Edit model card

wav2vec2-large-xls-r-300m-gn-hu-colab

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5578
  • Wer: 0.5363

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: 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
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Wer
5.9484 4.97 400 3.2466 1.0
1.5016 9.94 800 0.6870 0.8148
0.2838 14.91 1200 0.5523 0.6087
0.1566 19.88 1600 0.5592 0.5801
0.105 24.84 2000 0.5685 0.5655
0.0761 29.81 2400 0.5578 0.5363

Framework versions

  • Transformers 4.34.1
  • Pytorch 2.1.0+cu118
  • Datasets 2.14.6
  • Tokenizers 0.14.1
Downloads last month
8

Finetuned from

Evaluation results