Skip to content

gtanvi58/mass-shooter-classifier

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

mass-shooter-classifier

In this project, we explored advanced fine-tuning methodologies for large language models (LLMs) to enhance their performance in identifying potential mass shooters from social media comments.

Instructions on running the code:

  • Running the ensemble:

    In folder Ensemble_Model_Approach,

    Step 1: Run the following files on colab to train and test on specific trait models that form the ensemble: ensemble_mistral_manifesto.ipynb ensemble_supremacist_manifesto.ipynb ensemble_suicidal_manifesto.ipynb ensemble_terrorist_manifesto.ipynb

    Step 2: Download the generated label files from these codes.

    Step 3: Run the evaluate_ensemble.ipynb file to perform a majority vote and evaluate the model.

  • Running the code for multistage fine tuning on Mistral:

    In folder Multistage_Fine_Tuning_Approach_Mistral,

    This folder contains codes to perform the multistage fine-tuning in various combinations of train and test sets.

    Step 1: Run the fine_tune_on_3_traits.ipynb to fine tune baseline Mistral sequentially on the three traits.

    Step 2: Run Train_manifesto_test_manifesto.ipynb to train the fine tuned model earlier and test on the main dataset.

    Step 3: Run the files TrainOn_traits_TestOn_manifestos_cross_val.ipynb and TrainOn_traits_TestOn_manifestos_n_pair.ipynb to train the models on traits and test them on the final dataset using two different loss functions.

  • Running the code for multistage fine tuning on BERT:

    In folder Multistage_Fine_Tuning_Approach_BERT,

    Run the file Transformer_model_multistage_finetuning.ipynb

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •