File size: 2,155 Bytes
6228a62
 
b4a3d6d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6228a62
 
b4a3d6d
 
6228a62
b4a3d6d
6228a62
b4a3d6d
 
 
 
6228a62
b4a3d6d
6228a62
b4a3d6d
6228a62
b4a3d6d
6228a62
b4a3d6d
6228a62
b4a3d6d
6228a62
b4a3d6d
6228a62
b4a3d6d
6228a62
b4a3d6d
6228a62
b4a3d6d
 
 
 
 
 
 
 
 
 
6228a62
b4a3d6d
6228a62
b4a3d6d
 
 
 
 
 
 
 
 
6228a62
 
b4a3d6d
6228a62
b4a3d6d
 
 
 
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
---
library_name: transformers
license: apache-2.0
base_model: facebook/wav2vec2-base
tags:
- generated_from_trainer
datasets:
- timit_asr
metrics:
- wer
model-index:
- name: wav2vec2-base-timit-demo-google-colab
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: timit_asr
      type: timit_asr
      config: clean
      split: None
      args: clean
    metrics:
    - name: Wer
      type: wer
      value: 0.3354696437185583
---

<!-- 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. -->

# wav2vec2-base-timit-demo-google-colab

This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co./facebook/wav2vec2-base) on the timit_asr dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4743
- Wer: 0.3355

## 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: 0.0001
- train_batch_size: 32
- 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
- num_epochs: 30
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer    |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 3.4394        | 4.0   | 500  | 1.2662          | 0.8530 |
| 0.5192        | 8.0   | 1000 | 0.4308          | 0.4176 |
| 0.1896        | 12.0  | 1500 | 0.4249          | 0.3656 |
| 0.1158        | 16.0  | 2000 | 0.4405          | 0.3583 |
| 0.0791        | 20.0  | 2500 | 0.4949          | 0.3481 |
| 0.0578        | 24.0  | 3000 | 0.4895          | 0.3448 |
| 0.0462        | 28.0  | 3500 | 0.4743          | 0.3355 |


### Framework versions

- Transformers 4.45.1
- Pytorch 2.1.2
- Datasets 3.0.1
- Tokenizers 0.20.0