README.md

metadata

language:
  - ar
license: apache-2.0
base_model: openai/whisper-medium
tags:
  - generated_from_trainer
datasets:
  - Arbi-Houssem/comondov
metrics:
  - wer
model-index:
  - name: Whisper Tunisien
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: comondov
          type: Arbi-Houssem/comondov
          args: 'config: ar, split: test'
        metrics:
          - name: Wer
            type: wer
            value: 62.71186440677966

Whisper Tunisien

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

• Loss: 1.7085
• Wer: 62.7119

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: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1296	50.0	200	1.5371	63.0200
0.0012	100.0	400	1.6181	61.4792
0.0005	150.0	600	1.6781	63.3282
0.0004	200.0	800	1.7005	61.9414
0.0003	250.0	1000	1.7085	62.7119

Framework versions

• Transformers 4.41.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1