whisper-small-SplitEndMovie-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: 0.9463
• Cer: 38.1911
• Wer: 59.9299

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
3.4626	1.0	2620	1.1588	42.0462	68.3704
1.7429	2.0	5240	1.0783	44.3746	70.5728
1.475	3.0	7860	1.0208	42.7916	66.8927
1.3263	4.0	10480	0.9993	41.8993	64.1945
1.2324	5.0	13100	0.9805	39.9841	60.0842
1.1663	6.0	15720	0.9674	40.6211	62.8151
1.1149	7.0	18340	0.9556	38.5933	59.7241
1.0821	8.0	20960	0.9517	39.3292	61.9593
1.0589	9.0	23580	0.9463	38.1911	59.9299

Framework versions

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