This project is a machine learning model that categorizes images into two categories. After it has finished training, it is used to take screenshots and classify them as either productive or nonproductive.
This project was built and tested using Python 3.8. To run this program, you MUST use a version of Python 3.
TO RUN: use "python3 main.py"
TO CONFIGURE: open main.py and set variables in cnn.run() to your preferences
- epochs(any int): number of times the CNN goes through all the training data
- load_model(T/F): whether or not the program loads a previously trained model. If this parameter is true, the parameters save_model, train, plot, data_prep, and clear_data will automatically be false even if they are manually set to true
- save_model(T/F): whether or not the model is saved after training
- train(T/F): whether or not the program trains
- evaluate(T/F): whether or not the program tests the model on test data
- plot(T/F): whether or not the program plots the loss and accuracy over epochs
- data_prep(T/F): whether or not the program starts web scraping for training and testing data
- clear_data(T/F): whether or not the program resets all the data and asks for new ones
All modules can be installed via pip
- numpy
- time
- cv2
- pyautogui
- os
- requests
- tqdm
- bs4
- pickle
- matplotlib
- PIL