Skip to content

glouwa/kandinsky-pattern-generator

BranchesTags
 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

103 Commits
src
src
 
 
.gitignore
.gitignore
 
 
Kandinsky.ipynb
Kandinsky.ipynb
 
 
README.md
README.md
 
 
app-kandinsky-pattern-generator.zip
app-kandinsky-pattern-generator.zip
 
 
generate_kandinsky_patterns.ipynb
generate_kandinsky_patterns.ipynb
 
 
inception.ipynb
inception.ipynb
 
 
kandinsky.tar.gz
kandinsky.tar.gz
 
 
kandinsky_color_large.tar.gz
kandinsky_color_large.tar.gz
 
 
kandinsky_color_small.tar.gz
kandinsky_color_small.tar.gz
 
 
kandinsky_in_colab_Datagenerator_from_Frame.ipynb
kandinsky_in_colab_Datagenerator_from_Frame.ipynb
 
 
kandinsky_in_colab_circles_and_stars_count_red_stars.ipynb
kandinsky_in_colab_circles_and_stars_count_red_stars.ipynb
 
 
kandinsky_in_colab_circles_only_count_red.ipynb
kandinsky_in_colab_circles_only_count_red.ipynb
 
 
kandinsky_multilabel.ipynb
kandinsky_multilabel.ipynb
 
 
kandinsky_multilabel_circles_and_stars_count_red,_count_stars,_count_red_stars.ipynb
kandinsky_multilabel_circles_and_stars_count_red,_count_stars,_count_red_stars.ipynb
 
 
kandinsky_multilabel_circles_and_stars_count_red,_count_yellow.ipynb
kandinsky_multilabel_circles_and_stars_count_red,_count_yellow.ipynb
 
 
kandinsky_multilabel_circles_only_count_red_and_yellow_circles.ipynb
kandinsky_multilabel_circles_only_count_red_and_yellow_circles.ipynb
 
 
keras_test.ipynb
keras_test.ipynb
 
 
lib.py
lib.py
 
 
lstm_calculator.ipynb
lstm_calculator.ipynb
 
 
more_than.tar.gz
more_than.tar.gz
 
 
vgg16-2.ipynb
vgg16-2.ipynb
 
 
vgg16.ipynb
vgg16.ipynb
 
 
vgg16_count.ipynb
vgg16_count.ipynb
 
 
vgg16_more.ipynb
vgg16_more.ipynb
 
 

Repository files navigation

Based on the Kandinsky Pattern Generator this project trains a VGG16 network using tensorflow. Tests with output spaces are visualized with iNNvestigate, a Layer-wise Relevance Propagation toolbox.

Selected notebooks:

Train on 2 properites

Train on 3 properites

About

Layer-wise Relevance Propagation (iNNvestigate) evaluation of VGG16 trained on Kandinsky Figures.

colab.research.google.com/github/glouwa/app-kandinsky-pattern-generator/blob/master/kandinsky_multilabel_circles_and_stars_count_red%2C_count_yellow.ipynb

Resources

Readme
Activity

Stars

1 star

Watchers

1 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 99.9%
  • Python 0.1%