metadata
license: mit
base_model: microsoft/speecht5_tts
tags:
- generated_from_trainer
datasets:
- common_voice_13_0
model-index:
- name: speecht5_tts
results: []
speecht5_tts
This model is a fine-tuned version of microsoft/speecht5_tts on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.5720
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: 1e-05
- train_batch_size: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 20000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.0398 | 1.0 | 100 | 0.7784 |
| 0.8663 | 2.0 | 200 | 0.7059 |
| 0.7824 | 3.0 | 300 | 0.6734 |
| 0.7181 | 4.0 | 400 | 0.5776 |
| 0.6418 | 5.0 | 500 | 0.5584 |
| 0.6127 | 6.0 | 600 | 0.5452 |
| 0.5939 | 7.0 | 700 | 0.5386 |
| 0.5924 | 8.0 | 800 | 0.5420 |
| 0.5887 | 9.0 | 900 | 0.5392 |
| 0.5769 | 10.0 | 1000 | 0.5319 |
| 0.578 | 11.0 | 1100 | 0.5345 |
| 0.5799 | 12.0 | 1200 | 0.5257 |
| 0.5614 | 13.0 | 1300 | 0.5342 |
| 0.554 | 14.0 | 1400 | 0.5223 |
| 0.551 | 15.0 | 1500 | 0.5209 |
| 0.5587 | 16.0 | 1600 | 0.5221 |
| 0.5485 | 17.0 | 1700 | 0.5193 |
| 0.5354 | 18.0 | 1800 | 0.5216 |
| 0.5417 | 19.0 | 1900 | 0.5260 |
| 0.5319 | 20.0 | 2000 | 0.5218 |
| 0.5354 | 21.0 | 2100 | 0.5212 |
| 0.5316 | 22.0 | 2200 | 0.5233 |
| 0.5295 | 23.0 | 2300 | 0.5222 |
| 0.5407 | 24.0 | 2400 | 0.5317 |
| 0.5309 | 25.0 | 2500 | 0.5258 |
| 0.5196 | 26.0 | 2600 | 0.5317 |
| 0.5195 | 27.0 | 2700 | 0.5325 |
| 0.5134 | 28.0 | 2800 | 0.5193 |
| 0.5143 | 29.0 | 2900 | 0.5254 |
| 0.5227 | 30.0 | 3000 | 0.5260 |
| 0.5157 | 31.0 | 3100 | 0.5311 |
| 0.5214 | 32.0 | 3200 | 0.5292 |
| 0.5196 | 33.0 | 3300 | 0.5283 |
| 0.522 | 34.0 | 3400 | 0.5296 |
| 0.5193 | 35.0 | 3500 | 0.5252 |
| 0.5156 | 36.0 | 3600 | 0.5272 |
| 0.5182 | 37.0 | 3700 | 0.5318 |
| 0.5079 | 38.0 | 3800 | 0.5289 |
| 0.5103 | 39.0 | 3900 | 0.5374 |
| 0.5044 | 40.0 | 4000 | 0.5289 |
| 0.5021 | 41.0 | 4100 | 0.5372 |
| 0.5202 | 42.0 | 4200 | 0.5384 |
| 0.5022 | 43.0 | 4300 | 0.5281 |
| 0.498 | 44.0 | 4400 | 0.5327 |
| 0.4991 | 45.0 | 4500 | 0.5351 |
| 0.4972 | 46.0 | 4600 | 0.5383 |
| 0.5075 | 47.0 | 4700 | 0.5319 |
| 0.5063 | 48.0 | 4800 | 0.5365 |
| 0.4964 | 49.0 | 4900 | 0.5361 |
| 0.5021 | 50.0 | 5000 | 0.5353 |
| 0.4981 | 51.0 | 5100 | 0.5419 |
| 0.4914 | 52.0 | 5200 | 0.5398 |
| 0.5016 | 53.0 | 5300 | 0.5499 |
| 0.4911 | 54.0 | 5400 | 0.5484 |
| 0.5048 | 55.0 | 5500 | 0.5369 |
| 0.4828 | 56.0 | 5600 | 0.5452 |
| 0.4906 | 57.0 | 5700 | 0.5446 |
| 0.4922 | 58.0 | 5800 | 0.5451 |
| 0.4851 | 59.0 | 5900 | 0.5444 |
| 0.4898 | 60.0 | 6000 | 0.5461 |
| 0.4858 | 61.0 | 6100 | 0.5388 |
| 0.4966 | 62.0 | 6200 | 0.5408 |
| 0.4935 | 63.0 | 6300 | 0.5442 |
| 0.4824 | 64.0 | 6400 | 0.5466 |
| 0.4967 | 65.0 | 6500 | 0.5486 |
| 0.4789 | 66.0 | 6600 | 0.5429 |
| 0.481 | 67.0 | 6700 | 0.5516 |
| 0.4873 | 68.0 | 6800 | 0.5452 |
| 0.4816 | 69.0 | 6900 | 0.5497 |
| 0.4911 | 70.0 | 7000 | 0.5546 |
| 0.4805 | 71.0 | 7100 | 0.5460 |
| 0.4781 | 72.0 | 7200 | 0.5486 |
| 0.4923 | 73.0 | 7300 | 0.5479 |
| 0.4779 | 74.0 | 7400 | 0.5467 |
| 0.4778 | 75.0 | 7500 | 0.5513 |
| 0.4826 | 76.0 | 7600 | 0.5513 |
| 0.4756 | 77.0 | 7700 | 0.5509 |
| 0.4698 | 78.0 | 7800 | 0.5528 |
| 0.4868 | 79.0 | 7900 | 0.5559 |
| 0.478 | 80.0 | 8000 | 0.5523 |
| 0.472 | 81.0 | 8100 | 0.5570 |
| 0.4835 | 82.0 | 8200 | 0.5542 |
| 0.4813 | 83.0 | 8300 | 0.5538 |
| 0.472 | 84.0 | 8400 | 0.5503 |
| 0.4726 | 85.0 | 8500 | 0.5521 |
| 0.4804 | 86.0 | 8600 | 0.5577 |
| 0.4836 | 87.0 | 8700 | 0.5554 |
| 0.4786 | 88.0 | 8800 | 0.5603 |
| 0.471 | 89.0 | 8900 | 0.5597 |
| 0.4782 | 90.0 | 9000 | 0.5543 |
| 0.4713 | 91.0 | 9100 | 0.5549 |
| 0.4825 | 92.0 | 9200 | 0.5585 |
| 0.4749 | 93.0 | 9300 | 0.5598 |
| 0.4684 | 94.0 | 9400 | 0.5574 |
| 0.4732 | 95.0 | 9500 | 0.5577 |
| 0.4663 | 96.0 | 9600 | 0.5596 |
| 0.4618 | 97.0 | 9700 | 0.5555 |
| 0.4637 | 98.0 | 9800 | 0.5563 |
| 0.4731 | 99.0 | 9900 | 0.5578 |
| 0.485 | 100.0 | 10000 | 0.5591 |
| 0.475 | 101.0 | 10100 | 0.5598 |
| 0.4631 | 102.0 | 10200 | 0.5539 |
| 0.4636 | 103.0 | 10300 | 0.5567 |
| 0.4686 | 104.0 | 10400 | 0.5554 |
| 0.4677 | 105.0 | 10500 | 0.5530 |
| 0.4705 | 106.0 | 10600 | 0.5555 |
| 0.4596 | 107.0 | 10700 | 0.5567 |
| 0.4689 | 108.0 | 10800 | 0.5552 |
| 0.4698 | 109.0 | 10900 | 0.5591 |
| 0.4767 | 110.0 | 11000 | 0.5583 |
| 0.466 | 111.0 | 11100 | 0.5594 |
| 0.4792 | 112.0 | 11200 | 0.5604 |
| 0.4692 | 113.0 | 11300 | 0.5635 |
| 0.4675 | 114.0 | 11400 | 0.5597 |
| 0.467 | 115.0 | 11500 | 0.5587 |
| 0.4653 | 116.0 | 11600 | 0.5610 |
| 0.468 | 117.0 | 11700 | 0.5608 |
| 0.4649 | 118.0 | 11800 | 0.5625 |
| 0.4614 | 119.0 | 11900 | 0.5606 |
| 0.4663 | 120.0 | 12000 | 0.5626 |
| 0.4654 | 121.0 | 12100 | 0.5623 |
| 0.4582 | 122.0 | 12200 | 0.5613 |
| 0.4621 | 123.0 | 12300 | 0.5650 |
| 0.459 | 124.0 | 12400 | 0.5617 |
| 0.4538 | 125.0 | 12500 | 0.5609 |
| 0.4602 | 126.0 | 12600 | 0.5636 |
| 0.462 | 127.0 | 12700 | 0.5661 |
| 0.4647 | 128.0 | 12800 | 0.5585 |
| 0.4616 | 129.0 | 12900 | 0.5638 |
| 0.4691 | 130.0 | 13000 | 0.5658 |
| 0.4645 | 131.0 | 13100 | 0.5646 |
| 0.4581 | 132.0 | 13200 | 0.5638 |
| 0.4546 | 133.0 | 13300 | 0.5656 |
| 0.4633 | 134.0 | 13400 | 0.5651 |
| 0.4626 | 135.0 | 13500 | 0.5652 |
| 0.4663 | 136.0 | 13600 | 0.5657 |
| 0.4598 | 137.0 | 13700 | 0.5639 |
| 0.4711 | 138.0 | 13800 | 0.5650 |
| 0.4595 | 139.0 | 13900 | 0.5678 |
| 0.4586 | 140.0 | 14000 | 0.5638 |
| 0.4562 | 141.0 | 14100 | 0.5668 |
| 0.456 | 142.0 | 14200 | 0.5673 |
| 0.4561 | 143.0 | 14300 | 0.5694 |
| 0.4562 | 144.0 | 14400 | 0.5685 |
| 0.4583 | 145.0 | 14500 | 0.5642 |
| 0.446 | 146.0 | 14600 | 0.5690 |
| 0.4631 | 147.0 | 14700 | 0.5647 |
| 0.4553 | 148.0 | 14800 | 0.5673 |
| 0.4569 | 149.0 | 14900 | 0.5658 |
| 0.4618 | 150.0 | 15000 | 0.5645 |
| 0.4586 | 151.0 | 15100 | 0.5693 |
| 0.4474 | 152.0 | 15200 | 0.5683 |
| 0.4499 | 153.0 | 15300 | 0.5687 |
| 0.4533 | 154.0 | 15400 | 0.5687 |
| 0.452 | 155.0 | 15500 | 0.5693 |
| 0.4578 | 156.0 | 15600 | 0.5681 |
| 0.4534 | 157.0 | 15700 | 0.5697 |
| 0.4554 | 158.0 | 15800 | 0.5695 |
| 0.4532 | 159.0 | 15900 | 0.5728 |
| 0.4471 | 160.0 | 16000 | 0.5746 |
| 0.4528 | 161.0 | 16100 | 0.5715 |
| 0.4535 | 162.0 | 16200 | 0.5677 |
| 0.4487 | 163.0 | 16300 | 0.5719 |
| 0.4539 | 164.0 | 16400 | 0.5673 |
| 0.4493 | 165.0 | 16500 | 0.5722 |
| 0.4463 | 166.0 | 16600 | 0.5725 |
| 0.4547 | 167.0 | 16700 | 0.5693 |
| 0.4557 | 168.0 | 16800 | 0.5697 |
| 0.4548 | 169.0 | 16900 | 0.5727 |
| 0.4551 | 170.0 | 17000 | 0.5732 |
| 0.4633 | 171.0 | 17100 | 0.5725 |
| 0.4529 | 172.0 | 17200 | 0.5744 |
| 0.4542 | 173.0 | 17300 | 0.5745 |
| 0.4551 | 174.0 | 17400 | 0.5725 |
| 0.4562 | 175.0 | 17500 | 0.5724 |
| 0.4473 | 176.0 | 17600 | 0.5746 |
| 0.4491 | 177.0 | 17700 | 0.5714 |
| 0.4498 | 178.0 | 17800 | 0.5729 |
| 0.4612 | 179.0 | 17900 | 0.5704 |
| 0.4565 | 180.0 | 18000 | 0.5725 |
| 0.4571 | 181.0 | 18100 | 0.5716 |
| 0.4561 | 182.0 | 18200 | 0.5718 |
| 0.4542 | 183.0 | 18300 | 0.5726 |
| 0.4563 | 184.0 | 18400 | 0.5730 |
| 0.4517 | 185.0 | 18500 | 0.5726 |
| 0.449 | 186.0 | 18600 | 0.5715 |
| 0.4513 | 187.0 | 18700 | 0.5744 |
| 0.4487 | 188.0 | 18800 | 0.5769 |
| 0.4516 | 189.0 | 18900 | 0.5759 |
| 0.4524 | 190.0 | 19000 | 0.5741 |
| 0.4586 | 191.0 | 19100 | 0.5730 |
| 0.4507 | 192.0 | 19200 | 0.5748 |
| 0.4488 | 193.0 | 19300 | 0.5728 |
| 0.4635 | 194.0 | 19400 | 0.5739 |
| 0.4566 | 195.0 | 19500 | 0.5779 |
| 0.4556 | 196.0 | 19600 | 0.5745 |
| 0.4577 | 197.0 | 19700 | 0.5776 |
| 0.4481 | 198.0 | 19800 | 0.5746 |
| 0.4576 | 199.0 | 19900 | 0.5737 |
| 0.4523 | 200.0 | 20000 | 0.5720 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.14.6
- Tokenizers 0.14.1