This project is a machine learning pipeline that uses a convolutional neural network (CNN) to classify images into three categories: cats, dogs, and birds. By leveraging transfer learning with VGG16, robust data cleaning, and augmentation techniques, the model achieves high accuracy (~94–95%) and strong generalization. Designed for scalability and efficiency, the pipeline is easy to integrate into real-world workflows.
Features
- Data Cleaning: Automated removal of corrupt, empty, or invalid files.
- Transfer Learning: Built on the pre-trained VGG16 architecture for fast, reliable results.
- Data Augmentation: Enhances robustness through image transformations.
- Streamlined Training: Includes EarlyStopping and dynamic learning rate adjustment for efficiency.
Contributors: Isaiah Doan, Owen Downs, Chhi Ly
All rights reserved.
Doc: https://docs.google.com/document/d/1ILK8-M_i4C0ufWhDPROFcpcrN2Z7k_hXrKLDtTgbMN0/edit?usp=sharing