The goal of this project was to fine tune various large language models on the task of research paper summarization. This repository contains code for training models and for testing them.
For running pytorch models execute the following
conda create --name torch python=3.9
conda activate torch
pip install torch --index-url https://download.pytorch.org/whl/cu118
pip install datasets transformers matplotlib PyPDF2
For running tensorflow models execute the following
conda create --name tf python=3.9
conda activate tf
conda install -c conda-forge cudatoolkit=11.2 cudnn=8.1.0
pip install "tensorflow<2.11"
pip install datasets transformers matplotlib PyPDF2
Inside the working directory create a folder named data. Inside this folder create a cache folder and an experiments folder. Then for each model create a folder with the appropriate name(which can be found in the model file) in the experiments folder.
To compile the models is relatively simple. Just run the following
conda activate environment_name
python3 model_name.py