File size: 1,657 Bytes
e267fb6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8ecc78a
e267fb6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8ecc78a
 
 
 
e267fb6
 
 
 
 
 
 
 
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
---
language:
- tr
license: mit
base_model: microsoft/speecht5_tts
tags:
- text-to-speech
- turkish
- speech-generation
- generated_from_trainer
datasets:
- common_voice
model-index:
- name: SpeechT5 TTS Turkish
  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. -->

# SpeechT5 TTS Turkish

This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the CommonVoice dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4650

## 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: 1e-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
- training_steps: 4000

### Training results

| Training Loss | Epoch  | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 0.4726        | 111.11 | 1000 | 0.4570          |
| 0.4333        | 222.22 | 2000 | 0.4567          |
| 0.417         | 333.33 | 3000 | 0.4631          |
| 0.4327        | 444.44 | 4000 | 0.4650          |


### Framework versions

- Transformers 4.33.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.14.4
- Tokenizers 0.13.3