File size: 2,750 Bytes
34c6018
ab82e42
 
 
 
 
979a537
ab82e42
 
 
34c6018
 
ab82e42
 
 
 
 
 
 
42da1d9
 
ab82e42
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
42da1d9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ab82e42
 
 
 
 
 
 
 
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
---
license: mit
tags:
- generated_from_trainer
metrics:
- wer
base_model: facebook/w2v-bert-2.0
model-index:
- name: w2v-bert-2.0-malayalam_mixeddataset_thre
  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. -->

# w2v-bert-2.0-malayalam_mixeddataset_thre

This model is a fine-tuned version of [facebook/w2v-bert-2.0](https://huggingface.co/facebook/w2v-bert-2.0) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1709
- Wer: 0.1197

## 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: 5e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 10
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Wer    |
|:-------------:|:-----:|:-----:|:---------------:|:------:|
| 1.1907        | 0.47  | 600   | 0.3890          | 0.4765 |
| 0.1663        | 0.95  | 1200  | 0.2528          | 0.3528 |
| 0.1207        | 1.42  | 1800  | 0.2176          | 0.2849 |
| 0.1017        | 1.9   | 2400  | 0.2021          | 0.2625 |
| 0.0833        | 2.37  | 3000  | 0.2032          | 0.2456 |
| 0.076         | 2.85  | 3600  | 0.1880          | 0.2376 |
| 0.0625        | 3.32  | 4200  | 0.1946          | 0.2247 |
| 0.0552        | 3.8   | 4800  | 0.1701          | 0.2247 |
| 0.0441        | 4.27  | 5400  | 0.1627          | 0.1759 |
| 0.0392        | 4.74  | 6000  | 0.1629          | 0.1829 |
| 0.0362        | 5.22  | 6600  | 0.1723          | 0.1605 |
| 0.0278        | 5.69  | 7200  | 0.1600          | 0.1665 |
| 0.0248        | 6.17  | 7800  | 0.1557          | 0.1446 |
| 0.0197        | 6.64  | 8400  | 0.1524          | 0.1505 |
| 0.0176        | 7.12  | 9000  | 0.1580          | 0.1339 |
| 0.0129        | 7.59  | 9600  | 0.1528          | 0.1411 |
| 0.0125        | 8.07  | 10200 | 0.1502          | 0.1299 |
| 0.0076        | 8.54  | 10800 | 0.1711          | 0.1189 |
| 0.0076        | 9.02  | 11400 | 0.1689          | 0.1237 |
| 0.0041        | 9.49  | 12000 | 0.1708          | 0.1227 |
| 0.0041        | 9.96  | 12600 | 0.1709          | 0.1197 |


### Framework versions

- Transformers 4.39.3
- Pytorch 2.1.1+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1