How alicegupta24/fitai works
Project type and purpose
• FitAI is a Python-based full-stack web application (built on Streamlit) that serves as a virtual shopping assistant.
• It persists user history in a local SQLite database.
File structure
• Top level:
• README.md – project overview, tech stack, models
• ai.py, updatedai.py – entry points for the Streamlit UI and orchestration of inference pipelines
• Implicit components (not in file tree but described in README):
• Model artifacts (.h5, .pkl files) for CNN, NN, Random Forest, recommendation engine
• SQLite file for data storage
• Potential data directory for synthetic data generation & training scripts
README for additional insights
• Core libraries & frameworks: Streamlit (UI), TensorFlow & Keras, scikit-learn, MediaPipe, OpenCV
• External integrations: Google Shopping API, local SQLite for ‘User History’
• Offline ML pipeline: synthetic data generator → model training → model artifact storage