whisper-large-v3-turbo-nope-en

This model is a fine-tuned version of openai/whisper-large-v3-turbo on the common_voice_16_1 dataset. It achieves the following results on the evaluation set:

• Loss: 3.6282
• Wer: 109.3298
• Cer: 76.3723

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-05
• train_batch_size: 8
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• 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: 500
• training_steps: 5000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
No log	0	0	8.8537	1137.9325	452.8176
4.2788	0.1	500	4.0872	160.9724	102.4893
4.1267	0.2	1000	3.8964	106.7017	73.6642
4.0032	0.3	1500	3.8182	110.0745	73.1695
3.8969	0.4	2000	3.8130	108.3224	77.8156
3.8894	0.5	2500	3.6948	109.8117	75.5939
3.9464	0.6	3000	3.6782	111.5637	76.2183
3.8295	0.7	3500	3.6481	133.2457	100.4784
3.75	0.8	4000	3.6358	112.1770	78.8859
3.7714	0.9	4500	3.6292	110.1183	77.8480
3.8284	1.0	5000	3.6282	109.3298	76.3723

Framework versions

• Transformers 4.54.1
• Pytorch 2.8.0.dev20250319+cu128
• Datasets 3.6.0
• Tokenizers 0.21.4