convisualize

Visualizations for Convolutional Neural Networks (CNNs) in Pytorch

The corresponding article can be found here!

Requirements:

• Pytorch
• Torchvision
• Numpy
• Matplotlib
• Pillow

Note: In case you don't have a GPU, remove all instances of "cuda" and "cpu" from the notebook before running.

TODO

• Layer Outputs at all layers
• Filter outputs at a given layer
• Filter visualization at a given layer
• Image heatmap using Occlusion
• Image heatmap using Grad Cam
• Class specific saliency maps
• SmoothGrad
• Semantic segmentation using GrabCut
• Visualization of class models (Gradient Ascent)
• Regularization techniques for class models (L2, Clip, Blur, etc.)
• Guided Backprop
• Filter visualization (Gradient Ascent)
• Neural Texture Synthesis
• Deep Dream
• Neural Style Transfer

References

• cs231n
• cs231n Lecture 12
• Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps Karen Simonyan, Andrea Vedaldi, Andrew Zisserman
• Visualizing and Understanding Convolutional Networks Matthew D Zeiler, Rob Fergus
• SmoothGrad: removing noise by adding noise Daniel Smilkov, Nikhil Thorat, Been Kim, Fernanda Viégas, Martin Wattenberg
• Texture Synthesis Using Convolutional Neural Networks Leon A. Gatys, Alexander S. Ecker, Matthias Bethge