Hugging Face's logo

fsicoli
/
whisper-medium-pt-3000h-ct2

Generated from Trainer
Eval Results
Model card Files Community
fsicoli commited on
Commit
34f1250
·
verified ·
1 Parent(s): 0722f77

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +126 -0
README.md ADDED
@@ -0,0 +1,126 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: openai/whisper-medium
4
+ tags:
5
+ - generated_from_trainer
6
+ datasets:
7
+ - fsicoli/cv17-fleurs-coraa-mls-ted-alcaim-cf-cdc-lapsbm-lapsmail-sydney-lingualibre-voxforge-tatoeba
8
+ metrics:
9
+ - wer
10
+ model-index:
11
+ - name: whisper-medium-pt-3000h-ct2
12
+ results:
13
+ - task:
14
+ name: Automatic Speech Recognition
15
+ type: automatic-speech-recognition
16
+ dataset:
17
+ name: fsicoli/cv17-fleurs-coraa-mls-ted-alcaim-cf-cdc-lapsbm-lapsmail-sydney-lingualibre-voxforge-tatoeba
18
+ default
19
+ type: fsicoli/cv17-fleurs-coraa-mls-ted-alcaim-cf-cdc-lapsbm-lapsmail-sydney-lingualibre-voxforge-tatoeba
20
+ args: default
21
+ metrics:
22
+ - name: Wer
23
+ type: wer
24
+ value: 0.11007210455159983
25
+ ---
26
+
27
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
28
+ should probably proofread and complete it, then remove this comment. -->
29
+
30
+ # whisper-medium-pt-3000h-ct2
31
+
32
+ This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the fsicoli/cv17-fleurs-coraa-mls-ted-alcaim-cf-cdc-lapsbm-lapsmail-sydney-lingualibre-voxforge-tatoeba default dataset.
33
+ It was converted to the CTranslate2 format.
34
+ It achieves the following results on the evaluation set:
35
+ - Loss: 0.9306
36
+ - Wer: 0.1101
37
+
38
+ ## Model description
39
+
40
+ More information needed
41
+
42
+ ## Intended uses & limitations
43
+
44
+ More information needed
45
+
46
+ ## Training and evaluation data
47
+
48
+ More information needed
49
+
50
+ ## Training procedure
51
+
52
+ ### Training hyperparameters
53
+
54
+ The following hyperparameters were used during training:
55
+ - learning_rate: 5e-06
56
+ - train_batch_size: 8
57
+ - eval_batch_size: 8
58
+ - seed: 42
59
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
60
+ - lr_scheduler_type: linear
61
+ - lr_scheduler_warmup_steps: 10000
62
+ - num_epochs: 10.0
63
+ - mixed_precision_training: Native AMP
64
+
65
+ ### Training results
66
+
67
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
68
+ |:-------------:|:-----:|:-------:|:---------------:|:------:|
69
+ | 0.4423 | 0.2 | 20000 | 0.4723 | 0.1633 |
70
+ | 0.4963 | 0.39 | 40000 | 0.4921 | 0.1547 |
71
+ | 0.3853 | 0.59 | 60000 | 0.5099 | 0.1470 |
72
+ | 0.37 | 0.79 | 80000 | 0.4753 | 0.1439 |
73
+ | 0.3615 | 0.98 | 100000 | 0.5074 | 0.1386 |
74
+ | 0.2394 | 1.18 | 120000 | 0.4858 | 0.1341 |
75
+ | 0.227 | 1.38 | 140000 | 0.5758 | 0.1323 |
76
+ | 0.2461 | 1.57 | 160000 | 0.5067 | 0.1322 |
77
+ | 0.2078 | 1.77 | 180000 | 0.5087 | 0.1291 |
78
+ | 0.2138 | 1.97 | 200000 | 0.5201 | 0.1273 |
79
+ | 0.1188 | 2.16 | 220000 | 0.6359 | 0.1265 |
80
+ | 0.1009 | 2.36 | 240000 | 0.6229 | 0.1253 |
81
+ | 0.1394 | 2.56 | 260000 | 0.5734 | 0.1231 |
82
+ | 0.1383 | 2.75 | 280000 | 0.5914 | 0.1213 |
83
+ | 0.1332 | 2.95 | 300000 | 0.6174 | 0.1212 |
84
+ | 0.0634 | 3.15 | 320000 | 0.6461 | 0.1190 |
85
+ | 0.0667 | 3.34 | 340000 | 0.6330 | 0.1211 |
86
+ | 0.0546 | 3.54 | 360000 | 0.6927 | 0.1190 |
87
+ | 0.1029 | 3.74 | 380000 | 0.6777 | 0.1184 |
88
+ | 0.0664 | 3.93 | 400000 | 0.6367 | 0.1161 |
89
+ | 0.0665 | 4.13 | 420000 | 0.7467 | 0.1171 |
90
+ | 0.0695 | 4.33 | 440000 | 0.7332 | 0.1164 |
91
+ | 0.0708 | 4.52 | 460000 | 0.7141 | 0.1171 |
92
+ | 0.0695 | 4.72 | 480000 | 0.6869 | 0.1169 |
93
+ | 0.0758 | 4.92 | 500000 | 0.7360 | 0.1153 |
94
+ | 0.061 | 5.11 | 520000 | 0.7594 | 0.1161 |
95
+ | 0.0804 | 5.31 | 540000 | 0.7640 | 0.1158 |
96
+ | 0.0963 | 5.51 | 560000 | 0.7848 | 0.1157 |
97
+ | 0.0815 | 5.7 | 580000 | 0.7635 | 0.1145 |
98
+ | 0.0794 | 5.9 | 600000 | 0.7566 | 0.1134 |
99
+ | 0.0907 | 6.1 | 620000 | 0.8152 | 0.1147 |
100
+ | 0.0664 | 6.29 | 640000 | 0.8405 | 0.1123 |
101
+ | 0.0654 | 6.49 | 660000 | 0.8278 | 0.1119 |
102
+ | 0.0652 | 6.69 | 680000 | 0.8267 | 0.1134 |
103
+ | 0.1043 | 6.88 | 700000 | 0.8254 | 0.1122 |
104
+ | 0.0383 | 7.08 | 720000 | 0.8719 | 0.1122 |
105
+ | 0.0461 | 7.28 | 740000 | 0.8640 | 0.1130 |
106
+ | 0.0791 | 7.47 | 760000 | 0.8990 | 0.1122 |
107
+ | 0.0587 | 7.67 | 780000 | 0.9107 | 0.1122 |
108
+ | 0.0578 | 7.87 | 800000 | 0.9060 | 0.1124 |
109
+ | 0.0218 | 8.06 | 820000 | 0.8845 | 0.1111 |
110
+ | 0.0125 | 8.26 | 840000 | 0.9072 | 0.1112 |
111
+ | 0.0172 | 8.46 | 860000 | 0.8899 | 0.1107 |
112
+ | 0.0204 | 8.65 | 880000 | 0.9149 | 0.1108 |
113
+ | 0.0145 | 8.85 | 900000 | 0.9097 | 0.1103 |
114
+ | 0.0146 | 9.05 | 920000 | 0.9084 | 0.1107 |
115
+ | 0.0166 | 9.24 | 940000 | 0.9053 | 0.1103 |
116
+ | 0.0177 | 9.44 | 960000 | 0.9193 | 0.1100 |
117
+ | 0.0157 | 9.64 | 980000 | 0.9212 | 0.1101 |
118
+ | 0.0096 | 9.83 | 1000000 | 0.9313 | 0.1103 |
119
+
120
+
121
+ ### Framework versions
122
+
123
+ - Transformers 4.39.0.dev0
124
+ - Pytorch 2.2.1+cu121
125
+ - Datasets 2.18.1.dev0
126
+ - Tokenizers 0.15.0