README.md

metadata

license: apache-2.0
base_model: NbAiLab/nb-whisper-tiny
tags:
  - generated_from_trainer
datasets:
  - ravnursson_asr
metrics:
  - wer
model-index:
  - name: whisper-tiny-no-fo-5k_steps
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: ravnursson_asr
          type: ravnursson_asr
          config: ravnursson_asr
          split: test
          args: ravnursson_asr
        metrics:
          - name: Wer
            type: wer
            value: 39.16347896114355

whisper-tiny-no-fo-5k_steps

This model is a fine-tuned version of NbAiLab/nb-whisper-tiny on the ravnursson_asr dataset. It achieves the following results on the evaluation set:

• Loss: 0.5429
• Wer: 39.1635

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: 1e-05
• train_batch_size: 16
• 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: 500
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.0737	0.2320	1000	1.0306	58.6169
0.7129	0.4640	2000	0.7040	46.5522
0.6255	0.6961	3000	0.6029	42.4300
0.6188	0.9281	4000	0.5573	39.8430
0.5141	1.1601	5000	0.5429	39.1635

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1