whisper-tiny-minds14

This model is a fine-tuned version of openai/whisper-tiny on the PolyAI/minds14 dataset. It achieves the following results on the evaluation set:

• Loss: 0.6098
• Wer: 0.3448
• Wer Ortho: 0.3331

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Wer Ortho
0.1502	1.0	57	0.5102	0.3294	0.3122
0.0571	2.0	114	0.5214	0.3411	0.3282
0.0563	3.0	171	0.5403	0.3467	0.3300
0.0183	4.0	228	0.5591	0.3448	0.3288
0.0092	5.0	285	0.5730	0.3362	0.3202
0.0031	6.0	342	0.5926	0.3325	0.3177
0.0051	7.0	399	0.5965	0.3405	0.3282
0.0018	8.0	456	0.6059	0.3381	0.3257
0.0023	9.0	513	0.6084	0.3387	0.3263
0.0017	10.0	570	0.6098	0.3448	0.3331

Framework versions

• Transformers 5.16.1
• Pytorch 2.11.0+cu128
• Datasets 5.0.1
• Tokenizers 0.23.1