whisper-large-v3-turbo-nope

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: 0.4736
• Wer: 15.5497
• Cer: 8.5462

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
1.0467	0.1	500	1.6165	59.5707	44.0850
0.5278	0.2	1000	0.9622	29.9606	18.2843
0.3648	0.3	1500	0.7630	24.7043	14.1328
0.346	0.4	2000	0.6770	21.1126	11.6111
0.2755	0.5	2500	0.6325	21.1564	12.0409
0.3617	0.6	3000	0.5587	22.8208	11.4895
0.2343	0.7	3500	0.5154	16.8200	12.4625
0.2468	0.8	4000	0.4874	18.5721	13.8733
0.2459	0.9	4500	0.4765	15.5497	8.3516
0.2776	1.0	5000	0.4736	15.5497	8.5462

Framework versions

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