whisper-small-tf

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

• Loss: 0.4403
• Wer: 0.3811

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
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• 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
0.4849	1.7564	500	0.4948	0.4388
0.3147	3.5101	1000	0.4403	0.3811
0.1881	5.2639	1500	0.4694	0.3843
0.124	7.0176	2000	0.5287	0.3952
0.05	8.7740	2500	0.6124	0.4024
0.0166	10.5277	3000	0.6773	0.3989
0.0083	12.2814	3500	0.7468	0.4017

Framework versions

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