___n____model

This model is a fine-tuned version of openai/whisper-medium on the whsNect/n dataset. It achieves the following results on the evaluation set:

• Loss: 0.0145
• Cer: 0.8035

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: 5e-06
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 6000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.022	0.1743	500	0.0191	0.8758
0.0173	0.3486	1000	0.0171	0.8382
0.0185	0.5229	1500	0.0161	0.8369
0.0163	0.6972	2000	0.0164	0.8223
0.016	0.8715	2500	0.0157	0.8039
0.0143	1.0458	3000	0.0160	0.8420
0.016	1.2201	3500	0.0153	0.7954
0.0165	1.3945	4000	0.0152	0.8069
0.015	1.5688	4500	0.0151	0.8001
0.0144	1.7431	5000	0.0149	0.8022
0.0128	1.9174	5500	0.0146	0.8181
0.013	2.0917	6000	0.0145	0.8035

Framework versions

• Transformers 4.41.2
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.19.1