keyi10's picture
update model card README.md
29d2bb4
metadata
license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - lj_speech
model-index:
  - name: wav2vec2-model-training
    results: []

wav2vec2-model-training

This model is a fine-tuned version of facebook/wav2vec2-base on the lj_speech dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1258
  • Wer: 0.1446

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: 0.0001
  • train_batch_size: 32
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 1000
  • num_epochs: 5

Training results

Training Loss Epoch Step Validation Loss Wer
3.5099 1.52 500 1.1323 0.6740
0.3293 3.05 1000 0.1430 0.1851
0.1047 4.57 1500 0.1258 0.1446

Framework versions

  • Transformers 4.17.0
  • Pytorch 2.1.2
  • Datasets 2.15.0
  • Tokenizers 0.15.2