Deepfake Face Detector
A neural network model made in Python trained to classify and real pictures of human faces against fake faces generated by AI. Created using sklearn and pytorch along with standard Python libraries such as pandas and numpy. Adam optimizer and cross-entropy loss were initially used which resulted in higher metrics, but were ultimately scrapped due to runtime constraints on Google Colab. Created in collaboration with two other students.
Python notebook can be found here: https://github.com/ChauAbdul786/DeepfakeFaceDetector
Dataset can be found here: https://www.kaggle.com/datasets/sachchitkunichetty/rvf10k