Skip to content

Feature and class visualization with deep neural networks in tensorflow. Contains deepdream, style transfer, receptive field visualization, convolutional filter visualization, etc.

Notifications You must be signed in to change notification settings

timsainb/tensorflow-2-feature-visualization-notebooks

Repository files navigation

Tensorflow 2 Deep feature visualization

These are a series of self-contained notebook that show you how to visualize classes, features, and receptive fields, and a few other things in a neural network in tensorflow. I also wrote a short set of blog posts about this code.

Notebooks

  • Visualizing classes Open In Colab
  • Visualizing features Open In Colab
  • Reconstructing images from layer activations Open In Colab
  • Visualizing deepdream style transfer Open In Colab

Here are a few things I implemented:

Class visualization

class_visualization

Feature visualization

feature_visualization

feature_visualization

Receptive field visualization

rf_viz

rf_viz_zoom

Image reconstruction from filter activation

recon

Style transfer (starry night)

starry_night

style_transfer

About

Feature and class visualization with deep neural networks in tensorflow. Contains deepdream, style transfer, receptive field visualization, convolutional filter visualization, etc.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published