MelanieKoe's picture
Model save
9b9b9fd verified
|
raw
history blame
1.91 kB
metadata
license: apache-2.0
base_model: facebook/wav2vec2-base
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: w2v2-base-pretrained_lr1e-4_at0.9_da1
    results: []

w2v2-base-pretrained_lr1e-4_at0.9_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: 16.9074
  • Wer: 1.0

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: 60
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
25.5111 6.1 250 21.0169 1.0
8.8611 12.2 500 18.3862 1.0
8.2182 18.29 750 16.3061 1.0
8.1364 24.39 1000 15.6357 1.0
8.1315 30.49 1250 17.5602 1.0
8.1532 36.59 1500 16.3879 1.0
8.0249 42.68 1750 17.4829 1.0
8.0482 48.78 2000 16.9106 1.0
8.0999 54.88 2250 16.9074 1.0

Framework versions

  • Transformers 4.35.0
  • Pytorch 2.0.0
  • Datasets 2.14.6
  • Tokenizers 0.14.1