OwLim/whisper-java

Whisper Small Java

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

• Loss: 0.9356
• Wer: 38.3731

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• 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: 100
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.8832	0.1	100	0.9373	51.7965
0.3579	1.075	200	0.9986	51.4516
0.2348	2.05	300	0.9892	46.0765
0.1397	3.025	400	1.0404	47.0250
0.0836	3.125	500	0.9862	46.9531
0.0515	4.1	600	1.0148	42.2248
0.0222	5.075	700	0.9917	40.2846
0.0191	6.05	800	0.9665	39.3360
0.0078	7.025	900	0.9541	39.0486
0.0009	7.125	1000	0.9356	38.3731

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1