whisper-large-v3-turbo-16e-20000u

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: 0.6263
• Wer: 20.9247

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• 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: 2000
• training_steps: 20000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0	0	8.8155	100.0
0.6438	0.1	2000	1.1388	50.2930
0.5006	0.2	4000	0.9544	42.9781
0.3766	0.3	6000	0.8654	32.7328
0.3788	0.4	8000	0.7993	29.0862
0.3314	0.5	10000	0.7550	31.1048
0.2993	0.6	12000	0.7117	24.2240
0.3074	0.7	14000	0.6822	26.6768
0.3779	0.8	16000	0.6472	27.6319
0.3094	0.9	18000	0.6297	20.5774
0.3768	1.0	20000	0.6263	20.9247

Framework versions

• Transformers 4.54.0
• Pytorch 2.8.0.dev20250319+cu128
• Datasets 3.6.0
• Tokenizers 0.21.2