Hugging Face's logo

FelixK7
/
whisper-medium-lv

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Latvian
whisper
hf-asr-leaderboard
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Metrics Community
FelixK7 commited on
Commit
95d8e16
•
1 Parent(s): e2304e3

End of training

Browse files
Files changed (2) hide show
  1. README.md +62 -37
  2. model.safetensors +1 -1
README.md CHANGED
@@ -12,7 +12,7 @@ datasets:
12
  metrics:
13
  - wer
14
  model-index:
15
- - name: Whisper medium LV - Felikss Kleins
16
  results:
17
  - task:
18
  name: Automatic Speech Recognition
@@ -26,18 +26,18 @@ model-index:
26
  metrics:
27
  - name: Wer
28
  type: wer
29
- value: 13.506122934252765
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 [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the Common Voice 17.0 dataset.
38
  It achieves the following results on the evaluation set:
39
- - Loss: 0.1848
40
- - Wer: 13.5061
41
 
42
  ## Model description
43
 
@@ -56,47 +56,72 @@ More information needed
56
  ### Training hyperparameters
57
 
58
  The following hyperparameters were used during training:
59
- - learning_rate: 5e-06
60
- - train_batch_size: 6
61
- - eval_batch_size: 16
62
  - seed: 42
63
  - gradient_accumulation_steps: 2
64
- - total_train_batch_size: 12
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: 5000
69
  - mixed_precision_training: Native AMP
70
 
71
  ### Training results
72
 
73
- | Training Loss | Epoch | Step | Validation Loss | Wer |
74
- |:-------------:|:------:|:----:|:---------------:|:-------:|
75
- | No log | 0.04 | 200 | 0.8471 | 30.5453 |
76
- | 1.8421 | 0.08 | 400 | 0.3571 | 27.1906 |
77
- | 0.5851 | 0.12 | 600 | 0.3146 | 23.9508 |
78
- | 0.4787 | 0.16 | 800 | 0.2913 | 23.2454 |
79
- | 0.4845 | 0.2 | 1000 | 0.2736 | 20.9388 |
80
- | 0.4845 | 0.24 | 1200 | 0.2495 | 19.1951 |
81
- | 0.4004 | 0.28 | 1400 | 0.2416 | 18.4600 |
82
- | 0.357 | 1.018 | 1600 | 0.2300 | 17.8576 |
83
- | 0.3381 | 1.058 | 1800 | 0.2261 | 17.1125 |
84
- | 0.3052 | 1.098 | 2000 | 0.2151 | 16.6013 |
85
- | 0.3052 | 1.138 | 2200 | 0.2154 | 16.1673 |
86
- | 0.2636 | 1.178 | 2400 | 0.2256 | 17.3107 |
87
- | 0.2805 | 1.218 | 2600 | 0.2059 | 15.6482 |
88
- | 0.2331 | 1.258 | 2800 | 0.2022 | 15.4599 |
89
- | 0.2245 | 1.298 | 3000 | 0.1971 | 15.0953 |
90
- | 0.2245 | 2.036 | 3200 | 0.1988 | 14.7604 |
91
- | 0.2312 | 2.076 | 3400 | 0.1941 | 14.5623 |
92
- | 0.2077 | 2.116 | 3600 | 0.1915 | 14.2175 |
93
- | 0.1923 | 2.156 | 3800 | 0.1935 | 14.4018 |
94
- | 0.2012 | 2.196 | 4000 | 0.1906 | 14.4850 |
95
- | 0.2012 | 2.2360 | 4200 | 0.1880 | 13.9302 |
96
- | 0.1694 | 2.276 | 4400 | 0.1860 | 14.1640 |
97
- | 0.1548 | 3.014 | 4600 | 0.1855 | 13.6508 |
98
- | 0.1628 | 3.054 | 4800 | 0.1850 | 13.5557 |
99
- | 0.1529 | 3.094 | 5000 | 0.1848 | 13.5061 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
100
 
101
 
102
  ### Framework versions
 
12
  metrics:
13
  - wer
14
  model-index:
15
+ - name: Whisper medium LV - Felikss Kleins-{timestamp}
16
  results:
17
  - task:
18
  name: Automatic Speech Recognition
 
26
  metrics:
27
  - name: Wer
28
  type: wer
29
+ value: 10.601196845400864
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-{timestamp}
36
 
37
  This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the Common Voice 17.0 dataset.
38
  It achieves the following results on the evaluation set:
39
+ - Loss: 0.2428
40
+ - Wer: 10.6012
41
 
42
  ## Model description
43
 
 
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 | 0.2930 | 22.3576 |
76
+ | 0.9797 | 1.0116 | 400 | 0.2359 | 18.2083 |
77
+ | 0.357 | 2.0033 | 600 | 0.2274 | 16.4665 |
78
+ | 0.2582 | 2.0233 | 800 | 0.2111 | 15.6402 |
79
+ | 0.1718 | 3.0149 | 1000 | 0.2135 | 14.9883 |
80
+ | 0.1718 | 4.0066 | 1200 | 0.2090 | 14.2294 |
81
+ | 0.1355 | 4.0266 | 1400 | 0.2193 | 13.5537 |
82
+ | 0.1024 | 5.0183 | 1600 | 0.2255 | 14.5048 |
83
+ | 0.0836 | 6.0099 | 1800 | 0.2145 | 12.9751 |
84
+ | 0.0699 | 7.0015 | 2000 | 0.2232 | 13.2129 |
85
+ | 0.0699 | 7.0216 | 2200 | 0.2181 | 12.7155 |
86
+ | 0.0598 | 8.0132 | 2400 | 0.2192 | 12.7076 |
87
+ | 0.054 | 9.0048 | 2600 | 0.2348 | 13.0048 |
88
+ | 0.0452 | 9.0249 | 2800 | 0.2241 | 13.0940 |
89
+ | 0.0433 | 10.0165 | 3000 | 0.2406 | 12.6362 |
90
+ | 0.0433 | 11.0082 | 3200 | 0.2283 | 12.5332 |
91
+ | 0.0377 | 11.0282 | 3400 | 0.2293 | 12.2201 |
92
+ | 0.0317 | 12.0198 | 3600 | 0.2323 | 12.6144 |
93
+ | 0.0297 | 13.0114 | 3800 | 0.2309 | 12.2974 |
94
+ | 0.0267 | 14.0031 | 4000 | 0.2342 | 11.9011 |
95
+ | 0.0267 | 14.0231 | 4200 | 0.2286 | 12.1171 |
96
+ | 0.0243 | 15.0147 | 4400 | 0.2364 | 12.0854 |
97
+ | 0.0218 | 16.0064 | 4600 | 0.2405 | 12.1805 |
98
+ | 0.021 | 16.0264 | 4800 | 0.2422 | 12.0338 |
99
+ | 0.0173 | 17.0180 | 5000 | 0.2416 | 11.9387 |
100
+ | 0.0173 | 18.0097 | 5200 | 0.2421 | 12.0180 |
101
+ | 0.0175 | 19.0013 | 5400 | 0.2385 | 11.6613 |
102
+ | 0.0161 | 19.0214 | 5600 | 0.2442 | 11.9090 |
103
+ | 0.0136 | 20.013 | 5800 | 0.2411 | 11.4513 |
104
+ | 0.0135 | 21.0047 | 6000 | 0.2470 | 12.0418 |
105
+ | 0.0135 | 21.0247 | 6200 | 0.2446 | 11.5246 |
106
+ | 0.0117 | 22.0163 | 6400 | 0.2466 | 11.7386 |
107
+ | 0.0111 | 23.0080 | 6600 | 0.2498 | 12.0715 |
108
+ | 0.01 | 23.0280 | 6800 | 0.2487 | 12.0596 |
109
+ | 0.0094 | 24.0196 | 7000 | 0.2431 | 11.4315 |
110
+ | 0.0094 | 25.0113 | 7200 | 0.2460 | 11.5702 |
111
+ | 0.009 | 26.0029 | 7400 | 0.2436 | 11.2293 |
112
+ | 0.0077 | 26.0229 | 7600 | 0.2467 | 11.3423 |
113
+ | 0.007 | 27.0145 | 7800 | 0.2439 | 11.0054 |
114
+ | 0.0071 | 28.0062 | 8000 | 0.2430 | 11.1996 |
115
+ | 0.0071 | 28.0262 | 8200 | 0.2458 | 11.1798 |
116
+ | 0.0063 | 29.0178 | 8400 | 0.2456 | 11.0847 |
117
+ | 0.0049 | 30.0095 | 8600 | 0.2450 | 11.1303 |
118
+ | 0.0049 | 31.0011 | 8800 | 0.2467 | 11.0530 |
119
+ | 0.0053 | 31.0211 | 9000 | 0.2448 | 11.1085 |
120
+ | 0.0053 | 32.0128 | 9200 | 0.2467 | 11.2650 |
121
+ | 0.0041 | 33.0044 | 9400 | 0.2444 | 11.0728 |
122
+ | 0.0045 | 33.0245 | 9600 | 0.2426 | 10.6547 |
123
+ | 0.0036 | 34.0161 | 9800 | 0.2427 | 10.5972 |
124
+ | 0.004 | 35.0078 | 10000 | 0.2428 | 10.6012 |
125
 
126
 
127
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:92f985c3352b96d087a48db02cab59ff54319e4a1b20c4530453b92323471da7
3
  size 3055544304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2db2fbe0b23a85b7fd24f1cc479b0e02c6d6eaa86424d262ab0edd392dc5607a
3
  size 3055544304