results / README.md
Meitt's picture
End of training
56d79b6 verified
metadata
language:
  - fl
license: mit
base_model: microsoft/speecht5_tts
tags:
  - Tets only
  - generated_from_trainer
datasets:
  - mewu/test
model-index:
  - name: Mewu custom
    results: []

Mewu custom

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

  • Loss: 1.4956

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: 0.001
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 2
  • 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: 500
  • training_steps: 1000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
0.4967 66.6667 100 0.4655
0.504 133.3333 200 0.4540
0.5434 200.0 300 0.6293
1.5883 266.6667 400 1.5112
1.5286 333.3333 500 1.5048
1.4761 400.0 600 1.5227
1.4458 466.6667 700 1.4967
1.4453 533.3333 800 1.4935
1.4446 600.0 900 1.4960
1.4451 666.6667 1000 1.4956

Framework versions

  • Transformers 4.40.1
  • Pytorch 2.3.0+cu118
  • Datasets 2.19.1
  • Tokenizers 0.19.1