semantic-encodings

Beyond One-Hot-Encoding: Injecting Semantics to Drive Image Classifiers [Conference Paper] | [arXiv Version]

This is not the code used in the paper but a simple notebook to explain the core concepts.

Installation

We assume that you have Python 3.10 and jupyter-lab/jupyter-notebook installed on your system.

1. Clone this repository: git clone https://github.com/S1M0N38/semantic-encodings.git
2. Move inside the repository: cd semantic-encodings
3. Create a virtual environment: python -m venv venv
4. Activate the virtual environment: source venv/bin/activate
5. Install the dependencies: python -m pip install -r requirements.txt
6. Install the kernel: python -m ipykernel install --user --name=semantic-encodings

Usage

We assume that you have followed the installation instructions above.

1. Open the notebook: jupyter-lab semantic-encodings.ipynb or jupyter-notebook semantic-encodings.ipynb
2. Select the kernel: Kernel -> Change kernel -> semantic-encodings

Uninstall

If you need to uninstall this directory and all libraries installed in the virtual environment, simply remove the semantic-encodings directory (If you have followed the installation, the virtual environment directory is inside semantic-encodings).

Moreover, you need to remove the kernel installed. Run the command jupyter kernelspec uninstall semantic-encodings to remove the kernel.

Cite

@article{SemanticsEncPerott2023,
  author = {Perotti, Alan and Bertolotto, Simone and Pastor, Eliana and Panisson, Andr\'{e}},
  eprint = {2308.00607v1},
  month = {8},
  primaryclass = {cs.cv},
  title = {Beyond One-Hot-Encoding: Injecting Semantics to Drive Image Classifiers},
  url = {http://arxiv.org/abs/2308.00607v1},
  year = {2023},
}