🚀 speecht5_finetuned_google_fleurs_greek
This is a fine - tuned text - to - speech model based on microsoft/speecht5_tts, achieving a loss of 0.3920 on the evaluation set.
🚀 Quick Start
This model is a fine-tuned version of microsoft/speecht5_tts on the None dataset. It achieves the following results on the evaluation set:
📚 Documentation
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2.5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 40
Training results
Training Loss |
Epoch |
Step |
Validation Loss |
0.5174 |
1.0 |
583 |
0.4863 |
0.4948 |
2.0 |
1166 |
0.4611 |
0.4723 |
3.0 |
1749 |
0.4503 |
0.4763 |
4.0 |
2333 |
0.4438 |
0.4614 |
5.0 |
2916 |
0.4407 |
0.4569 |
6.0 |
3499 |
0.4387 |
0.4538 |
7.0 |
4082 |
0.4306 |
0.4539 |
8.0 |
4666 |
0.4282 |
0.4564 |
9.0 |
5249 |
0.4230 |
0.4493 |
10.0 |
5832 |
0.4222 |
0.445 |
11.0 |
6415 |
0.4190 |
0.4564 |
12.0 |
6999 |
0.4195 |
0.4381 |
13.0 |
7582 |
0.4161 |
0.4328 |
14.0 |
8165 |
0.4147 |
0.4424 |
15.0 |
8748 |
0.4140 |
0.4282 |
16.0 |
9332 |
0.4117 |
0.4349 |
17.0 |
9915 |
0.4090 |
0.4381 |
18.0 |
10498 |
0.4090 |
0.4328 |
19.0 |
11081 |
0.4073 |
0.4347 |
20.0 |
11665 |
0.4079 |
0.4293 |
21.0 |
12248 |
0.4055 |
0.4251 |
22.0 |
12831 |
0.4052 |
0.4359 |
23.0 |
13414 |
0.4023 |
0.4311 |
24.0 |
13998 |
0.4016 |
0.421 |
25.0 |
14581 |
0.4014 |
0.4162 |
26.0 |
15164 |
0.3991 |
0.4219 |
27.0 |
15747 |
0.3990 |
0.4247 |
28.0 |
16331 |
0.3989 |
0.4188 |
29.0 |
16914 |
0.3974 |
0.4229 |
30.0 |
17497 |
0.3976 |
0.4246 |
31.0 |
18080 |
0.3960 |
0.4219 |
32.0 |
18664 |
0.3956 |
0.4228 |
33.0 |
19247 |
0.3951 |
0.4183 |
34.0 |
19830 |
0.3946 |
0.4097 |
35.0 |
20413 |
0.3936 |
0.4245 |
36.0 |
20997 |
0.3935 |
0.4184 |
37.0 |
21580 |
0.3930 |
0.4198 |
38.0 |
22163 |
0.3937 |
0.4193 |
39.0 |
22746 |
0.3925 |
0.4096 |
39.98 |
23320 |
0.3920 |
Framework versions
- Transformers 4.30.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.13.1
- Tokenizers 0.13.3
📄 License
This project is licensed under the MIT license.