w2v2-base-pretrained_lr5e-5_at1_da1

This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 2.3753
  • Wer: 0.1837

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: 5e-05
  • 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
  • training_steps: 4000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
18.9508 6.1 250 4.5541 1.0
3.4223 12.2 500 3.2365 1.0
3.1022 18.29 750 3.1107 1.0
2.0891 24.39 1000 1.2943 0.4930
0.2477 30.49 1250 1.4675 0.2208
0.1323 36.59 1500 1.8152 0.2067
0.0908 42.68 1750 1.8821 0.1931
0.0721 48.78 2000 2.1984 0.2008
0.0584 54.88 2250 2.0544 0.1927
0.0483 60.98 2500 2.0092 0.1845
0.0437 67.07 2750 2.1545 0.1837
0.0347 73.17 3000 2.3775 0.1918
0.0331 79.27 3250 2.4051 0.1880
0.0302 85.37 3500 2.3556 0.1790
0.0279 91.46 3750 2.3822 0.1858
0.026 97.56 4000 2.3753 0.1837

Framework versions

  • Transformers 4.35.0
  • Pytorch 2.0.0
  • Datasets 2.14.6
  • Tokenizers 0.14.1
Downloads last month
15
Safetensors
Model size
94.4M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for MelanieKoe/w2v2-base-pretrained_lr5e-5_at1_da1

Finetuned
(688)
this model