whisper-tiny-for-nigerian-common-languages

This model is a fine-tuned version of openai/whisper-tiny on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1096
• Wer: 86.2787

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.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
• training_steps: 500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
4.6091	0.7143	100	0.2441	105.2157
0.1921	1.4286	200	0.1646	105.6208
0.1473	2.1429	300	0.1348	89.4652
0.12	2.8571	400	0.1194	91.8664
0.0936	3.5714	500	0.1096	86.2787

Framework versions

• Transformers 4.50.1
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1