File size: 2,088 Bytes
ed6093e
ee8ebcd
 
ed6093e
 
ee8ebcd
ed6093e
 
ee8ebcd
ed6093e
 
 
ee8ebcd
ed6093e
 
 
 
 
ee8ebcd
 
ed6093e
 
 
 
 
 
ee8ebcd
ed6093e
 
 
 
 
ee8ebcd
ed6093e
ee8ebcd
ed6093e
ee8ebcd
 
ed6093e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
---
language:
- or
license: apache-2.0
tags:
- whisper-event
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_11_0
metrics:
- wer
model-index:
- name: Whisper Small Odia
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: mozilla-foundation/common_voice_11_0 or
      type: mozilla-foundation/common_voice_11_0
      config: or
      split: test
      args: or
    metrics:
    - name: Wer
      type: wer
      value: 27.02397743300423
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# Whisper Small Odia

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the mozilla-foundation/common_voice_11_0 or dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4245
- Wer: 27.0240

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer     |
|:-------------:|:-----:|:----:|:---------------:|:-------:|
| 0.0021        | 49.0  | 1000 | 0.4245          | 27.0240 |
| 0.0001        | 99.0  | 2000 | 0.7338          | 28.1241 |
| 0.0           | 149.0 | 3000 | 0.8594          | 28.6601 |
| 0.0           | 199.0 | 4000 | 0.9103          | 28.3498 |
| 0.0           | 249.0 | 5000 | 0.9329          | 28.2934 |


### Framework versions

- Transformers 4.26.0.dev0
- Pytorch 1.13.0
- Datasets 2.7.1
- Tokenizers 0.13.2