whisper-tiny-banking-en

This model is a fine-tuned version of openai/whisper-tiny on the PolyAI/minds14 dataset. It achieves the following results on the evaluation set:

• Loss: 0.5118
• Wer Ortho: 0.2988
• Wer: 0.2913

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: 3e-06
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• 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: cosine
• lr_scheduler_warmup_steps: 150
• training_steps: 400
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
6.8838	3.5714	50	3.1258	0.4805	0.3463
3.7031	7.1429	100	1.4470	0.4004	0.3430
1.1011	10.7143	150	0.5527	0.3143	0.3036
0.7571	14.2857	200	0.5194	0.3163	0.3068
0.6139	17.8571	250	0.5125	0.2894	0.2817
0.5393	21.4286	300	0.5131	0.3015	0.2946
0.4649	25.0	350	0.5118	0.2981	0.2913
0.4781	28.5714	400	0.5118	0.2988	0.2913

Framework versions

• Transformers 4.48.0.dev0
• Pytorch 2.6.0+cu126
• Datasets 3.2.0
• Tokenizers 0.21.0