whisper-large-v3-DODa

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: 0.7765
• Wer: 41.8438

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: 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: 500
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	1.0	359	0.7839	57.4236
1.1556	2.0	718	0.6461	49.8481
0.5055	3.0	1077	0.6097	47.1860
0.5055	4.0	1436	0.5795	44.7025
0.2883	5.0	1795	0.5925	43.7913
0.1566	6.0	2154	0.6317	43.4161
0.079	7.0	2513	0.6748	43.2196
0.079	8.0	2872	0.7083	43.1660
0.0281	9.0	3231	0.7577	42.5942
0.0094	10.0	3590	0.7765	41.8438

Framework versions

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