ap-3un1lfHfqPpJtZAAFiSEbP

This model is a fine-tuned version of openai/whisper-large-v3-turbo on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4069
• Model Preparation Time: 0.0106
• Wer: 0.1135

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• 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: 400
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Wer
0.2629	0.9858	52	0.3305	0.0106	0.1155
0.1722	1.9858	104	0.2931	0.0106	0.1089
0.1442	2.9858	156	0.2942	0.0106	0.1088
0.1056	3.9858	208	0.3149	0.0106	0.1073
0.0736	4.9858	260	0.3371	0.0106	0.1108
0.0853	5.9858	312	0.3656	0.0106	0.1193
0.0741	6.9858	364	0.3886	0.0106	0.1122
0.0515	7.9858	416	0.4183	0.0106	0.1229
0.0326	8.9858	468	0.4145	0.0106	0.1572
0.0111	9.9858	520	0.4069	0.0106	0.1135

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0