Model save
Browse files
README.md
CHANGED
|
@@ -17,6 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 17 |
# speecht5_tts_ta
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the common_voice_17_0 dataset.
|
|
|
|
|
|
|
| 20 |
|
| 21 |
## Model description
|
| 22 |
|
|
@@ -46,6 +48,16 @@ The following hyperparameters were used during training:
|
|
| 46 |
- training_steps: 4000
|
| 47 |
- mixed_precision_training: Native AMP
|
| 48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
### Framework versions
|
| 50 |
|
| 51 |
- Transformers 4.49.0.dev0
|
|
|
|
| 17 |
# speecht5_tts_ta
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the common_voice_17_0 dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.5390
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
|
|
| 48 |
- training_steps: 4000
|
| 49 |
- mixed_precision_training: Native AMP
|
| 50 |
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 54 |
+
|:-------------:|:------:|:----:|:---------------:|
|
| 55 |
+
| 0.357 | 500.0 | 1000 | 0.5013 |
|
| 56 |
+
| 0.324 | 1000.0 | 2000 | 0.4800 |
|
| 57 |
+
| 0.2914 | 1500.0 | 3000 | 0.4997 |
|
| 58 |
+
| 0.283 | 2000.0 | 4000 | 0.5390 |
|
| 59 |
+
|
| 60 |
+
|
| 61 |
### Framework versions
|
| 62 |
|
| 63 |
- Transformers 4.49.0.dev0
|