cyberseclabs/bert-classify-urls-v0.05

test_trainer_full

This model is a fine-tuned version of cyberseclabs/bert-classify-urls-v0.005 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1000
• Accuracy: 0.9823
• Precision: 0.9744
• Recall: 0.9732
• F1: 0.9738
• Roc Auc: 0.9965

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Roc Auc
0.1474	0.3846	500	0.1319	0.9704	0.9679	0.9437	0.9556	0.9900
0.1275	0.7692	1000	0.0938	0.9758	0.9739	0.9539	0.9638	0.9957
0.1005	1.1538	1500	0.0905	0.9767	0.9717	0.9590	0.9653	0.9964
0.0637	1.5385	2000	0.0842	0.9815	0.9689	0.9767	0.9728	0.9969
0.0568	1.9231	2500	0.0884	0.9794	0.9666	0.9727	0.9696	0.9972
0.0357	2.3077	3000	0.1019	0.9810	0.9786	0.9647	0.9716	0.9964
0.0272	2.6923	3500	0.1000	0.9823	0.9744	0.9732	0.9738	0.9965

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1