Skip to content

Conversation

borisgin
Copy link

@borisgin borisgin commented Sep 9, 2015

New layer, which produces Confusion Matrix (https://en.wikipedia.org/wiki/Confusion_matrix ):

  • input: the same as for accuracy layer
  • output: 2D matrix ( Blob) [CxC], where C is number of classes

I also modified Solver::Test() to print 2D outputs of test net into log file.

The example of using new layer:

  • examples/cifar10/cifar_cm_train_test.prototxt

The example of Visualization of confusion matrix:

  • plot_conf_matrrix.py

@ducha-aiki
Copy link
Contributor

Nice!

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Specific to your environment. Has to be fixed.

@Austriker Austriker mentioned this pull request Mar 30, 2016
@journey-wang
Copy link

not merged ?

@shelhamer
Copy link
Member

Closing as better addressed by a Python layer. There are many metrics out there. Thanks @borisgin for the metric effort all the same.

@shelhamer shelhamer closed this Apr 14, 2017
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants