Convolutional neural network implementations for image classification.
pytorch_cnn.py— PyTorch CNN for SVHN street view house numbers (Conv2d, MaxPool, RMSprop)tf2/res_net_cifar10/— TF2 ResNet on CIFAR-10 with data augmentation and learning rate schedulingtf_cnn.py,echo.py,edge_detection.py,blr.py— NumPy signal processing and edge detection samples
The SVHN .mat files can be downloaded from http://ufldl.stanford.edu/housenumbers/