File size: 3,205 Bytes
ff914a0
 
941974a
 
 
 
 
 
 
28e0128
2ba89aa
 
b454968
2ba89aa
 
 
 
b454968
f061c9b
 
2ba89aa
 
 
 
 
 
 
 
 
 
fbff6c0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
28e0128
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2ba89aa
ff914a0
 
 
 
 
c63748d
ff914a0
5020ac4
941974a
bfb0ec2
10af7a3
 
ff914a0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d26b7c8
10af7a3
ff914a0
 
 
 
 
 
10af7a3
 
 
 
 
 
ff914a0
 
 
 
 
 
 
941974a
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
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
---
license: apache-2.0
datasets:
- thennal/IMaSC
- vrclc/openslr63
- vrclc/festvox-iiith-ml
- smcproject/MSC
language:
- ml
- en
base_model: openai/whisper-medium
model-index:
- name: vrclc/Whisper-med-ml - Bajiyo Baiju
  results:
  - task:
      type: automatic-speech-recognition
      name: Automatic Speech Recognition
    dataset: 
      name: Common Voice 13 Malayalam
      type: mozilla-foundation/common_voice_13_0
      config: ml
      split: test
      args: ml
    metrics:
    - type: wer
      value: 63.64
      name: WER
    - type: cer
      value: 13.61
      name: CER
  - task:
      type: automatic-speech-recognition
      name: Automatic Speech Recognition
    dataset: 
      name: Common Voice 16 Malayalam
      type: mozilla-foundation/common_voice_16_1
      config: ml
      split: test
      args: ml
    metrics:
    - type: wer
      value: 64.63
      name: WER
    - type: cer
      value: 14.07
      name: CER
  - task:
      type: automatic-speech-recognition
      name: Automatic Speech Recognition
    dataset:
      name: OpenSLR Malayalam -Test
      type: vrclc/openslr63
      config: ml
      split: test
      args: ml
    metrics:
    - type: wer
      value: 14.65
      name: WER
    - type: cer
      value: 2.59
      name: CER
library_name: transformers
---

<!-- 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-med-ml

This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the datasets: [IMASC](https://huggingface.co/datasets/thennal/IMaSC), [MSC](https://huggingface.co/datasets/smcproject/MSC), [OpenSLR Malayalam Train split](https://huggingface.co/datasets/vrclc/openslr63), [Festvox Malayalam](https://huggingface.co/datasets/vrclc/openslr63)  .

It achieves the following results on the validation set : [OpenSLR-Test](https://huggingface.co/vrclc/openslr63)
- Loss: 0.0318
- Wer: 14.7300

## 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: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- training_steps: 6000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer     |
|:-------------:|:-----:|:----:|:---------------:|:-------:|
| 0.0599        | 0.4   | 1000 | 0.0910          | 42.4981 |
| 0.0341        | 0.79  | 2000 | 0.0584          | 30.0572 |
| 0.0183        | 1.19  | 3000 | 0.0439          | 23.1650 |
| 0.0147        | 1.58  | 4000 | 0.0363          | 18.7360 |
| 0.0107        | 1.98  | 5000 | 0.0322          | 16.4220 |
| 0.0032        | 2.37  | 6000 | 0.0318          | 14.7300 |


### Framework versions

- Transformers 4.37.2
- Pytorch 2.1.1+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1