gte-large-en-v1.5-based-ft-prompt-injection-detection-241205Weighted

This model is a fine-tuned version of Alibaba-NLP/gte-large-en-v1.5 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3220
• F1: 0.9471

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1
0.4432	0.2527	100	0.2279	0.9104
0.1996	0.5054	200	0.1793	0.9343
0.165	0.7581	300	0.1437	0.9450
0.1528	1.0107	400	0.1273	0.9531
0.1062	1.2634	500	0.1355	0.9490
0.1127	1.5161	600	0.1349	0.9544
0.1186	1.7688	700	0.1523	0.9496
0.1173	2.0215	800	0.1516	0.9483
0.0785	2.2742	900	0.1503	0.9528
0.0849	2.5268	1000	0.1623	0.9514
0.0898	2.7795	1100	0.1539	0.9460
0.0891	3.0322	1200	0.2415	0.9515
0.065	3.2849	1300	0.1589	0.9541
0.062	3.5376	1400	0.1499	0.9470
0.0677	3.7903	1500	0.1788	0.9445
0.0638	4.0430	1600	0.3220	0.9471

Framework versions

• Transformers 4.45.2
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.3