README.md

metadata

language:
  - ar
license: apache-2.0
tags:
  - generated_from_trainer
base_model: tarteel-ai/whisper-base-ar-quran
datasets:
  - zolfa
metrics:
  - wer
model-index:
  - name: Whisper-raghadomar
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: Zolfa Dataset
          type: zolfa
          args: 'config: ar, split: test'
        metrics:
          - type: wer
            value: 6.896551724137931
            name: Wer

Whisper-raghadomar

This model is a fine-tuned version of tarteel-ai/whisper-base-ar-quran on the Zolfa Dataset dataset. It achieves the following results on the evaluation set:

• Loss: 0.0196
• Wer: 6.8966

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: 5
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0084	2.8571	100	0.0224	6.8966
0.0017	5.7143	200	0.0177	6.8966
0.0004	8.5714	300	0.0161	6.8966
0.0001	11.4286	400	0.0187	6.8966
0.0003	14.2857	500	0.0171	6.8966
0.0001	17.1429	600	0.0200	6.8966
0.0001	20.0	700	0.0169	6.8966
0.0002	22.8571	800	0.0187	6.8966
0.0002	25.7143	900	0.0186	6.8966
0.0001	28.5714	1000	0.0196	6.8966

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1