volkan-aslan/whisper-base-tr-v2

Whisper Base TR V2 - Fast - Volkan ASLAN

This model is a fine-tuned version of openai/whisper-base on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4148
• Wer: 30.8729

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• 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: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 6000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2299	0.6894	1000	0.4219	34.5602
0.1368	1.3785	2000	0.4092	32.7446
0.0736	2.0676	3000	0.3968	31.2152
0.0695	2.7570	4000	0.3973	31.0636
0.0403	3.4461	5000	0.4112	31.0398
0.0241	4.1351	6000	0.4148	30.8729

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1