whisper-large-v3-DODa2

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: 0.7930
• Wer: 40.9862

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: 32
• eval_batch_size: 16
• seed: 42
• 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: 150
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	1.0	359	0.7424	59.3532
0.9478	2.0	718	0.6449	48.3473
0.4349	3.0	1077	0.6237	46.5964
0.4349	4.0	1436	0.6294	45.4351
0.2436	5.0	1795	0.6367	43.5948
0.1274	6.0	2154	0.6712	44.1844
0.0638	7.0	2513	0.7189	43.4340
0.0638	8.0	2872	0.7456	42.1297
0.0227	9.0	3231	0.7789	41.7902
0.0074	10.0	3590	0.7930	40.9862

Framework versions

• Transformers 4.46.2
• Pytorch 2.4.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.0