rishavranaut / Llama2_7B_fact_updates

Llama2_7B_fact_updates

This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6150

• Balanced Accuracy: 0.7155

• Accuracy: 0.7328

• Micro F1: 0.7328

• Macro F1: 0.6683

• Weighted F1: 0.7522

• Classification Report: precision recall f1-score support

       0       0.90      0.75      0.81       857
       1       0.42      0.69      0.52       232
  

  accuracy 0.73 1089 macro avg 0.66 0.72 0.67 1089

weighted avg 0.80 0.73 0.75 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: 16
• 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

       0       0.87      0.81      0.84       857
       1       0.45      0.55      0.49       232

accuracy                           0.76      1089

macro avg 0.66 0.68 0.67 1089 weighted avg 0.78 0.76 0.77 1089 | 0.5476 | 0.6674 | 0.7585 | 0.7673 | | 0.6444 | 0.6079 | 200 | 0.7355 | 0.7125 | precision recall f1-score support

       0       0.89      0.75      0.82       857
       1       0.42      0.67      0.52       232

accuracy                           0.74      1089

macro avg 0.66 0.71 0.67 1089 weighted avg 0.79 0.74 0.75 1089 | 0.6209 | 0.6687 | 0.7355 | 0.7541 | | 0.5961 | 0.9119 | 300 | 0.7420 | 0.7292 | precision recall f1-score support

       0       0.90      0.75      0.82       857
       1       0.44      0.71      0.54       232

accuracy                           0.74      1089

macro avg 0.67 0.73 0.68 1089 weighted avg 0.80 0.74 0.76 1089 | 0.5747 | 0.6797 | 0.7420 | 0.7608 | | 0.5961 | 1.0 | 329 | 0.5454 | 0.7302 | 0.7585 | 0.7585 | 0.6906 | 0.7738 | precision recall f1-score support

       0       0.90      0.78      0.84       857
       1       0.46      0.68      0.55       232

accuracy                           0.76      1089

macro avg 0.68 0.73 0.69 1089 weighted avg 0.81 0.76 0.77 1089 | | 0.5028 | 2.0 | 658 | 0.4805 | 0.7297 | 0.7897 | 0.7897 | 0.7104 | 0.7974 | precision recall f1-score support

       0       0.89      0.83      0.86       857
       1       0.51      0.62      0.56       232

accuracy                           0.79      1089

macro avg 0.70 0.73 0.71 1089 weighted avg 0.81 0.79 0.80 1089 | | 0.4393 | 3.0 | 987 | 0.7759 | 0.6933 | 0.6657 | 0.6657 | 0.6191 | 0.6956 | precision recall f1-score support

       0       0.90      0.65      0.75       857
       1       0.36      0.74      0.49       232

accuracy                           0.67      1089

macro avg 0.63 0.69 0.62 1089 weighted avg 0.79 0.67 0.70 1089 | | 0.3784 | 4.0 | 1316 | 0.5960 | 0.7205 | 0.7456 | 0.7456 | 0.6784 | 0.7628 | precision recall f1-score support

       0       0.90      0.76      0.83       857
       1       0.44      0.68      0.53       232

accuracy                           0.75      1089

macro avg 0.67 0.72 0.68 1089 weighted avg 0.80 0.75 0.76 1089 | | 0.2856 | 5.0 | 1645 | 0.6150 | 0.7155 | 0.7328 | 0.7328 | 0.6683 | 0.7522 | precision recall f1-score support

       0       0.90      0.75      0.81       857
       1       0.42      0.69      0.52       232

accuracy                           0.73      1089

macro avg 0.66 0.72 0.67 1089 weighted avg 0.80 0.73 0.75 1089 |

Framework versions

• PEFT 0.11.1
• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1