This application demonstrates the implementation of a convolutional neural network (CNN) programmed with Python class structures. A menu user interface is provided by PySide6, while the network model is created by Keras.

The application includes training, testing, and tuning tasks for fitting the Keras network model. Matplotlib is employed to make plots of accuracy, loss, confusion matrix, and sample dataset images. Gridsearchcv allows optimizaton across combinations of hyper-parameters. Long running tasks are delegated to threads, with signal feedback to update GUI progess bars.

Read the project description in Python-Keras-CNN.pdf found in the docs folder.

Follow the installation instructions in uv-instructions.txt contained in the docs folder.