ninninz/whisper-ckm-1

whisper-large-v3-croatian-v3

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

• Loss: 2.8726
• Wer: 74.3118

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: 1.25e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0539	13.89	1000	2.3264	87.0594
0.0116	27.78	2000	2.5778	91.6517
0.0072	41.67	3000	2.8216	76.4729
0.0074	55.56	4000	2.8726	74.3118

Framework versions

• Transformers 4.36.1
• Pytorch 2.1.1+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0