whisper-finetuned-fullsample-v1

This model is a fine-tuned version of openai/whisper-large-v3-turbo on the common_voice_17_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.3719
• Wer: 11.3120

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 8
• total_train_batch_size: 256
• total_eval_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: 600
• training_steps: 6000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0094	8.1384	1000	0.2714	24.2485
0.0008	16.2767	2000	0.3292	25.8955
0.0011	24.4151	3000	0.3289	12.6679
0.0003	32.5534	4000	0.3546	12.0631
0.0015	40.6918	5000	0.3405	12.0647
0.0002	48.8301	6000	0.3719	11.3120

Framework versions

• Transformers 4.46.0.dev0
• Pytorch 2.4.1+cu124
• Datasets 3.0.2.dev0
• Tokenizers 0.20.0