CasperMuz/whisper-small-sna

Whisper Small Shona - Casper Muziri

This model is a fine-tuned version of openai/whisper-small on the Google WAXAL dataset. It achieves the following results on the evaluation set:

• Loss: 0.4376
• Wer: 37.1131

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 adamw_torch_fused 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: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.3940	1.1338	1000	0.4608	41.3111
0.2771	2.2676	2000	0.4146	36.3338
0.2642	3.4014	3000	0.4148	36.6475
0.1771	4.5351	4000	0.4276	36.9636
0.1395	5.6689	5000	0.4376	37.1131

Framework versions

• Transformers 5.10.1
• Pytorch 2.12.0+cu130
• Datasets 4.8.5
• Tokenizers 0.22.2