distilbert-base-uncased-lora-text-classification-jackhao-jailbreak

This model is a fine-tuned version of distilbert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0744
• F1: 0.9809
• Auprc: 0.9877

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: 0.001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Auprc
No log	1.0	131	0.1594	0.9313	0.9696
No log	2.0	262	0.0829	0.9847	0.9894
No log	3.0	393	0.0571	0.9809	0.9877
0.0806	4.0	524	0.0699	0.9847	0.9894
0.0806	5.0	655	0.0744	0.9809	0.9877

Framework versions

• PEFT 0.11.1
• Transformers 4.40.2
• Pytorch 2.3.0+cu118
• Datasets 2.19.2
• Tokenizers 0.19.1