This repository is based on an assignment made for a Deep Learning (National University of Ireland, Galway) module.
Parts of this project are based on my previous work DeepNeural-JavNet.
The project was written using PyCharm but should work just as any other notebook provided the dependencies are available.
You can install the dependencies using anaconda
conda env create --file .\environment.yml
or manually using the below list of packages.
- python = 3.9
- ipython = 8.4.0
- cupy = 11.0.0
- numpy = 1.23.1
- pandas = 1.4.3
- matplotlib = 3.5.3
- jupyterlab = 3.4.5
The CIFAR-10 dataset was used for training and evaluation, it can be downloaded from this link. Simply unpack the contents of the downloaded file to the project folder.
By default, the horse and truck classes are used by this project.
Alternatively you can select a different dataset such as the CIFAR-100 dataset on the CIFAR homepage.
The remaining two datasets blobs300.csv and circles600.csv are already included in the repository since their size is relatively small.