whisper-id-jv-multilingual

This model is a fine-tuned version of openai/whisper-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4037
• Wer: 75.5167

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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
• training_steps: 4000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.6758	0.8418	500	0.6661	73.2909
0.3777	1.6835	1000	0.4950	56.4123
0.2271	2.5253	1500	0.4440	56.2798
0.1269	3.3670	2000	0.4188	60.1749
0.0739	4.2088	2500	0.4087	73.1849
0.0589	5.0505	3000	0.4039	71.8866
0.0473	5.8923	3500	0.4022	80.4716
0.0353	6.7340	4000	0.4037	75.5167

Framework versions

• Transformers 4.57.2
• Pytorch 2.9.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1