llama3.2-qlora-rank8-hurtmongoose

This model is a fine-tuned version of meta-llama/Llama-3.2-3B-Instruct on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4098
• Accuracy: 0.9033
• Mcc: 0.8073
• Recall: 0.8833
• Fpr: 0.0767
• Auc: 0.9575

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: 0.0002
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Use paged_adamw_8bit with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Mcc	Recall	Fpr	Auc
0.9055	1.0	313	0.3082	0.8983	0.7967	0.8967	0.1	0.9625
1.4793	2.0	626	0.2815	0.915	0.8302	0.9033	0.0733	0.9707
0.3112	3.0	939	0.7091	0.9	0.8052	0.8433	0.0433	0.9661
0.1549	4.0	1252	1.4908	0.8767	0.7678	0.78	0.0267	0.9626
0.0021	5.0	1565	1.3141	0.8883	0.7864	0.81	0.0333	0.9647

Framework versions

• PEFT 0.18.0
• Transformers 4.57.3
• Pytorch 2.9.1+cu128
• Datasets 2.15.0
• Tokenizers 0.22.1