End of training
Browse files
README.md
CHANGED
|
@@ -19,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 19 |
|
| 20 |
This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the TTS_English_data dataset.
|
| 21 |
It achieves the following results on the evaluation set:
|
| 22 |
-
- Loss: 0.
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
@@ -38,45 +38,30 @@ More information needed
|
|
| 38 |
### Training hyperparameters
|
| 39 |
|
| 40 |
The following hyperparameters were used during training:
|
| 41 |
-
- learning_rate:
|
| 42 |
-
- train_batch_size:
|
| 43 |
- eval_batch_size: 10
|
| 44 |
- seed: 42
|
| 45 |
-
- gradient_accumulation_steps:
|
| 46 |
-
- total_train_batch_size:
|
| 47 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
- lr_scheduler_type: linear
|
| 49 |
-
- lr_scheduler_warmup_steps:
|
| 50 |
-
- training_steps:
|
| 51 |
- mixed_precision_training: Native AMP
|
| 52 |
|
| 53 |
### Training results
|
| 54 |
|
| 55 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 56 |
|:-------------:|:-------:|:----:|:---------------:|
|
| 57 |
-
|
|
| 58 |
-
| 0.
|
| 59 |
-
| 0.
|
| 60 |
-
| 0.6633 | 4.0 | 704 | 0.4828 |
|
| 61 |
-
| 0.5635 | 5.0 | 880 | 0.4694 |
|
| 62 |
-
| 0.5364 | 6.0 | 1056 | 0.4601 |
|
| 63 |
-
| 0.5364 | 7.0 | 1232 | 0.4609 |
|
| 64 |
-
| 0.5155 | 8.0 | 1408 | 0.4463 |
|
| 65 |
-
| 0.5025 | 9.0 | 1584 | 0.4459 |
|
| 66 |
-
| 0.4883 | 10.0 | 1760 | 0.4399 |
|
| 67 |
-
| 0.4883 | 11.0 | 1936 | 0.4342 |
|
| 68 |
-
| 0.4792 | 12.0 | 2112 | 0.4359 |
|
| 69 |
-
| 0.4715 | 13.0 | 2288 | 0.4264 |
|
| 70 |
-
| 0.4715 | 14.0 | 2464 | 0.4273 |
|
| 71 |
-
| 0.4646 | 15.0 | 2640 | 0.4238 |
|
| 72 |
-
| 0.4598 | 16.0 | 2816 | 0.4231 |
|
| 73 |
-
| 0.4598 | 17.0 | 2992 | 0.4227 |
|
| 74 |
-
| 0.452 | 17.0455 | 3000 | 0.4244 |
|
| 75 |
|
| 76 |
|
| 77 |
### Framework versions
|
| 78 |
|
| 79 |
- Transformers 4.44.2
|
| 80 |
- Pytorch 2.5.0+cu121
|
| 81 |
-
- Datasets 3.0
|
| 82 |
- Tokenizers 0.19.1
|
|
|
|
| 19 |
|
| 20 |
This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the TTS_English_data dataset.
|
| 21 |
It achieves the following results on the evaluation set:
|
| 22 |
+
- Loss: 0.4859
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
|
|
| 38 |
### Training hyperparameters
|
| 39 |
|
| 40 |
The following hyperparameters were used during training:
|
| 41 |
+
- learning_rate: 2e-05
|
| 42 |
+
- train_batch_size: 14
|
| 43 |
- eval_batch_size: 10
|
| 44 |
- seed: 42
|
| 45 |
+
- gradient_accumulation_steps: 3
|
| 46 |
+
- total_train_batch_size: 42
|
| 47 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
- lr_scheduler_type: linear
|
| 49 |
+
- lr_scheduler_warmup_steps: 500
|
| 50 |
+
- training_steps: 1500
|
| 51 |
- mixed_precision_training: Native AMP
|
| 52 |
|
| 53 |
### Training results
|
| 54 |
|
| 55 |
| Training Loss | Epoch | Step | Validation Loss |
|
| 56 |
|:-------------:|:-------:|:----:|:---------------:|
|
| 57 |
+
| 0.5661 | 4.3353 | 500 | 0.5135 |
|
| 58 |
+
| 0.5339 | 8.6705 | 1000 | 0.4927 |
|
| 59 |
+
| 0.5155 | 13.0058 | 1500 | 0.4859 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 60 |
|
| 61 |
|
| 62 |
### Framework versions
|
| 63 |
|
| 64 |
- Transformers 4.44.2
|
| 65 |
- Pytorch 2.5.0+cu121
|
| 66 |
+
- Datasets 3.1.0
|
| 67 |
- Tokenizers 0.19.1
|