code:start | CILVR Lab @ NYUOverFeat is an image recognizer and feature extractor built around a convolutional network. The OverFeat convolutional net was trained on the ImageNet 1K dataset. It participated in the ImangeNet Large Scale Recognition Challenge 2013 under the name “OverFeat NYU”. This release provides C/C++ code to run the network and output class probabilities or feature vectors. It also includes a webcam-based
