ahishamm/finetuned-whisper-quranic-large-v3-10000

QURANIC Whisper Large V3 - 10000

This model is a fine-tuned version of openai/whisper-large-v3 on the QURANICWhisperDataset dataset. It achieves the following results on the evaluation set:

• Loss: 0.2528
• Wer: 99.9391

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: 8
• 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: 10000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0907	2.0	1000	0.1326	107.4287
0.0545	4.0	2000	0.1366	156.4231
0.0211	6.0	3000	0.1515	245.3308
0.0076	8.0	4000	0.1627	330.6630
0.0031	10.0	5000	0.1788	170.7794
0.0035	12.0	6000	0.1947	107.0630
0.0006	14.0	7000	0.2107	98.0091
0.0	16.0	8000	0.2208	97.8533
0.0	18.0	9000	0.2426	99.7833
0.0	20.0	10000	0.2528	99.9391

Framework versions

• Transformers 4.39.2
• Pytorch 2.2.0
• Datasets 2.18.0
• Tokenizers 0.15.1