whisper-small-en

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: 0.9996
• Wer: 68.4497

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: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• 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: 100
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0326	1.0	100	0.8999	49.6422
0.0297	2.0	200	0.9166	58.1942
0.0388	3.0	300	0.9314	73.6627
0.0164	4.0	400	0.9372	60.9881
0.0163	5.0	500	0.9578	64.2930
0.0258	6.0	600	0.9645	66.7121
0.0261	7.0	700	0.9639	65.5877
0.0092	8.0	800	0.9843	64.2589
0.0069	9.0	900	0.9951	65.6899
0.0056	10.0	1000	0.9996	68.4497

Framework versions

• Transformers 4.57.3
• Pytorch 2.9.0+cu126
• Datasets 4.4.1
• Tokenizers 0.22.1