File size: 1,908 Bytes
c8beb8e
b1a09d0
 
c8beb8e
 
 
b1a09d0
c8beb8e
 
 
 
b1a09d0
c8beb8e
 
 
 
 
 
b1a09d0
c8beb8e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
language:
- it
license: apache-2.0
base_model: openai/whisper-small
tags:
- hf-asr-leaderboard
- generated_from_trainer
metrics:
- wer
model-index:
- name: Whisper Small IT
  results: []
---

<!-- 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 IT

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5372
- Wer: 130.3266

## 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: 500
- training_steps: 4000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer      |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.1972        | 1.6   | 500  | 0.4266          | 105.1927 |
| 0.033         | 3.19  | 1000 | 0.4507          | 209.0820 |
| 0.0167        | 4.79  | 1500 | 0.4738          | 96.0643  |
| 0.0052        | 6.39  | 2000 | 0.4948          | 143.7616 |
| 0.0035        | 7.99  | 2500 | 0.5144          | 126.8133 |
| 0.0047        | 9.58  | 3000 | 0.5273          | 133.8966 |
| 0.0033        | 11.18 | 3500 | 0.5349          | 137.8580 |
| 0.0026        | 12.78 | 4000 | 0.5372          | 130.3266 |


### Framework versions

- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0