File size: 1,800 Bytes
a4c9140
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
tags:
- generated_from_trainer
model-index:
- name: ''
  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. -->

# 

This model was trained from scratch on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5575
- Wer: 0.5219

## 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: 0.0003
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- 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: 400
- num_epochs: 5.0
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer    |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 2.3036        | 0.51  | 400  | 4.2199          | 0.9500 |
| 2.836         | 1.02  | 800  | 2.7748          | 1.2758 |
| 2.7021        | 1.53  | 1200 | 2.6211          | 1.1725 |
| 2.2322        | 2.04  | 1600 | 2.5714          | 1.1042 |
| 2.1435        | 2.55  | 2000 | 2.4205          | 1.0093 |
| 1.1205        | 3.06  | 2400 | 2.7612          | 0.9098 |
| 0.7821        | 3.57  | 2800 | 2.2900          | 0.8478 |
| 1.0018        | 4.08  | 3200 | 0.7970          | 0.6275 |
| 0.6312        | 4.59  | 3600 | 0.5575          | 0.5219 |


### Framework versions

- Transformers 4.17.0.dev0
- Pytorch 1.10.0+cu113
- Datasets 1.18.1
- Tokenizers 0.11.0