File size: 2,138 Bytes
767b2ae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
language:
- ta
license: apache-2.0
tags:
- whisper-event
metrics:
- wer
model-index:
- name: Whisper Tamil Small - Vasista Sai Lodagala
  results:
  - task:
      type: automatic-speech-recognition
      name: Automatic Speech Recognition
    dataset:
      name: google/fleurs
      type: google/fleurs
      config: ta_in
      split: test
    metrics:
    - type: wer
      value: 10.44
      name: WER
  - task:
      type: automatic-speech-recognition
      name: Automatic Speech Recognition
    dataset:
      name: mozilla-foundation/common_voice_11_0
      type: mozilla-foundation/common_voice_11_0
      config: ta
      split: test
    metrics:
    - type: wer
      value: 8.75
      name: WER
---

<!-- 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 Tamil Small

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Tamil data available from multiple publicly available ASR corpuses.
It has been fine-tuned as a part of the Whisper fine-tuning sprint.

## Training and evaluation data at Speech Lab, IITM

Training Data: MILE ASR Corpus, ULCA ASR Corpus, Shrutilipi ASR Corpus, Microsoft Research Tamil Corpus (Train+Dev), Babel ASR Corpus, Google/Fleurs (Train+Dev) set.

Evaluation Data: MILE ASR Corpus Test, Babel Test, Microsoft Research Tamil Corpus Test, Google/Fleurs Test set.

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 1.7e-05
- train_batch_size: 48
- eval_batch_size: 32
- seed: 22
- optimizer: adamw_bnb_8bit
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 17500
- training_steps: 29659 (Initially set to 84740 steps)
- mixed_precision_training: True

## Acknowledgement
This work was done at Speech Lab, IITM. The compute resources for this work were funded by "Bhashini: National Language translation Mission" project of the Ministry of Electronics and Information Technology (MeitY), Government of India.