whisper_small_no_overfit

This model is a fine-tuned version of Lingua-Connect/whisper-small-sw-bible on the kazeric/Giriama_bible_dataset_no_split dataset. It achieves the following results on the evaluation set:

• eval_loss: 0.0773
• eval_wer_ortho: 7.4069
• eval_wer: 3.9417
• eval_runtime: 741.4145
• eval_samples_per_second: 1.935
• eval_steps_per_second: 0.121
• epoch: 0.4190
• step: 150

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: 16
• 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: constant_with_warmup
• lr_scheduler_warmup_steps: 50
• training_steps: 500
• mixed_precision_training: Native AMP

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.5.1
• Tokenizers 0.21.1