Hugging Face's logo

Edmon02
/
speecht5_finetuned_hy

Text-to-Speech
Transformers
TensorBoard
Safetensors
Armenian
English
Dutch
speecht5
text-to-audio
Generated from Trainer
Model card Files Metrics Community
Edmon02 commited on
Commit
49a49f8
•
1 Parent(s): 69c48e9

End of training

Browse files
Files changed (1) hide show
  1. README.md +11 -20
README.md CHANGED
@@ -1,6 +1,6 @@
1
  ---
2
  license: mit
3
- base_model: microsoft/speecht5_tts
4
  tags:
5
  - generated_from_trainer
6
  model-index:
@@ -13,9 +13,9 @@ should probably proofread and complete it, then remove this comment. -->
13
 
14
  # speecht5_finetuned_voxpopuli_hy
15
 
16
- This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the None dataset.
17
  It achieves the following results on the evaluation set:
18
- - Loss: 0.6020
19
 
20
  ## Model description
21
 
@@ -42,31 +42,22 @@ The following hyperparameters were used during training:
42
  - total_train_batch_size: 32
43
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
44
  - lr_scheduler_type: linear
45
- - lr_scheduler_warmup_steps: 62
46
- - training_steps: 1500
47
- - mixed_precision_training: Native AMP
48
 
49
  ### Training results
50
 
51
  | Training Loss | Epoch | Step | Validation Loss |
52
  |:-------------:|:------:|:----:|:---------------:|
53
- | 0.3261 | 83.0 | 125 | 0.7073 |
54
- | 0.3063 | 166.67 | 250 | 0.6427 |
55
- | 0.2848 | 250.0 | 375 | 0.6256 |
56
- | 0.2857 | 333.33 | 500 | 0.6221 |
57
- | 0.2838 | 417.0 | 625 | 0.6205 |
58
- | 0.2526 | 500.0 | 750 | 0.6013 |
59
- | 0.2429 | 583.0 | 875 | 0.5997 |
60
- | 0.2666 | 666.67 | 1000 | 0.6097 |
61
- | 0.2531 | 750.0 | 1125 | 0.6119 |
62
- | 0.336 | 833.33 | 1250 | 0.6109 |
63
- | 0.2299 | 917.0 | 1375 | 0.6049 |
64
- | 0.2501 | 1000.0 | 1500 | 0.6020 |
65
 
66
 
67
  ### Framework versions
68
 
69
  - Transformers 4.35.2
70
- - Pytorch 2.1.0+cu118
71
- - Datasets 2.15.0
72
  - Tokenizers 0.15.0
 
1
  ---
2
  license: mit
3
+ base_model: Edmon02/speecht5_finetuned_voxpopuli_hy
4
  tags:
5
  - generated_from_trainer
6
  model-index:
 
13
 
14
  # speecht5_finetuned_voxpopuli_hy
15
 
16
+ This model is a fine-tuned version of [Edmon02/speecht5_finetuned_voxpopuli_hy](https://huggingface.co/Edmon02/speecht5_finetuned_voxpopuli_hy) on the None dataset.
17
  It achieves the following results on the evaluation set:
18
+ - Loss: 0.4473
19
 
20
  ## Model description
21
 
 
42
  - total_train_batch_size: 32
43
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
44
  - lr_scheduler_type: linear
45
+ - lr_scheduler_warmup_steps: 125
46
+ - training_steps: 1000
 
47
 
48
  ### Training results
49
 
50
  | Training Loss | Epoch | Step | Validation Loss |
51
  |:-------------:|:------:|:----:|:---------------:|
52
+ | 0.4549 | 83.33 | 250 | 0.4517 |
53
+ | 0.4474 | 166.67 | 500 | 0.4474 |
54
+ | 0.4668 | 250.0 | 750 | 0.4475 |
55
+ | 0.4551 | 333.33 | 1000 | 0.4473 |
 
 
 
 
 
 
 
 
56
 
57
 
58
  ### Framework versions
59
 
60
  - Transformers 4.35.2
61
+ - Pytorch 2.1.0+cu121
62
+ - Datasets 2.16.0
63
  - Tokenizers 0.15.0