HACNet: End-to-end learning of tabular-to-image converter and convolutional neural network

This repository contains the code for the following paper: https://doi.org/10.1016/j.knosys.2023.111293

Please cite our paper if you use this code for your research.

Directory

.
├── README.md
├── src (source code of HACNet)
├── datasets (Please put datasets here)
└── template (Template images we used)

Requirements

You can utilize Pipfile or requirements.txt to set up the environment.

We checked the code works in the following environment:

• Ubuntu 18.04 LTS
• GPU: NVIDIA A100
• CUDA version 11.5
• Python 3.9
• PyTorch 1.9.0

Usage

You can download datasets from these links. Please preprocess downloaded data and put correctly to read (ref).

You can determine the setting from config.yaml file.

Then, run the program as

$ cd src
$ python main.py