bart-base-lora / README.md
xshubhamx's picture
Upload folder using huggingface_hub
9d2f03f verified
metadata
license: apache-2.0
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - precision
  - recall
base_model: facebook/bart-base
model-index:
  - name: bart-base-lora
    results: []

bart-base-lora

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

  • Loss: 0.6655
  • Accuracy: 0.7963
  • Precision: 0.7841
  • Recall: 0.7963
  • Precision Macro: 0.5968
  • Recall Macro: 0.6325
  • Macro Fpr: 0.0186
  • Weighted Fpr: 0.0179
  • Weighted Specificity: 0.9749
  • Macro Specificity: 0.9847
  • Weighted Sensitivity: 0.7963
  • Macro Sensitivity: 0.6325
  • F1 Micro: 0.7963
  • F1 Macro: 0.6074
  • F1 Weighted: 0.7859

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: 5e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 32
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 15

Training results

Training Loss Epoch Step Validation Loss Accuracy Precision Recall Precision Macro Recall Macro Macro Fpr Weighted Fpr Weighted Specificity Macro Specificity Weighted Sensitivity Macro Sensitivity F1 Micro F1 Macro F1 Weighted
No log 1.0 160 1.2642 0.6313 0.5477 0.6313 0.3009 0.3127 0.0428 0.0400 0.9351 0.9711 0.6313 0.3127 0.6313 0.2941 0.5769
No log 2.0 321 0.8962 0.7119 0.6939 0.7119 0.3937 0.4525 0.0285 0.0281 0.9669 0.9786 0.7119 0.4525 0.7119 0.4107 0.6960
No log 3.0 482 0.8204 0.7196 0.6953 0.7196 0.3974 0.4468 0.0278 0.0271 0.9653 0.9790 0.7196 0.4468 0.7196 0.3998 0.6885
1.2731 4.0 643 0.7519 0.7436 0.7186 0.7436 0.4131 0.4673 0.0244 0.0240 0.9695 0.9809 0.7436 0.4673 0.7436 0.4272 0.7248
1.2731 5.0 803 0.7364 0.7475 0.7524 0.7475 0.6132 0.5050 0.0243 0.0236 0.9679 0.9810 0.7475 0.5050 0.7475 0.4905 0.7286
1.2731 6.0 964 0.7273 0.7514 0.7423 0.7514 0.5784 0.5258 0.0237 0.0231 0.9699 0.9814 0.7514 0.5258 0.7514 0.5150 0.7311
0.7243 7.0 1125 0.6993 0.7645 0.7478 0.7645 0.5498 0.5565 0.0222 0.0215 0.9721 0.9824 0.7645 0.5565 0.7645 0.5453 0.7538
0.7243 8.0 1286 0.6952 0.7769 0.7639 0.7769 0.5682 0.5888 0.0207 0.0201 0.9731 0.9833 0.7769 0.5888 0.7769 0.5700 0.7649
0.7243 9.0 1446 0.6759 0.7823 0.7708 0.7823 0.5764 0.5877 0.0201 0.0195 0.9739 0.9838 0.7823 0.5877 0.7823 0.5699 0.7697
0.6098 10.0 1607 0.6705 0.7847 0.7720 0.7847 0.5899 0.6176 0.0199 0.0192 0.9732 0.9839 0.7847 0.6176 0.7847 0.5935 0.7724
0.6098 11.0 1768 0.6794 0.7909 0.7737 0.7909 0.5882 0.6237 0.0193 0.0185 0.9736 0.9843 0.7909 0.6237 0.7909 0.5988 0.7773
0.6098 12.0 1929 0.6836 0.7909 0.7816 0.7909 0.5973 0.6285 0.0192 0.0185 0.9742 0.9843 0.7909 0.6285 0.7909 0.6034 0.7802
0.5239 13.0 2089 0.6508 0.7932 0.7783 0.7932 0.5965 0.6273 0.0189 0.0183 0.9738 0.9845 0.7932 0.6273 0.7932 0.6046 0.7821
0.5239 14.0 2250 0.6588 0.7963 0.7823 0.7963 0.5957 0.6290 0.0186 0.0179 0.9746 0.9847 0.7963 0.6290 0.7963 0.6055 0.7852
0.5239 14.93 2400 0.6655 0.7963 0.7841 0.7963 0.5968 0.6325 0.0186 0.0179 0.9749 0.9847 0.7963 0.6325 0.7963 0.6074 0.7859

Framework versions

  • Transformers 4.35.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.1