llama2_13B_LORA_FOR_CLASSIFICATION

This model is a fine-tuned version of meta-llama/Llama-2-13b-hf on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5708

• Balanced Accuracy: 0.7079

• Accuracy: 0.7530

• Micro F1: 0.7530

• Macro F1: 0.6771

• Weighted F1: 0.7669

• Classification Report: precision recall f1-score support

       0       0.89      0.79      0.83       857
       1       0.44      0.63      0.52       232
  

  accuracy 0.75 1089 macro avg 0.67 0.71 0.68 1089

weighted avg 0.79 0.75 0.77 1089

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: 24
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Accuracy	Balanced Accuracy	Classification Report	Validation Loss	Macro F1	Micro F1	Weighted F1
0.4853	2.0	522	0.7750	0.7297	precision recall f1-score support

       0       0.90      0.81      0.85       857
       1       0.48      0.65      0.55       232

accuracy                           0.78      1089

macro avg 0.69 0.73 0.70 1089 weighted avg 0.81 0.78 0.79 1089 | 0.5482 | 0.7009 | 0.7750 | 0.7864 | | 0.4116 | 3.0 | 783 | 0.7668 | 0.7182 | precision recall f1-score support

       0       0.89      0.80      0.84       857
       1       0.47      0.63      0.54       232

accuracy                           0.77      1089

macro avg 0.68 0.72 0.69 1089 weighted avg 0.80 0.77 0.78 1089 | 0.5497 | 0.6903 | 0.7668 | 0.7786 | | 0.3224 | 4.0 | 1044 | 0.5708 | 0.7079 | 0.7530 | 0.7530 | 0.6771 | 0.7669 | precision recall f1-score support

       0       0.89      0.79      0.83       857
       1       0.44      0.63      0.52       232

accuracy                           0.75      1089

macro avg 0.67 0.71 0.68 1089 weighted avg 0.79 0.75 0.77 1089 |

Framework versions

• PEFT 0.11.1
• Transformers 4.41.2
• Pytorch 2.3.0+cu118
• Datasets 2.20.0
• Tokenizers 0.19.1