UDA-LIDI-Whisper-large-v3-turbo-ECU-911

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

• Loss: 0.8685
• Wer: 40.1779

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
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.7289	1.0	91	0.6513	40.7708
0.4426	2.0	182	0.6487	40.1779
0.298	3.0	273	0.6699	40.1186
0.2058	4.0	364	0.6912	42.6285
0.1435	5.0	455	0.7103	39.6838
0.1022	6.0	546	0.7852	41.8379
0.0735	7.0	637	0.8315	40.6324
0.0568	8.0	728	0.8265	40.6126
0.0444	9.0	819	0.8538	40.0198
0.0399	9.8950	900	0.8685	40.1779

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0