mrYou/whisper-small-ar

Whisper Small ar - younes matrab

This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.8027
• Wer: 61.6541

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: 2e-05
• train_batch_size: 16
• 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: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.7188	0.4167	10	2.7773	67.1053
1.6979	0.8333	20	2.4033	66.5414
1.3932	1.25	30	1.9422	66.3534
1.0467	1.6667	40	1.6225	65.2256
0.8824	2.0833	50	1.3586	64.4737
0.5935	2.5	60	1.0915	62.4060
0.4491	2.9167	70	0.8986	63.3459
0.3438	3.3333	80	0.8473	61.6541
0.2915	3.75	90	0.8132	60.1504
0.2391	4.1667	100	0.8027	61.6541

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1