This repository contains a boilerplate code for getting started with the Face Recognition Udemy course by Anis Koubaa. The code provides a starting point for building your own face recognition application and ensures that you have all the necessary requirements to get started.
First, make sure to use Python 3.7, which is the version used to develop the code. Although it can work with other Python version, I do not give any guarantees to face issues.
You can find the list of requirements in the following files:
requirements_face.txt: A pip requirements file that includes all the necessary dependencies.environment_face.yaml: A conda environment file that includes the necessary packages.
Note that it is important to install the same version of TensorFlow and other packages to ensure that the face recognition application works correctly.
To use the boilerplate code, follow these steps:
- Clone the repository to your local machine.
- Place the file
facenet_keras_128.h5inside themodelsfolder. This file is necessary to run the embedding function for the face.
pip install -r requirements_face.txt
or
conda env create -f environment_face.yaml
- Run the
test_environment.pyfile to test that everything is working correctly.
Unfortunately, Google Colab no longer supports older versions of TensorFlow such as TensorFlow 2.0. Therefore, it is required to work locally on your machine to use this boilerplate code. However, if you have successfully installed all the necessary requirements on Colab, then the code should work fine on Colab as well.
We hope that this boilerplate code provides a helpful starting point for building your own face recognition application. If you have any questions or issues, please feel free to open an issue on GitHub.