Hugging Face's logo

mikhail-panzo
/
zlm_b64_le5_s8000

Text-to-Audio
Transformers
TensorBoard
Safetensors
speecht5
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
mikhail-panzo commited on
Commit
5d258ba
·
verified ·
1 Parent(s): deb8684

End of training

Browse files
Files changed (1) hide show
  1. README.md +18 -19
README.md CHANGED
@@ -15,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on an unknown dataset.
17
  It achieves the following results on the evaluation set:
18
- - Loss: 0.3630
19
 
20
  ## Model description
21
 
@@ -43,30 +43,29 @@ The following hyperparameters were used during training:
43
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
44
  - lr_scheduler_type: linear
45
  - lr_scheduler_warmup_steps: 2000
46
- - training_steps: 8500
47
  - mixed_precision_training: Native AMP
48
 
49
  ### Training results
50
 
51
  | Training Loss | Epoch | Step | Validation Loss |
52
  |:-------------:|:------:|:----:|:---------------:|
53
- | 0.4525 | 0.4188 | 500 | 0.4043 |
54
- | 0.4541 | 0.8375 | 1000 | 0.3992 |
55
- | 0.4355 | 1.2563 | 1500 | 0.3946 |
56
- | 0.4315 | 1.6750 | 2000 | 0.3966 |
57
- | 0.4329 | 2.0938 | 2500 | 0.3881 |
58
- | 0.4235 | 2.5126 | 3000 | 0.3829 |
59
- | 0.4179 | 2.9313 | 3500 | 0.3775 |
60
- | 0.4116 | 3.3501 | 4000 | 0.3739 |
61
- | 0.4107 | 3.7688 | 4500 | 0.3721 |
62
- | 0.4029 | 4.1876 | 5000 | 0.3693 |
63
- | 0.409 | 4.6064 | 5500 | 0.3680 |
64
- | 0.4061 | 5.0251 | 6000 | 0.3662 |
65
- | 0.403 | 5.4439 | 6500 | 0.3654 |
66
- | 0.3958 | 5.8626 | 7000 | 0.3630 |
67
- | 0.3952 | 6.2814 | 7500 | 0.3635 |
68
- | 0.3971 | 6.7002 | 8000 | 0.3627 |
69
- | 0.4004 | 7.1189 | 8500 | 0.3630 |
70
 
71
 
72
  ### Framework versions
 
15
 
16
  This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on an unknown dataset.
17
  It achieves the following results on the evaluation set:
18
+ - Loss: 0.3771
19
 
20
  ## Model description
21
 
 
43
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
44
  - lr_scheduler_type: linear
45
  - lr_scheduler_warmup_steps: 2000
46
+ - training_steps: 8000
47
  - mixed_precision_training: Native AMP
48
 
49
  ### Training results
50
 
51
  | Training Loss | Epoch | Step | Validation Loss |
52
  |:-------------:|:------:|:----:|:---------------:|
53
+ | 0.7074 | 0.4188 | 500 | 0.6029 |
54
+ | 0.5916 | 0.8375 | 1000 | 0.4968 |
55
+ | 0.5206 | 1.2563 | 1500 | 0.4592 |
56
+ | 0.4979 | 1.6750 | 2000 | 0.4388 |
57
+ | 0.4852 | 2.0938 | 2500 | 0.4211 |
58
+ | 0.4615 | 2.5126 | 3000 | 0.4088 |
59
+ | 0.4521 | 2.9313 | 3500 | 0.4002 |
60
+ | 0.4431 | 3.3501 | 4000 | 0.3948 |
61
+ | 0.4393 | 3.7688 | 4500 | 0.3914 |
62
+ | 0.4271 | 4.1876 | 5000 | 0.3861 |
63
+ | 0.4317 | 4.6064 | 5500 | 0.3836 |
64
+ | 0.4265 | 5.0251 | 6000 | 0.3809 |
65
+ | 0.424 | 5.4439 | 6500 | 0.3794 |
66
+ | 0.4123 | 5.8626 | 7000 | 0.3786 |
67
+ | 0.4117 | 6.2814 | 7500 | 0.3776 |
68
+ | 0.4155 | 6.7002 | 8000 | 0.3771 |
 
69
 
70
 
71
  ### Framework versions