nik-o/subtitri-whisper-ka

Whisper ka

This model is a fine-tuned version of openai/whisper-small on the Subtitri voice 0.1v dataset. It achieves the following results on the evaluation set:

• Loss: 0.1584
• Wer: 42.4434

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 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: 500
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0471	2.9070	1000	0.0875	46.0260
0.0058	5.8140	2000	0.1194	44.0363
0.0004	8.7209	3000	0.1400	42.8457
0.0001	11.6279	4000	0.1529	42.5829
0.0	14.5349	5000	0.1584	42.4434

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0