The data archive of SEER which is publicly available on Zenodo requires roughly 3 GBs of disk space. Moreover, all neural models have been trained on a Virtual Machine with a GPU on Google Cloud Platform. Using a GPU will significantly improve the performance of model. The details of our GPU is given in RQ4 of the paper.
We have developed and tested SEER using Ubuntu 18.04 LTS (Bionic Beaver), Python 3.6.9, and Java 8. SEER has some software dependencies listed below:
- matplotlib
- numpy
- pandas
- scikit learn
- tables
- torch
- tqdm
Please install all dependencies by executing the following:
pip3 install -r requirements.txt