Edit model card

wav2vec2-large-xls-r-300m-ml-superb-xty

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

  • Loss: 1.6099
  • Wer: 0.8114

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: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 30
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss Wer
3.1825 0.8219 30 3.2071 1.0
3.0491 1.6438 60 3.0508 1.0
2.9717 2.4658 90 3.0385 1.0
2.93 3.2877 120 2.9222 1.0
2.6444 4.1096 150 2.3753 0.9931
2.05 4.9315 180 1.9591 0.9868
1.6856 5.7534 210 1.7810 0.9478
1.4182 6.5753 240 1.6843 0.8843
1.1773 7.3973 270 1.6370 0.8554
1.0521 8.2192 300 1.5868 0.8215
0.881 9.0411 330 1.5935 0.8202
0.7605 9.8630 360 1.6099 0.8114

Framework versions

  • Transformers 4.40.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1
Downloads last month
2
Safetensors
Model size
315M params
Tensor type
F32
·

Finetuned from

Evaluation results