File size: 1,828 Bytes
ae36ecf
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: mit
base_model: microsoft/speecht5_tts
tags:
- generated_from_trainer
model-index:
- name: fil_b64_le4_s4000
  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. -->

# fil_b64_le4_s4000

This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4134

## 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: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- 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: 2000
- training_steps: 4000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch    | Step | Validation Loss |
|:-------------:|:--------:|:----:|:---------------:|
| 0.4725        | 22.2222  | 500  | 0.4372          |
| 0.4415        | 44.4444  | 1000 | 0.4203          |
| 0.423         | 66.6667  | 1500 | 0.4169          |
| 0.4108        | 88.8889  | 2000 | 0.4183          |
| 0.3934        | 111.1111 | 2500 | 0.4111          |
| 0.3821        | 133.3333 | 3000 | 0.4164          |
| 0.3743        | 155.5556 | 3500 | 0.4127          |
| 0.3714        | 177.7778 | 4000 | 0.4134          |


### Framework versions

- Transformers 4.41.0.dev0
- Pytorch 2.2.1+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1