File size: 2,831 Bytes
4a81027
 
 
 
 
c38ed4a
 
 
4a81027
 
 
 
 
 
 
 
 
 
 
 
 
c38ed4a
4a81027
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
library_name: transformers
license: cc-by-nc-4.0
base_model: facebook/mms-1b-all
tags:
- automatic-speech-recognition
- lozgen
- mms
- generated_from_trainer
metrics:
- wer
model-index:
- name: mms-1b-lozgen-male-model
  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. -->

# mms-1b-lozgen-male-model

This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the LOZGEN - LOZ dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3107
- Wer: 0.3264

## 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: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 30.0
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch   | Step | Validation Loss | Wer    |
|:-------------:|:-------:|:----:|:---------------:|:------:|
| 6.7665        | 0.7634  | 100  | 3.1610          | 0.9990 |
| 2.6476        | 1.5267  | 200  | 2.2410          | 0.8953 |
| 1.5694        | 2.2901  | 300  | 0.5585          | 0.7599 |
| 0.6761        | 3.0534  | 400  | 0.4566          | 0.6319 |
| 0.5793        | 3.8168  | 500  | 0.4013          | 0.4822 |
| 0.5392        | 4.5802  | 600  | 0.3795          | 0.4554 |
| 0.4809        | 5.3435  | 700  | 0.3730          | 0.4333 |
| 0.4813        | 6.1069  | 800  | 0.3597          | 0.4230 |
| 0.4484        | 6.8702  | 900  | 0.3432          | 0.3925 |
| 0.4418        | 7.6336  | 1000 | 0.3391          | 0.3947 |
| 0.4322        | 8.3969  | 1100 | 0.3339          | 0.3841 |
| 0.3963        | 9.1603  | 1200 | 0.3294          | 0.3669 |
| 0.4104        | 9.9237  | 1300 | 0.3217          | 0.3635 |
| 0.3777        | 10.6870 | 1400 | 0.3177          | 0.3610 |
| 0.3785        | 11.4504 | 1500 | 0.3236          | 0.3539 |
| 0.3682        | 12.2137 | 1600 | 0.3144          | 0.3468 |
| 0.3654        | 12.9771 | 1700 | 0.3122          | 0.3529 |
| 0.3509        | 13.7405 | 1800 | 0.3088          | 0.3463 |
| 0.3412        | 14.5038 | 1900 | 0.3146          | 0.3347 |
| 0.3344        | 15.2672 | 2000 | 0.3108          | 0.3416 |
| 0.3351        | 16.0305 | 2100 | 0.3107          | 0.3264 |


### Framework versions

- Transformers 4.48.0.dev0
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0