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

QURANIC Whisper Large V3 - full

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: 0.0375
• Wer: 121.0055

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
• 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
0.1349	0.2	1000	0.1227	256.8256
0.1098	0.4	2000	0.0918	438.2193
0.1071	0.6	3000	0.0839	286.1663
0.0837	0.8	4000	0.0737	295.5091
0.0672	1.0	5000	0.0611	293.6147
0.03	1.2	6000	0.0559	204.9680
0.0104	1.4	7000	0.0485	189.5761
0.0245	1.6	8000	0.0456	141.0698
0.0446	1.8	9000	0.0398	134.5774
0.0231	2.0	10000	0.0375	121.0055

Framework versions

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