zibib commited on
Commit
5e3ff42
·
verified ·
1 Parent(s): 6656a55

Model save

Browse files
Files changed (2) hide show
  1. README.md +115 -9
  2. generation_config.json +2 -90
README.md CHANGED
@@ -1,12 +1,118 @@
1
  ---
2
  library_name: transformers
3
- datasets:
4
- - imvladikon/hebrew_speech_coursera
5
- language:
6
- - he
7
- metrics:
8
- - wer
9
- base_model:
10
- - imvladikon/hebrew_speech_coursera
11
- pipeline_tag: automatic-speech-recognition
12
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  library_name: transformers
3
+ base_model: ivrit-ai/whisper-large-v3-turbo
4
+ tags:
5
+ - generated_from_trainer
6
+ model-index:
7
+ - name: whisper-large-v3-turbo-ivrit-ai-coursera-fine-tuned
8
+ results: []
 
 
 
9
  ---
10
+
11
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
12
+ should probably proofread and complete it, then remove this comment. -->
13
+
14
+ # whisper-large-v3-turbo-ivrit-ai-coursera-fine-tuned
15
+
16
+ This model is a fine-tuned version of [ivrit-ai/whisper-large-v3-turbo](https://huggingface.co/ivrit-ai/whisper-large-v3-turbo) on an unknown dataset.
17
+ It achieves the following results on the evaluation set:
18
+ - Loss: 0.2829
19
+
20
+ ## Model description
21
+
22
+ More information needed
23
+
24
+ ## Intended uses & limitations
25
+
26
+ More information needed
27
+
28
+ ## Training and evaluation data
29
+
30
+ More information needed
31
+
32
+ ## Training procedure
33
+
34
+ ### Training hyperparameters
35
+
36
+ The following hyperparameters were used during training:
37
+ - learning_rate: 1e-05
38
+ - train_batch_size: 5
39
+ - eval_batch_size: 5
40
+ - seed: 42
41
+ - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
42
+ - lr_scheduler_type: constant_with_warmup
43
+ - lr_scheduler_warmup_ratio: 0.1
44
+ - lr_scheduler_warmup_steps: 500
45
+ - num_epochs: 10
46
+
47
+ ### Training results
48
+
49
+ | Training Loss | Epoch | Step | Validation Loss |
50
+ |:-------------:|:------:|:-----:|:---------------:|
51
+ | 0.1907 | 0.1641 | 500 | 0.2266 |
52
+ | 0.2283 | 0.3283 | 1000 | 0.2217 |
53
+ | 0.2253 | 0.4924 | 1500 | 0.2154 |
54
+ | 0.2257 | 0.6566 | 2000 | 0.2080 |
55
+ | 0.2138 | 0.8207 | 2500 | 0.2102 |
56
+ | 0.2153 | 0.9849 | 3000 | 0.2056 |
57
+ | 0.1615 | 1.1490 | 3500 | 0.2128 |
58
+ | 0.1588 | 1.3132 | 4000 | 0.1677 |
59
+ | 0.1628 | 1.4773 | 4500 | 0.1656 |
60
+ | 0.168 | 1.6415 | 5000 | 0.1798 |
61
+ | 0.167 | 1.8056 | 5500 | 0.1710 |
62
+ | 0.1663 | 1.9698 | 6000 | 0.1828 |
63
+ | 0.1297 | 2.1339 | 6500 | 0.1722 |
64
+ | 0.1196 | 2.2981 | 7000 | 0.1762 |
65
+ | 0.1336 | 2.4622 | 7500 | 0.1779 |
66
+ | 0.1258 | 2.6264 | 8000 | 0.1821 |
67
+ | 0.1275 | 2.7905 | 8500 | 0.1796 |
68
+ | 0.1331 | 2.9547 | 9000 | 0.1786 |
69
+ | 0.0988 | 3.1188 | 9500 | 0.1982 |
70
+ | 0.0933 | 3.2830 | 10000 | 0.1888 |
71
+ | 0.0963 | 3.4471 | 10500 | 0.1927 |
72
+ | 0.0946 | 3.6113 | 11000 | 0.1979 |
73
+ | 0.1018 | 3.7754 | 11500 | 0.2031 |
74
+ | 0.1027 | 3.9396 | 12000 | 0.1971 |
75
+ | 0.0795 | 4.1037 | 12500 | 0.2016 |
76
+ | 0.0698 | 4.2679 | 13000 | 0.2017 |
77
+ | 0.0736 | 4.4320 | 13500 | 0.2058 |
78
+ | 0.0747 | 4.5962 | 14000 | 0.2033 |
79
+ | 0.0768 | 4.7603 | 14500 | 0.2057 |
80
+ | 0.0801 | 4.9245 | 15000 | 0.2076 |
81
+ | 0.067 | 5.0886 | 15500 | 0.2196 |
82
+ | 0.0539 | 5.2528 | 16000 | 0.2185 |
83
+ | 0.0563 | 5.4169 | 16500 | 0.2220 |
84
+ | 0.0594 | 5.5811 | 17000 | 0.2265 |
85
+ | 0.0651 | 5.7452 | 17500 | 0.2176 |
86
+ | 0.0655 | 5.9094 | 18000 | 0.2227 |
87
+ | 0.0533 | 6.0735 | 18500 | 0.2387 |
88
+ | 0.0441 | 6.2377 | 19000 | 0.2334 |
89
+ | 0.0474 | 6.4018 | 19500 | 0.2343 |
90
+ | 0.0506 | 6.5660 | 20000 | 0.2387 |
91
+ | 0.0504 | 6.7301 | 20500 | 0.2373 |
92
+ | 0.0502 | 6.8943 | 21000 | 0.2318 |
93
+ | 0.0441 | 7.0584 | 21500 | 0.2524 |
94
+ | 0.0375 | 7.2226 | 22000 | 0.2533 |
95
+ | 0.0379 | 7.3867 | 22500 | 0.2491 |
96
+ | 0.0382 | 7.5509 | 23000 | 0.2635 |
97
+ | 0.0427 | 7.7150 | 23500 | 0.2506 |
98
+ | 0.0439 | 7.8792 | 24000 | 0.2430 |
99
+ | 0.043 | 8.0433 | 24500 | 0.2575 |
100
+ | 0.0296 | 8.2075 | 25000 | 0.2617 |
101
+ | 0.0309 | 8.3716 | 25500 | 0.2797 |
102
+ | 0.0366 | 8.5358 | 26000 | 0.2689 |
103
+ | 0.0351 | 8.6999 | 26500 | 0.2687 |
104
+ | 0.0384 | 8.8641 | 27000 | 0.2643 |
105
+ | 0.0365 | 9.0282 | 27500 | 0.2688 |
106
+ | 0.0265 | 9.1924 | 28000 | 0.2903 |
107
+ | 0.0299 | 9.3565 | 28500 | 0.2742 |
108
+ | 0.0347 | 9.5207 | 29000 | 0.2754 |
109
+ | 0.0311 | 9.6848 | 29500 | 0.2744 |
110
+ | 0.0345 | 9.8490 | 30000 | 0.2829 |
111
+
112
+
113
+ ### Framework versions
114
+
115
+ - Transformers 4.48.1
116
+ - Pytorch 2.6.0+cu124
117
+ - Datasets 3.4.1
118
+ - Tokenizers 0.21.1
generation_config.json CHANGED
@@ -25,6 +25,7 @@
25
  14
26
  ]
27
  ],
 
28
  "begin_suppress_tokens": [
29
  220,
30
  50257
@@ -151,96 +152,7 @@
151
  "pad_token_id": 50257,
152
  "prev_sot_token_id": 50362,
153
  "return_timestamps": false,
154
- "suppress_tokens": [
155
- 1,
156
- 2,
157
- 7,
158
- 8,
159
- 9,
160
- 10,
161
- 14,
162
- 25,
163
- 26,
164
- 27,
165
- 28,
166
- 29,
167
- 31,
168
- 58,
169
- 59,
170
- 60,
171
- 61,
172
- 62,
173
- 63,
174
- 90,
175
- 91,
176
- 92,
177
- 93,
178
- 359,
179
- 503,
180
- 522,
181
- 542,
182
- 873,
183
- 893,
184
- 902,
185
- 918,
186
- 922,
187
- 931,
188
- 1350,
189
- 1853,
190
- 1982,
191
- 2460,
192
- 2627,
193
- 3246,
194
- 3253,
195
- 3268,
196
- 3536,
197
- 3846,
198
- 3961,
199
- 4183,
200
- 4667,
201
- 6585,
202
- 6647,
203
- 7273,
204
- 9061,
205
- 9383,
206
- 10428,
207
- 10929,
208
- 11938,
209
- 12033,
210
- 12331,
211
- 12562,
212
- 13793,
213
- 14157,
214
- 14635,
215
- 15265,
216
- 15618,
217
- 16553,
218
- 16604,
219
- 18362,
220
- 18956,
221
- 20075,
222
- 21675,
223
- 22520,
224
- 26130,
225
- 26161,
226
- 26435,
227
- 28279,
228
- 29464,
229
- 31650,
230
- 32302,
231
- 32470,
232
- 36865,
233
- 42863,
234
- 47425,
235
- 49870,
236
- 50254,
237
- 50258,
238
- 50359,
239
- 50360,
240
- 50361,
241
- 50362,
242
- 50363
243
- ],
244
  "task_to_id": {
245
  "transcribe": 50360,
246
  "translate": 50359
 
25
  14
26
  ]
27
  ],
28
+ "attn_implementation": null,
29
  "begin_suppress_tokens": [
30
  220,
31
  50257
 
152
  "pad_token_id": 50257,
153
  "prev_sot_token_id": 50362,
154
  "return_timestamps": false,
155
+ "suppress_tokens": [],
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
156
  "task_to_id": {
157
  "transcribe": 50360,
158
  "translate": 50359