mixstyle_multi-s

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

• Loss: 0.4159
• Cer: 105.1323

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAFACTOR and the args are: No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 912
• training_steps: 9129
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.4009	0.9997	3043	0.5068	102.8680
0.292	1.9993	6086	0.4280	108.6212
0.2075	2.9990	9129	0.4159	105.1323

Framework versions

• Transformers 4.47.1
• Pytorch 2.8.0+cu128
• Datasets 2.21.0
• Tokenizers 0.21.4