FelixK7 commited on
Commit
b8c0079
1 Parent(s): 3763598

End of training

Browse files
Files changed (3) hide show
  1. README.md +132 -0
  2. generation_config.json +249 -0
  3. model.safetensors +1 -1
README.md ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ language:
4
+ - lv
5
+ license: apache-2.0
6
+ base_model: FelixK7/whisper-medium-lv
7
+ tags:
8
+ - hf-asr-leaderboard
9
+ - generated_from_trainer
10
+ datasets:
11
+ - mozilla-foundation/common_voice_16_1
12
+ metrics:
13
+ - wer
14
+ model-index:
15
+ - name: Whisper medium LV - Felikss Kleins
16
+ results:
17
+ - task:
18
+ name: Automatic Speech Recognition
19
+ type: automatic-speech-recognition
20
+ dataset:
21
+ name: Common Voice 16.1
22
+ type: mozilla-foundation/common_voice_16_1
23
+ config: lv
24
+ split: None
25
+ args: 'config: lv, split: test'
26
+ metrics:
27
+ - name: Wer
28
+ type: wer
29
+ value: 9.459716154242761
30
+ ---
31
+
32
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
33
+ should probably proofread and complete it, then remove this comment. -->
34
+
35
+ # Whisper medium LV - Felikss Kleins
36
+
37
+ This model is a fine-tuned version of [FelixK7/whisper-medium-lv](https://huggingface.co/FelixK7/whisper-medium-lv) on the Common Voice 16.1 dataset.
38
+ It achieves the following results on the evaluation set:
39
+ - Loss: 0.2053
40
+ - Wer: 9.4597
41
+
42
+ ## Model description
43
+
44
+ More information needed
45
+
46
+ ## Intended uses & limitations
47
+
48
+ More information needed
49
+
50
+ ## Training and evaluation data
51
+
52
+ More information needed
53
+
54
+ ## Training procedure
55
+
56
+ ### Training hyperparameters
57
+
58
+ The following hyperparameters were used during training:
59
+ - learning_rate: 3e-05
60
+ - train_batch_size: 32
61
+ - eval_batch_size: 32
62
+ - seed: 42
63
+ - gradient_accumulation_steps: 2
64
+ - total_train_batch_size: 64
65
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
66
+ - lr_scheduler_type: linear
67
+ - lr_scheduler_warmup_steps: 500
68
+ - training_steps: 10000
69
+ - mixed_precision_training: Native AMP
70
+
71
+ ### Training results
72
+
73
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
74
+ |:-------------:|:-------:|:-----:|:---------------:|:-------:|
75
+ | No log | 0.02 | 200 | 0.1318 | 7.6741 |
76
+ | 0.0445 | 1.0199 | 400 | 0.1527 | 8.4475 |
77
+ | 0.0338 | 2.0199 | 600 | 0.1703 | 9.7148 |
78
+ | 0.0345 | 3.0198 | 800 | 0.1725 | 9.7392 |
79
+ | 0.0311 | 4.0198 | 1000 | 0.1789 | 9.8830 |
80
+ | 0.0311 | 5.0198 | 1200 | 0.1792 | 10.0187 |
81
+ | 0.0288 | 6.0197 | 1400 | 0.1858 | 9.6063 |
82
+ | 0.0237 | 7.0197 | 1600 | 0.1839 | 9.8803 |
83
+ | 0.022 | 8.0196 | 1800 | 0.1847 | 10.2955 |
84
+ | 0.0198 | 9.0196 | 2000 | 0.1878 | 9.8885 |
85
+ | 0.0198 | 10.0195 | 2200 | 0.1909 | 9.9237 |
86
+ | 0.0183 | 11.0195 | 2400 | 0.1948 | 10.1924 |
87
+ | 0.0161 | 12.0194 | 2600 | 0.1951 | 10.4122 |
88
+ | 0.0154 | 13.0193 | 2800 | 0.1952 | 9.9997 |
89
+ | 0.0141 | 14.0193 | 3000 | 0.1972 | 10.1001 |
90
+ | 0.0141 | 15.0192 | 3200 | 0.1976 | 10.1544 |
91
+ | 0.0118 | 16.0192 | 3400 | 0.2014 | 10.4258 |
92
+ | 0.0115 | 17.0191 | 3600 | 0.2021 | 10.6890 |
93
+ | 0.0106 | 18.0191 | 3800 | 0.2005 | 10.1951 |
94
+ | 0.0092 | 19.0191 | 4000 | 0.2022 | 10.4638 |
95
+ | 0.0092 | 20.019 | 4200 | 0.2003 | 10.0947 |
96
+ | 0.0089 | 21.0190 | 4400 | 0.2043 | 9.8776 |
97
+ | 0.0085 | 22.0189 | 4600 | 0.2063 | 10.4719 |
98
+ | 0.0083 | 23.0189 | 4800 | 0.2067 | 10.0540 |
99
+ | 0.0069 | 24.0188 | 5000 | 0.2058 | 9.7908 |
100
+ | 0.0069 | 25.0188 | 5200 | 0.2056 | 10.4583 |
101
+ | 0.0078 | 26.0187 | 5400 | 0.2090 | 10.1843 |
102
+ | 0.0063 | 27.0187 | 5600 | 0.2096 | 10.2250 |
103
+ | 0.0058 | 28.0186 | 5800 | 0.2047 | 10.2602 |
104
+ | 0.0052 | 29.0186 | 6000 | 0.2087 | 9.9319 |
105
+ | 0.0052 | 30.0185 | 6200 | 0.2040 | 10.0811 |
106
+ | 0.0054 | 31.0185 | 6400 | 0.2081 | 9.9482 |
107
+ | 0.0045 | 32.0184 | 6600 | 0.2063 | 9.6849 |
108
+ | 0.004 | 33.0183 | 6800 | 0.2077 | 10.0052 |
109
+ | 0.0035 | 34.0183 | 7000 | 0.2105 | 10.1056 |
110
+ | 0.0035 | 35.0183 | 7200 | 0.2075 | 9.6985 |
111
+ | 0.0035 | 36.0182 | 7400 | 0.2075 | 9.6063 |
112
+ | 0.003 | 37.0181 | 7600 | 0.2115 | 9.8396 |
113
+ | 0.0027 | 38.0181 | 7800 | 0.2061 | 9.5601 |
114
+ | 0.0025 | 39.0181 | 8000 | 0.2082 | 9.6252 |
115
+ | 0.0025 | 40.018 | 8200 | 0.2052 | 9.5520 |
116
+ | 0.0023 | 41.0179 | 8400 | 0.2060 | 9.7826 |
117
+ | 0.0024 | 42.0179 | 8600 | 0.2083 | 9.6361 |
118
+ | 0.002 | 43.0179 | 8800 | 0.2069 | 9.5981 |
119
+ | 0.0021 | 44.0178 | 9000 | 0.2051 | 9.3892 |
120
+ | 0.0021 | 45.0177 | 9200 | 0.2054 | 9.3756 |
121
+ | 0.0019 | 46.0177 | 9400 | 0.2049 | 9.5167 |
122
+ | 0.0017 | 47.0177 | 9600 | 0.2051 | 9.4733 |
123
+ | 0.0017 | 48.0176 | 9800 | 0.2050 | 9.4923 |
124
+ | 0.0014 | 49.0175 | 10000 | 0.2053 | 9.4597 |
125
+
126
+
127
+ ### Framework versions
128
+
129
+ - Transformers 4.45.0.dev0
130
+ - Pytorch 2.0.1
131
+ - Datasets 3.0.0
132
+ - Tokenizers 0.19.1
generation_config.json ADDED
@@ -0,0 +1,249 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alignment_heads": [
3
+ [
4
+ 13,
5
+ 15
6
+ ],
7
+ [
8
+ 15,
9
+ 4
10
+ ],
11
+ [
12
+ 15,
13
+ 15
14
+ ],
15
+ [
16
+ 16,
17
+ 1
18
+ ],
19
+ [
20
+ 20,
21
+ 0
22
+ ],
23
+ [
24
+ 23,
25
+ 4
26
+ ]
27
+ ],
28
+ "begin_suppress_tokens": [
29
+ 220,
30
+ 50257
31
+ ],
32
+ "bos_token_id": 50257,
33
+ "decoder_start_token_id": 50258,
34
+ "eos_token_id": 50257,
35
+ "forced_decoder_ids": [
36
+ [
37
+ 1,
38
+ null
39
+ ],
40
+ [
41
+ 2,
42
+ 50359
43
+ ]
44
+ ],
45
+ "is_multilingual": true,
46
+ "lang_to_id": {
47
+ "<|af|>": 50327,
48
+ "<|am|>": 50334,
49
+ "<|ar|>": 50272,
50
+ "<|as|>": 50350,
51
+ "<|az|>": 50304,
52
+ "<|ba|>": 50355,
53
+ "<|be|>": 50330,
54
+ "<|bg|>": 50292,
55
+ "<|bn|>": 50302,
56
+ "<|bo|>": 50347,
57
+ "<|br|>": 50309,
58
+ "<|bs|>": 50315,
59
+ "<|ca|>": 50270,
60
+ "<|cs|>": 50283,
61
+ "<|cy|>": 50297,
62
+ "<|da|>": 50285,
63
+ "<|de|>": 50261,
64
+ "<|el|>": 50281,
65
+ "<|en|>": 50259,
66
+ "<|es|>": 50262,
67
+ "<|et|>": 50307,
68
+ "<|eu|>": 50310,
69
+ "<|fa|>": 50300,
70
+ "<|fi|>": 50277,
71
+ "<|fo|>": 50338,
72
+ "<|fr|>": 50265,
73
+ "<|gl|>": 50319,
74
+ "<|gu|>": 50333,
75
+ "<|haw|>": 50352,
76
+ "<|ha|>": 50354,
77
+ "<|he|>": 50279,
78
+ "<|hi|>": 50276,
79
+ "<|hr|>": 50291,
80
+ "<|ht|>": 50339,
81
+ "<|hu|>": 50286,
82
+ "<|hy|>": 50312,
83
+ "<|id|>": 50275,
84
+ "<|is|>": 50311,
85
+ "<|it|>": 50274,
86
+ "<|ja|>": 50266,
87
+ "<|jw|>": 50356,
88
+ "<|ka|>": 50329,
89
+ "<|kk|>": 50316,
90
+ "<|km|>": 50323,
91
+ "<|kn|>": 50306,
92
+ "<|ko|>": 50264,
93
+ "<|la|>": 50294,
94
+ "<|lb|>": 50345,
95
+ "<|ln|>": 50353,
96
+ "<|lo|>": 50336,
97
+ "<|lt|>": 50293,
98
+ "<|lv|>": 50301,
99
+ "<|mg|>": 50349,
100
+ "<|mi|>": 50295,
101
+ "<|mk|>": 50308,
102
+ "<|ml|>": 50296,
103
+ "<|mn|>": 50314,
104
+ "<|mr|>": 50320,
105
+ "<|ms|>": 50282,
106
+ "<|mt|>": 50343,
107
+ "<|my|>": 50346,
108
+ "<|ne|>": 50313,
109
+ "<|nl|>": 50271,
110
+ "<|nn|>": 50342,
111
+ "<|no|>": 50288,
112
+ "<|oc|>": 50328,
113
+ "<|pa|>": 50321,
114
+ "<|pl|>": 50269,
115
+ "<|ps|>": 50340,
116
+ "<|pt|>": 50267,
117
+ "<|ro|>": 50284,
118
+ "<|ru|>": 50263,
119
+ "<|sa|>": 50344,
120
+ "<|sd|>": 50332,
121
+ "<|si|>": 50322,
122
+ "<|sk|>": 50298,
123
+ "<|sl|>": 50305,
124
+ "<|sn|>": 50324,
125
+ "<|so|>": 50326,
126
+ "<|sq|>": 50317,
127
+ "<|sr|>": 50303,
128
+ "<|su|>": 50357,
129
+ "<|sv|>": 50273,
130
+ "<|sw|>": 50318,
131
+ "<|ta|>": 50287,
132
+ "<|te|>": 50299,
133
+ "<|tg|>": 50331,
134
+ "<|th|>": 50289,
135
+ "<|tk|>": 50341,
136
+ "<|tl|>": 50348,
137
+ "<|tr|>": 50268,
138
+ "<|tt|>": 50351,
139
+ "<|uk|>": 50280,
140
+ "<|ur|>": 50290,
141
+ "<|uz|>": 50337,
142
+ "<|vi|>": 50278,
143
+ "<|yi|>": 50335,
144
+ "<|yo|>": 50325,
145
+ "<|zh|>": 50260
146
+ },
147
+ "max_initial_timestamp_index": 50,
148
+ "max_length": 448,
149
+ "no_timestamps_token_id": 50363,
150
+ "pad_token_id": 50257,
151
+ "prev_sot_token_id": 50361,
152
+ "return_timestamps": false,
153
+ "suppress_tokens": [
154
+ 1,
155
+ 2,
156
+ 7,
157
+ 8,
158
+ 9,
159
+ 10,
160
+ 14,
161
+ 25,
162
+ 26,
163
+ 27,
164
+ 28,
165
+ 29,
166
+ 31,
167
+ 58,
168
+ 59,
169
+ 60,
170
+ 61,
171
+ 62,
172
+ 63,
173
+ 90,
174
+ 91,
175
+ 92,
176
+ 93,
177
+ 359,
178
+ 503,
179
+ 522,
180
+ 542,
181
+ 873,
182
+ 893,
183
+ 902,
184
+ 918,
185
+ 922,
186
+ 931,
187
+ 1350,
188
+ 1853,
189
+ 1982,
190
+ 2460,
191
+ 2627,
192
+ 3246,
193
+ 3253,
194
+ 3268,
195
+ 3536,
196
+ 3846,
197
+ 3961,
198
+ 4183,
199
+ 4667,
200
+ 6585,
201
+ 6647,
202
+ 7273,
203
+ 9061,
204
+ 9383,
205
+ 10428,
206
+ 10929,
207
+ 11938,
208
+ 12033,
209
+ 12331,
210
+ 12562,
211
+ 13793,
212
+ 14157,
213
+ 14635,
214
+ 15265,
215
+ 15618,
216
+ 16553,
217
+ 16604,
218
+ 18362,
219
+ 18956,
220
+ 20075,
221
+ 21675,
222
+ 22520,
223
+ 26130,
224
+ 26161,
225
+ 26435,
226
+ 28279,
227
+ 29464,
228
+ 31650,
229
+ 32302,
230
+ 32470,
231
+ 36865,
232
+ 42863,
233
+ 47425,
234
+ 49870,
235
+ 50254,
236
+ 50258,
237
+ 50358,
238
+ 50359,
239
+ 50360,
240
+ 50361,
241
+ 50362
242
+ ],
243
+ "task_to_id": {
244
+ "transcribe": 50359,
245
+ "translate": 50358
246
+ },
247
+ "transformers_version": "4.45.0.dev0",
248
+ "use_cache": false
249
+ }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7cdf25210daec74fabf4a911afd45fa0348bd706b292bb5de7940f626b4d82d9
3
  size 3055544304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c6171f0b9f700d11294fcc5d809c02ace38eac5de6bee81635a46574526dc75
3
  size 3055544304