p4b commited on
Commit
2d49c2f
1 Parent(s): 61e64aa

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +83 -0
README.md ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - ko
4
+ license: apache-2.0
5
+ tags:
6
+ - whisper-event
7
+ - generated_from_trainer
8
+ datasets:
9
+ - kr_dialect_speech
10
+ metrics:
11
+ - wer
12
+ model-index:
13
+ - name: Whisper Small Ko(Gyungsang dialect) - p4b
14
+ results:
15
+ - task:
16
+ name: Automatic Speech Recognition
17
+ type: automatic-speech-recognition
18
+ dataset:
19
+ name: KR Dialect Speech - gyungsang
20
+ type: kr_dialect_speech
21
+ config: gyungsang
22
+ split: validation
23
+ args: gyungsang
24
+ metrics:
25
+ - name: Wer
26
+ type: wer
27
+ value: 15.930018416206263
28
+ ---
29
+
30
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
31
+ should probably proofread and complete it, then remove this comment. -->
32
+
33
+ # Whisper Small Ko(Gyungsang dialect) - p4b
34
+
35
+ This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the KR Dialect Speech - gyungsang dataset.
36
+ It achieves the following results on the evaluation set:
37
+ - Loss: 0.2017
38
+ - Wer: 15.9300
39
+
40
+ ## Model description
41
+
42
+ More information needed
43
+
44
+ ## Intended uses & limitations
45
+
46
+ More information needed
47
+
48
+ ## Training and evaluation data
49
+
50
+ More information needed
51
+
52
+ ## Training procedure
53
+
54
+ ### Training hyperparameters
55
+
56
+ The following hyperparameters were used during training:
57
+ - learning_rate: 5e-07
58
+ - train_batch_size: 96
59
+ - eval_batch_size: 64
60
+ - seed: 42
61
+ - distributed_type: multi-GPU
62
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
63
+ - lr_scheduler_type: cosine
64
+ - lr_scheduler_warmup_steps: 500
65
+ - training_steps: 5000
66
+
67
+ ### Training results
68
+
69
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
70
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|
71
+ | 0.5909 | 0.2 | 1000 | 0.4133 | 211.6022 |
72
+ | 0.3612 | 0.4 | 2000 | 0.2137 | 16.9429 |
73
+ | 0.5373 | 0.6 | 3000 | 0.2063 | 15.8379 |
74
+ | 0.2909 | 0.8 | 4000 | 0.2012 | 15.8379 |
75
+ | 0.3317 | 1.0 | 5000 | 0.2017 | 15.9300 |
76
+
77
+
78
+ ### Framework versions
79
+
80
+ - Transformers 4.26.0.dev0
81
+ - Pytorch 1.14.0.dev20221208+cu116
82
+ - Datasets 2.7.1.dev0
83
+ - Tokenizers 0.13.2