Edit model card

Whisper Medium VI - Multi - Augmented

This model is a fine-tuned version of openai/whisper-medium on the following datasets:

It achieves the following results on the evaluation set:

  • Loss: 0.3696
  • Wer: 16.6594
  • Cer: 7.7625

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

Training:

Evaluation:

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 1e-05
  • train_batch_size: 32
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • training_steps: 5000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer Cer
0.1992 1.8 1000 0.2726 17.4929 8.2562
0.0402 3.6 2000 0.3317 17.4929 8.2588
0.0073 5.4 3000 0.3429 17.6793 8.8913
0.0014 7.19 4000 0.3599 19.0283 9.5103
0.0006 8.99 5000 0.3696 16.6594 7.7625

Framework versions

  • Transformers 4.26.0.dev0
  • Pytorch 1.13.1+cu117
  • Datasets 2.7.1.dev0
  • Tokenizers 0.13.2
Downloads last month
14
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.

Datasets used to train Scrya/whisper-medium-vi-augmented

Evaluation results