File size: 1,947 Bytes
8d87e6a 89cd82e 8d87e6a 89cd82e 8d87e6a 89cd82e 8d87e6a 89cd82e 8d87e6a 89cd82e 8d87e6a |
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 |
---
library_name: transformers
license: mit
base_model: microsoft/speecht5_tts
tags:
- generated_from_trainer
model-index:
- name: finetuned_speecht5_50Kdata
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. -->
# finetuned_speecht5_50Kdata
This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3766
## 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: 3e-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: 5000
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 0.5055 | 0.3505 | 500 | 0.4585 |
| 0.4679 | 0.7010 | 1000 | 0.4248 |
| 0.4444 | 1.0515 | 1500 | 0.4140 |
| 0.4356 | 1.4020 | 2000 | 0.4109 |
| 0.434 | 1.7525 | 2500 | 0.4082 |
| 0.428 | 2.1030 | 3000 | 0.4023 |
| 0.423 | 2.4536 | 3500 | 0.4003 |
| 0.4177 | 2.8041 | 4000 | 0.3835 |
| 0.4077 | 3.1546 | 4500 | 0.3766 |
| 0.4096 | 3.5051 | 5000 | 0.3781 |
### Framework versions
- Transformers 4.44.2
- Pytorch 2.5.0+cu121
- Datasets 3.0.2
- Tokenizers 0.19.1
|