arkanalexei/whisper-tiny-sundanese-pretrained

BisiX: Sundanese Whisper (Reset Params)

This model is a fine-tuned version of openai/whisper-tiny.en on the SU ID ASR dataset. It achieves the following results on the evaluation set:

• Loss: 10.8462
• Wer: 100.0
• Cer: 100.0

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 30
• training_steps: 150
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
10.856	0.3529	30	10.8546	1611.8562	2230.5547
10.853	0.7059	60	10.8517	100.0	83.0372
10.8498	1.0588	90	10.8486	100.0	95.0543
10.8475	1.4118	120	10.8467	100.0	100.0
10.8463	1.7647	150	10.8462	100.0	100.0

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.1+cu124
• Datasets 3.0.1
• Tokenizers 0.20.0