whisper-small-Split-Sentences-specAug

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: 0.8726
• Cer: 31.4957
• Wer: 52.7152

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: 5e-06
• train_batch_size: 4
• eval_batch_size: 8
• 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
• lr_scheduler_warmup_steps: 500
• num_epochs: 9
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer	Wer
3.4095	1.0	2706	1.0676	35.6254	61.2682
1.7232	2.0	5412	0.9871	35.7510	58.4606
1.4588	3.0	8118	0.9348	38.6172	61.8845
1.3133	4.0	10824	0.9094	31.8923	52.5577
1.221	5.0	13530	0.8874	31.6711	52.2838
1.1556	6.0	16236	0.8726	31.4957	52.7152
1.1084	7.0	18942	0.8692	32.7075	53.9341
1.0673	8.0	21648	0.8635	32.1992	51.4620
1.0478	9.0	24354	0.8614	32.5779	53.4616

Framework versions

• Transformers 4.56.1
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.22.0