README.md

metadata

license: apache-2.0
base_model: mayssakorbi/whisper-tiny-ar2
tags:
  - generated_from_trainer
datasets:
  - common_voice_16_1
metrics:
  - wer
model-index:
  - name: whisper-tiny-ar2
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: common_voice_16_1
          type: common_voice_16_1
          config: ar
          split: test[500:1000]
          args: ar
        metrics:
          - name: Wer
            type: wer
            value: 92.15686274509804

whisper-tiny-ar2

This model is a fine-tuned version of mayssakorbi/whisper-tiny-ar2 on the common_voice_16_1 dataset. It achieves the following results on the evaluation set:

• Loss: 0.9912
• Wer Ortho: 77.8635
• Wer: 92.1569

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_steps: 10
• training_steps: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
0.5668	0.4	50	0.9912	77.8635	92.1569

Framework versions

• Transformers 4.42.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1