sammylina/alx_final_portfolio

Amharic hand-written character recognition using Deep Learning # Amharic hand-written character recognition This is ALX final webstack portfolio project that can recognize Amharic hand-written characters using **Convolutional Neural Network** architecture, **Fast API** backend and **React** frontend. > The ML model has **89%** accuracy on test dataset Project Demo Video ## Installation Clone the project ```bash git clone github.com ``` Setup backend (FastAPI) ```bash cd app # This is backend of the application` pip install -r requirements.txt uvicorn main:app --reload ``` Setup Frontend (React) ```bash cd web_app npm install npm run dev ``` You can see the application running on `localhost:5173` Model training > CNN_Amharic ## Usage Send a request using `curl` ```bash curl {host}:{backend_port}/predict -X POST -F "req=@{path_to_img}.png" ``` Backend API response ##