numpy-cnn

Convolution network implemented in NumPy

Things to do

• Weight adjustments (momentum, variable learning rate, ...)
• Evaluation (precision, recall, confusion matrix, ...)
• Visualization

Layout

• Layer classes and Model are defined in layers.py
• Cnn class is defined in cnn.py
• Training loop for Cnn class is in cnn.ipynb and main.py
• Tests to check for bugs are in test_cnn and test_dense