File size: 2,515 Bytes
3cc38a5
 
 
e294886
3cc38a5
 
e294886
3cc38a5
 
 
e294886
3cc38a5
 
 
 
 
e294886
 
3cc38a5
 
 
 
 
 
674d21a
3cc38a5
 
 
 
 
e294886
3cc38a5
e294886
3cc38a5
674d21a
 
3cc38a5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b914315
3cc38a5
 
 
 
 
 
 
674d21a
 
3cc38a5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
674d21a
 
3cc38a5
 
 
 
 
 
 
 
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
---
license: apache-2.0
tags:
- whisper-event
- generated_from_trainer
datasets:
- google/fleurs
metrics:
- wer
model-index:
- name: Whisper Medium Pashto
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: google/fleurs ps_af
      type: google/fleurs
      config: ps_af
      split: test
      args: ps_af
    metrics:
    - name: Wer
      type: wer
      value: 50.544794188861985
---

<!-- 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 Medium Pashto

This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the google/fleurs ps_af dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4807
- Wer: 50.5448

## 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: 3e-07
- train_batch_size: 32
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 10
- training_steps: 1200
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Wer     |
|:-------------:|:------:|:----:|:---------------:|:-------:|
| 0.0334        | 14.29  | 100  | 1.0348          | 50.0908 |
| 0.0021        | 28.57  | 200  | 1.1971          | 49.4855 |
| 0.0007        | 42.86  | 300  | 1.2651          | 49.7352 |
| 0.0006        | 57.14  | 400  | 1.3084          | 49.9697 |
| 0.0005        | 71.43  | 500  | 1.3479          | 50.0605 |
| 0.0004        | 85.71  | 600  | 1.3835          | 50.3027 |
| 0.0004        | 100.0  | 700  | 1.4139          | 50.4540 |
| 0.0004        | 114.29 | 800  | 1.4382          | 50.4616 |
| 0.0004        | 128.57 | 900  | 1.4545          | 50.5297 |
| 0.0003        | 142.86 | 1000 | 1.4603          | 50.5675 |
| 0.0003        | 157.14 | 1100 | 1.4750          | 50.5599 |
| 0.0003        | 171.43 | 1200 | 1.4807          | 50.5448 |


### Framework versions

- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu117
- Datasets 2.7.1.dev0
- Tokenizers 0.13.2