leenag/KLuke_Malasar

This model is a fine-tuned version of openai/whisper-small on the Spoken Bible Corpus: Malasar dataset. It achieves the following results on the evaluation set:

• Loss: 0.6042
• Wer: 48.6008

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0372	11.3636	250	0.4525	55.4540
0.0108	22.7273	500	0.4908	52.9412
0.002	34.0909	750	0.5524	51.3992
0.0002	45.4545	1000	0.5567	49.1719
0.0	56.8182	1250	0.5861	48.7721
0.0	68.1818	1500	0.5971	48.8292
0.0	79.5455	1750	0.6025	48.6008
0.0	90.9091	2000	0.6042	48.6008

Framework versions

• Transformers 4.41.2
• Pytorch 2.0.1+cu117
• Datasets 2.16.0
• Tokenizers 0.19.1