Hugging Face's logo

miosipof
/
speecht5_tts_commonvoice_it_v2

TensorBoard
Safetensors
speecht5
Generated from Trainer
Model card Files Metrics Community
miosipof commited on
Commit
c338c49
•
1 Parent(s): d8f10be

End of training

Browse files
Files changed (2) hide show
  1. README.md +30 -38
  2. model.safetensors +1 -1
README.md CHANGED
@@ -17,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the common_voice_13_0 dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 0.5076
21
 
22
  ## Model description
23
 
@@ -38,53 +38,45 @@ More information needed
38
  The following hyperparameters were used during training:
39
  - learning_rate: 1e-06
40
  - train_batch_size: 32
41
- - eval_batch_size: 4
42
  - seed: 42
 
 
43
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
44
  - lr_scheduler_type: linear
45
- - lr_scheduler_warmup_steps: 1000
46
- - num_epochs: 3
47
  - mixed_precision_training: Native AMP
48
 
49
  ### Training results
50
 
51
- | Training Loss | Epoch | Step | Validation Loss |
52
- |:-------------:|:------:|:-----:|:---------------:|
53
- | 0.9213 | 0.0994 | 500 | 0.7823 |
54
- | 0.8356 | 0.1987 | 1000 | 0.7026 |
55
- | 0.6804 | 0.2981 | 1500 | 0.6003 |
56
- | 0.6518 | 0.3975 | 2000 | 0.5751 |
57
- | 0.6242 | 0.4968 | 2500 | 0.5594 |
58
- | 0.6237 | 0.5962 | 3000 | 0.5514 |
59
- | 0.6122 | 0.6955 | 3500 | 0.5414 |
60
- | 0.597 | 0.7949 | 4000 | 0.5335 |
61
- | 0.5909 | 0.8943 | 4500 | 0.5322 |
62
- | 0.6009 | 0.9936 | 5000 | 0.5283 |
63
- | 0.6086 | 1.0930 | 5500 | 0.5258 |
64
- | 0.5812 | 1.1924 | 6000 | 0.5209 |
65
- | 0.5868 | 1.2917 | 6500 | 0.5191 |
66
- | 0.5689 | 1.3911 | 7000 | 0.5177 |
67
- | 0.5777 | 1.4905 | 7500 | 0.5182 |
68
- | 0.577 | 1.5898 | 8000 | 0.5169 |
69
- | 0.5594 | 1.6892 | 8500 | 0.5150 |
70
- | 0.5728 | 1.7886 | 9000 | 0.5144 |
71
- | 0.571 | 1.8879 | 9500 | 0.5125 |
72
- | 0.5739 | 1.9873 | 10000 | 0.5116 |
73
- | 0.5819 | 2.0866 | 10500 | 0.5102 |
74
- | 0.5633 | 2.1860 | 11000 | 0.5102 |
75
- | 0.5635 | 2.2854 | 11500 | 0.5093 |
76
- | 0.5809 | 2.3847 | 12000 | 0.5094 |
77
- | 0.5647 | 2.4841 | 12500 | 0.5086 |
78
- | 0.5593 | 2.5835 | 13000 | 0.5065 |
79
- | 0.5639 | 2.6828 | 13500 | 0.5077 |
80
- | 0.5511 | 2.7822 | 14000 | 0.5073 |
81
- | 0.5534 | 2.8816 | 14500 | 0.5071 |
82
- | 0.5532 | 2.9809 | 15000 | 0.5076 |
83
 
84
 
85
  ### Framework versions
86
 
87
  - Transformers 4.43.1
88
- - Pytorch 2.4.1+cu121
89
- - Datasets 3.0.0
90
  - Tokenizers 0.19.1
 
17
 
18
  This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the common_voice_13_0 dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 0.5556
21
 
22
  ## Model description
23
 
 
38
  The following hyperparameters were used during training:
39
  - learning_rate: 1e-06
40
  - train_batch_size: 32
41
+ - eval_batch_size: 16
42
  - seed: 42
43
+ - gradient_accumulation_steps: 2
44
+ - total_train_batch_size: 64
45
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
46
  - lr_scheduler_type: linear
47
+ - lr_scheduler_warmup_steps: 200
48
+ - training_steps: 2000
49
  - mixed_precision_training: Native AMP
50
 
51
  ### Training results
52
 
53
+ | Training Loss | Epoch | Step | Validation Loss |
54
+ |:-------------:|:------:|:----:|:---------------:|
55
+ | 0.9351 | 0.1894 | 100 | 0.8355 |
56
+ | 0.8426 | 0.3788 | 200 | 0.7500 |
57
+ | 0.8314 | 0.5682 | 300 | 0.7244 |
58
+ | 0.7912 | 0.7576 | 400 | 0.7078 |
59
+ | 0.778 | 0.9470 | 500 | 0.6908 |
60
+ | 0.7205 | 1.1364 | 600 | 0.6744 |
61
+ | 0.7272 | 1.3258 | 700 | 0.6469 |
62
+ | 0.7394 | 1.5152 | 800 | 0.6176 |
63
+ | 0.6816 | 1.7045 | 900 | 0.5874 |
64
+ | 0.6653 | 1.8939 | 1000 | 0.5748 |
65
+ | 0.658 | 2.0833 | 1100 | 0.5683 |
66
+ | 0.628 | 2.2727 | 1200 | 0.5662 |
67
+ | 0.6376 | 2.4621 | 1300 | 0.5632 |
68
+ | 0.6232 | 2.6515 | 1400 | 0.5612 |
69
+ | 0.625 | 2.8409 | 1500 | 0.5583 |
70
+ | 0.63 | 3.0303 | 1600 | 0.5588 |
71
+ | 0.6299 | 3.2197 | 1700 | 0.5567 |
72
+ | 0.6332 | 3.4091 | 1800 | 0.5558 |
73
+ | 0.6083 | 3.5985 | 1900 | 0.5551 |
74
+ | 0.6161 | 3.7879 | 2000 | 0.5556 |
 
 
 
 
 
 
 
 
 
 
75
 
76
 
77
  ### Framework versions
78
 
79
  - Transformers 4.43.1
80
+ - Pytorch 2.2.0
81
+ - Datasets 3.0.1
82
  - Tokenizers 0.19.1
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:806e40033caa8fb0c2c1f44f22259d5d7a39a7311d642c6067908cbf842c6d43
3
  size 577789320
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bcfae44bbce693bedff535ece5e77ed1abad23e227168d317aa8eaf76b146fa3
3
  size 577789320