File size: 1,690 Bytes
8d0f6dd |
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 |
---
language:
- arr
license: mit
base_model: MBZUAI/speecht5_tts_clartts_ar
tags:
- generated_from_trainer
datasets:
- Arbi-Houssem/datasetSTT-TTS
model-index:
- name: SpeechT5 TTS Tunisien
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 Tunisien
This model is a fine-tuned version of [MBZUAI/speecht5_tts_clartts_ar](https://huggingface.co/MBZUAI/speecht5_tts_clartts_ar) on the datasetSTT-TTS dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5464
## 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: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 1000
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-------:|:----:|:---------------:|
| 0.544 | 14.2857 | 200 | 0.6401 |
| 0.4869 | 28.5714 | 400 | 0.5907 |
| 0.4384 | 42.8571 | 600 | 0.5684 |
| 0.4069 | 57.1429 | 800 | 0.5577 |
| 0.3992 | 71.4286 | 1000 | 0.5464 |
### Framework versions
- Transformers 4.42.0.dev0
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1
|