README.md

metadata

language:
  - en
license: apache-2.0
base_model: openai/whisper-small
tags:
  - generated_from_trainer
datasets:
  - jmcastelo17/FIFA_commentary
metrics:
  - wer
model-index:
  - name: Whisper Small FIFA_commentary
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: FIFA_commentary
          type: jmcastelo17/FIFA_commentary
        metrics:
          - name: Wer
            type: wer
            value: 24.600638977635782

Whisper Small FIFA_commentary

This model is a fine-tuned version of openai/whisper-small on the FIFA_commentary dataset. It achieves the following results on the evaluation set:

• Loss: 1.3180
• Wer: 24.6006

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: 50
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
5.3487	1.0	12	4.7958	46.0064
4.6559	2.0	24	3.5936	38.9776
3.5187	3.0	36	2.0147	30.3514
2.2037	4.0	48	1.4929	30.0319
1.054	5.0	60	1.3180	24.6006

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2