SBoM

Software Bill-of-Materials by Chain-of-Experts

Setup

Use conda to manage environments

conda create -n sem_seg python=3.8
conda activate sem_seg
# Install pytorch with cuda
pip install numpy scikit-learn # tools
pip install transformers
pip install tensorboard

Data Preparation

Download the data.tar.gz from release page and extract the tarball as following

mkdir data
tar -xzvf data.tar.gz -C data

The data has already been separated into train, validation, and test. Each file has already been preprocessed into the following format List[(segment:str, label_name:str, source_string:str)]

Running the model

Check experiments folder for scripts to train/test the model.