ditwoo's picture
update model card README.md
433ea64
metadata
license: mit
tags:
  - generated_from_trainer
datasets:
  - facebook/voxpopuli
model-index:
  - name: speecht5_finetuned_voxpopuli_pl_20epochs
    results: []

speecht5_finetuned_voxpopuli_pl_20epochs

This model is a fine-tuned version of microsoft/speecht5_tts on the facebook/voxpopuli dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4458

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: 4
  • eval_batch_size: 2
  • seed: 42
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 32
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 20

Training results

Training Loss Epoch Step Validation Loss
0.6474 1.0 210 0.5754
0.558 2.0 421 0.4976
0.5292 3.0 632 0.4788
0.5174 4.0 843 0.4693
0.5063 5.0 1053 0.4639
0.5053 6.0 1264 0.4599
0.4968 7.0 1475 0.4570
0.4913 8.0 1686 0.4549
0.4895 9.0 1896 0.4532
0.4876 10.0 2107 0.4520
0.487 11.0 2318 0.4501
0.4805 12.0 2529 0.4484
0.4789 13.0 2739 0.4495
0.4853 14.0 2950 0.4469
0.4797 15.0 3161 0.4468
0.4758 16.0 3372 0.4458
0.4729 17.0 3582 0.4458
0.4744 18.0 3793 0.4461
0.4759 19.0 4004 0.4462
0.4759 19.93 4200 0.4461

Framework versions

  • Transformers 4.29.2
  • Pytorch 2.0.0
  • Datasets 2.13.1
  • Tokenizers 0.13.3