README.md

metadata

language:
  - en
license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - alxfng/noisycommonvoice
metrics:
  - wer
model-index:
  - name: Whisper Base Noisy
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Noisy Common Voice
          type: alxfng/noisycommonvoice
          config: en
          split: None
        metrics:
          - name: Wer
            type: wer
            value: 59.32123598390824

Whisper Base Noisy

This model is a fine-tuned version of openai/whisper-base on the Noisy Common Voice dataset. It achieves the following results on the evaluation set:

• Loss: 1.4454
• Wer: 59.3212

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: 4000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.3125	3.19	1000	1.0918	56.8476
0.0585	6.39	2000	1.2650	58.9703
0.0153	9.58	3000	1.3946	58.3412
0.0066	12.78	4000	1.4454	59.3212

Framework versions

• Transformers 4.30.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3