whisper-bible

This model is a fine-tuned version of abiyo27/whisper-small-ewe-2 on the Leonel-Maia/ewe_dataset dataset. It achieves the following results on the evaluation set:

• Loss: 0.4076
• Wer: 0.3640

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: 1
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 60.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.5366	0.4398	500	0.5260	0.4396
0.4324	0.8796	1000	0.4524	0.3907
0.3911	1.3193	1500	0.4261	0.3697
0.3889	1.7591	2000	0.4111	0.3650
0.3005	2.1988	2500	0.4157	0.3608
0.3021	2.6386	3000	0.4076	0.3640
0.2166	3.0783	3500	0.4138	0.3609
0.2123	3.5181	4000	0.4201	0.3602
0.2596	3.9579	4500	0.4182	0.3598
0.1703	4.3976	5000	0.4448	0.3698
0.179	4.8374	5500	0.4421	0.3798

Framework versions

• Transformers 4.50.3
• Pytorch 2.7.0+cu126
• Datasets 3.5.0
• Tokenizers 0.21.1