wav2vec2-speechcommands-german

This model is a fine-tuned version of facebook/wav2vec2-base-10k-voxpopuli-ft-de on an unkown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2481
  • Accuracy: 0.9180

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: 0.0003
  • train_batch_size: 16
  • eval_batch_size: 16
  • 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
  • num_epochs: 3.0
  • mixed_precision_training: Native AMP

Training results

Framework versions

  • Transformers 4.9.1
  • Pytorch 1.8.1
  • Datasets 1.7.0
  • Tokenizers 0.10.2
Downloads last month
4
Hosted inference API

Unable to determine this model’s pipeline type. Check the docs .