | 8200 | 10.7035 | 31.25 |
138
+ | 1.3975 | 138.3333 | 8300 | 10.7001 | 31.25 |
139
+ | 1.3970 | 140.0 | 8400 | 10.6934 | 31.7708 |
140
+ | 1.3971 | 141.6667 | 8500 | 10.6830 | 31.7708 |
141
+ | 1.3965 | 143.3333 | 8600 | 10.6856 | 31.7708 |
142
+ | 1.3964 | 145.0 | 8700 | 10.6801 | 32.2917 |
143
+ | 1.3964 | 146.6667 | 8800 | 10.6654 | 31.25 |
144
+ | 1.3962 | 148.3333 | 8900 | 10.6820 | 32.2917 |
145
+ | 1.3960 | 150.0 | 9000 | 10.6621 | 33.8542 |
146
+ | 1.3959 | 151.6667 | 9100 | 10.6500 | 32.8125 |
147
+ | 1.3957 | 153.3333 | 9200 | 10.6611 | 31.7708 |
148
+ | 1.3954 | 155.0 | 9300 | 10.6551 | 33.8542 |
149
+ | 1.3954 | 156.6667 | 9400 | 10.6411 | 32.8125 |
150
+ | 1.3953 | 158.3333 | 9500 | 10.6423 | 34.375 |
151
+ | 1.3949 | 160.0 | 9600 | 10.6422 | 34.375 |
152
+ | 1.3948 | 161.6667 | 9700 | 10.6324 | 34.375 |
153
+ | 1.3947 | 163.3333 | 9800 | 10.6346 | 34.375 |
154
+ | 1.3944 | 165.0 | 9900 | 10.6289 | 34.375 |
155
+ | 1.3944 | 166.6667 | 10000 | 10.6224 | 34.375 |
156
+ | 1.3946 | 168.3333 | 10100 | 10.6211 | 34.375 |
157
+ | 1.3940 | 170.0 | 10200 | 10.6190 | 34.8958 |
158
+ | 1.3941 | 171.6667 | 10300 | 10.6132 | 34.8958 |
159
+ | 1.3938 | 173.3333 | 10400 | 10.6122 | 34.375 |
160
+ | 1.3938 | 175.0 | 10500 | 10.6104 | 34.8958 |
161
+ | 1.3938 | 176.6667 | 10600 | 10.6116 | 34.8958 |
162
+ | 1.3937 | 178.3333 | 10700 | 10.5989 | 34.8958 |
163
+ | 1.3934 | 180.0 | 10800 | 10.5986 | 35.4167 |
164
+ | 1.3934 | 181.6667 | 10900 | 10.5940 | 34.8958 |
165
+ | 1.3932 | 183.3333 | 11000 | 10.5894 | 35.9375 |
166
+ | 1.3930 | 185.0 | 11100 | 10.5905 | 36.4583 |
167
+ | 1.3930 | 186.6667 | 11200 | 10.5899 | 35.4167 |
168
+ | 1.3928 | 188.3333 | 11300 | 10.5872 | 36.4583 |
169
+ | 1.3927 | 190.0 | 11400 | 10.5828 | 35.9375 |
170
+ | 1.3927 | 191.6667 | 11500 | 10.5804 | 36.9792 |
171
+ | 1.3926 | 193.3333 | 11600 | 10.5769 | 39.0625 |
172
+ | 1.3924 | 195.0 | 11700 | 10.5745 | 38.5417 |
173
+ | 1.3925 | 196.6667 | 11800 | 10.5750 | 36.9792 |
174
+ | 1.3924 | 198.3333 | 11900 | 10.5698 | 36.4583 |
175
+ | 1.3922 | 200.0 | 12000 | 10.5762 | 38.5417 |
176
+ | 1.3923 | 201.6667 | 12100 | 10.5659 | 39.0625 |
177
+ | 1.3923 | 203.3333 | 12200 | 10.5648 | 39.0625 |
178
+ | 1.3918 | 205.0 | 12300 | 10.5621 | 38.0208 |
179
+ | 1.3920 | 206.6667 | 12400 | 10.5602 | 39.0625 |
180
+ | 1.3919 | 208.3333 | 12500 | 10.5582 | 38.5417 |
181
+ | 1.3917 | 210.0 | 12600 | 10.5566 | 37.5 |
182
+ | 1.3919 | 211.6667 | 12700 | 10.5561 | 39.0625 |
183
+ | 1.3918 | 213.3333 | 12800 | 10.5544 | 38.0208 |
184
+ | 1.3915 | 215.0 | 12900 | 10.5528 | 39.0625 |
185
+ | 1.3915 | 216.6667 | 13000 | 10.5500 | 39.0625 |
186
+ | 1.3913 | 218.3333 | 13100 | 10.5513 | 38.5417 |
187
+ | 1.3914 | 220.0 | 13200 | 10.5464 | 39.0625 |
188
+ | 1.3914 | 221.6667 | 13300 | 10.5456 | 37.5 |
189
+ | 1.3914 | 223.3333 | 13400 | 10.5433 | 38.5417 |
190
+ | 1.3912 | 225.0 | 13500 | 10.5426 | 39.0625 |
191
+ | 1.3912 | 226.6667 | 13600 | 10.5414 | 37.5 |
192
+ | 1.3912 | 228.3333 | 13700 | 10.5405 | 38.5417 |
193
+ | 1.3912 | 230.0 | 13800 | 10.5399 | 38.5417 |
194
+ | 1.3912 | 231.6667 | 13900 | 10.5391 | 37.5 |
195
+ | 1.3911 | 233.3333 | 14000 | 10.5369 | 36.9792 |
196
+ | 1.3910 | 235.0 | 14100 | 10.5375 | 38.0208 |
197
+ | 1.3913 | 236.6667 | 14200 | 10.5385 | 38.5417 |
198
+ | 1.3909 | 238.3333 | 14300 | 10.5371 | 38.5417 |
199
+ | 1.3910 | 240.0 | 14400 | 10.5352 | 38.0208 |
200
+ | 1.3910 | 241.6667 | 14500 | 10.5349 | 38.0208 |
201
+ | 1.3909 | 243.3333 | 14600 | 10.5345 | 38.0208 |
202
+ | 1.3908 | 245.0 | 14700 | 10.5347 | 38.0208 |
203
+ | 1.3910 | 246.6667 | 14800 | 10.5346 | 38.0208 |
204
+ | 1.3908 | 248.3333 | 14900 | 10.5355 | 38.0208 |
205
+ | 1.3909 | 250.0 | 15000 | 10.5343 | 38.0208 |
206
+
207
+
208
+ ### Framework versions
209
+
210
+ - Transformers 5.12.1
211
+ - Pytorch 2.11.0+cu128
212
+ - Datasets 5.0.0
213
+ - Tokenizers 0.22.2
final/config.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "MoonshineForConditionalGeneration"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "decoder_hidden_act": "silu",
9
+ "decoder_num_attention_heads": 8,
10
+ "decoder_num_hidden_layers": 6,
11
+ "decoder_num_key_value_heads": 8,
12
+ "decoder_start_token_id": 1,
13
+ "dtype": "float32",
14
+ "encoder_hidden_act": "gelu",
15
+ "encoder_num_attention_heads": 8,
16
+ "encoder_num_hidden_layers": 6,
17
+ "encoder_num_key_value_heads": 8,
18
+ "eos_token_id": 2,
19
+ "hidden_size": 288,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 1152,
22
+ "is_encoder_decoder": true,
23
+ "max_position_embeddings": 194,
24
+ "model_type": "moonshine",
25
+ "pad_head_dim_to_multiple_of": 8,
26
+ "pad_token_id": 2,
27
+ "partial_rotary_factor": 0.9,
28
+ "rope_parameters": {
29
+ "partial_rotary_factor": 0.9,
30
+ "rope_theta": 10000.0,
31
+ "rope_type": "default"
32
+ },
33
+ "tie_word_embeddings": true,
34
+ "transformers_version": "5.12.1",
35
+ "use_cache": false,
36
+ "vocab_size": 32768
37
+ }
final/generation_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "decoder_start_token_id": 1,
5
+ "do_sample": false,
6
+ "early_stopping": true,
7
+ "eos_token_id": [
8
+ 2
9
+ ],
10
+ "length_penalty": 1.2,
11
+ "max_length": 194,
12
+ "no_repeat_ngram_size": 2,
13
+ "num_beams": 5,
14
+ "pad_token_id": 2,
15
+ "repetition_penalty": 1.2,
16
+ "transformers_version": "5.12.1"
17
+ }
final/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e5e08f92e588327c8e56f43206f375ab79f2cc22279027ce00d40235c762b81
3
+ size 108389192
final/processor_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "feature_extractor": {
3
+ "do_normalize": false,
4
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
5
+ "feature_size": 1,
6
+ "padding_side": "right",
7
+ "padding_value": 0.0,
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ },
11
+ "processor_class": "Wav2Vec2Processor"
12
+ }
final/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
final/tokenizer_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "bos_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "is_local": false,
6
+ "local_files_only": false,
7
+ "model_max_length": 1000000000000000019884624838656,
8
+ "pad_token": "</s>",
9
+ "processor_class": "Wav2Vec2Processor",
10
+ "tokenizer_class": "TokenizersBackend"
11
+ }
final/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c2a78e5b7f5b3317907b7a24c89083c70c60b9a196099a65fba40bef5203b1c
3
+ size 5393
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:cb36e3f3e2c764a3af5b8d14b6ee63211894e2cdf5a20853f5d5a4ed56450eff
3
  size 108389192
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e5e08f92e588327c8e56f43206f375ab79f2cc22279027ce00d40235c762b81
3
  size 108389192