whisper_aii_nomimose

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: 1.2393
• Wer: 27.4280

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: 4
• eval_batch_size: 4
• 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: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.6668	1.0	114	0.9803	28.6671
1.4613	2.0	228	3.6964	117.2137
2.2089	3.0	342	1.8917	57.8366
0.8314	4.0	456	0.9927	29.9397
0.1661	5.0	570	0.9954	27.1266
0.0637	6.0	684	1.0471	26.7917
0.0269	7.0	798	1.1033	27.2271
0.0111	8.0	912	1.1958	27.0931
0.0078	9.0	1026	1.2212	27.4280
0.007	10.0	1140	1.2393	27.4280

Framework versions

• Transformers 4.52.0.dev0
• Pytorch 2.5.1+cu124
• Datasets 3.5.0
• Tokenizers 0.21.0