Digit recognition Web app using Flask and TensorFlow. This project was created as a part of 4th year module of Software Development course in GMIT.
For this web application we used Flask framework and Tensorflow library for machine learning.
- Python Tested on version 3.6
- Flask
pip install flask - Numpy library
pip istall numpy - Tensorflow
pip install tensorflow - PIL
pip install pillow
- Navigate to folder you downloaded the project.
- Make sure all dependencies are installed.
- Run
python main.py - Navigate your browser to
localhost:5000
Prediction model was trained using MNIST data set.
Fully trained model is located in /mlearning/data/ folder under convolutional.ckpt file.
Model will predict with high probability canvas or file input. If prediction is incorrect, user can correct it and app will adjust, retrain the model.
