The Dog App is a single-page application (SPA) that I initially developed as a personal challenge, which eventually became an invaluable learning experience in the field of machine learning and transfer learning. The core functionality of the app revolves around a fine-tuned image classification model that I created. This model utilizes the convolutional layers from MobileNetV3 and incorporates my own custom classifier. It has been trained specifically on images of various dog breeds, enabling it to accurately recognize and classify the breed of a dog from an uploaded image.
Developed using React JS, the Dog App showcases my proficiency in front-end development. Despite its intentionally minimalistic styling, the app delivers a seamless user experience with its intuitive interface. It employs responsive design principles to ensure optimal performance across different devices and screen sizes.
Furthermore, the Dog App leverages TensorFlow JS, a powerful library for machine learning in JavaScript, to execute the image classification on the client-side. This eliminates the need for server-side processing, resulting in faster and more efficient classification results.
To ensure the reliability and accuracy of the app, I implemented unit testing using Jest, a popular JavaScript testing framework. This approach guarantees the functionality of the app's core features and enhances its overall quality.
Some of the key skills demonstrated in this project include:
HTML
CSS
React JS
Firebase
Responsive Design
TensorFlow JS
Transfer Learning
Keras
Jest
Unit Testing