nassimmasi01/whisper-tiny-en_v7

Whisper tiny En v7 Naji

This model is a fine-tuned version of openai/whisper-tiny on the Common Voice 1 dataset. It achieves the following results on the evaluation set:

• Loss: 4.9970
• Wer Ortho: 155.2273
• Wer: 151.1430

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-08
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_steps: 50
• training_steps: 500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
3.4745	0.0824	50	4.9978	155.6762	151.5608
3.4696	0.1647	100	4.9978	155.6473	151.5326
3.453	0.2471	150	4.9976	155.5652	151.4528
3.379	0.3295	200	4.9974	155.4783	151.3684
3.3153	0.4119	250	4.9973	155.4687	151.3637
3.0962	0.4942	300	4.9972	155.1791	151.0820
3.8311	0.5766	350	4.9971	156.1348	152.0161
3.7729	0.6590	400	4.9970	155.2273	151.1430

Framework versions

• Transformers 4.52.4
• Pytorch 2.7.1+cu126
• Datasets 2.14.6
• Tokenizers 0.21.1