whisper-small-de-finetuned

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

• Loss: 0.1801
• Wer: 10.0733

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_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: 300
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2107	0.2962	500	0.2104	12.7071
0.2122	0.5924	1000	0.1997	11.8993
0.1820	0.8886	1500	0.1885	11.2831
0.0833	1.1848	2000	0.1840	10.9938
0.0812	1.4810	2500	0.1822	10.6481
0.0723	1.7773	3000	0.1795	10.3588
0.0293	2.0735	3500	0.1825	10.4828
0.0279	2.3697	4000	0.1818	10.3513
0.0285	2.6659	4500	0.1805	10.1822
0.0265	2.9621	5000	0.1801	10.0958
0.0267	3.0	5064	0.1801	10.0733

Framework versions

• Transformers 5.7.0
• Pytorch 2.10.0+cu128
• Datasets 4.8.5
• Tokenizers 0.22.2