nn_visualizer

Visualizing Tensorflow neural networks.

Please see the examples.ipynb notebook for an example on how to use the package.

Currently supporting:

• Input layers:
  • For dense networks with the shape (x_size)
  • For convolution with the shape (x_size, y_size, n_channels)
• Convolutional Network Layers:
  • Conv2D
  • MaxPooling2D
  • AveragePooling2D
• Dense Layers:
  • Dense (also as output)
  • Flatten

Changes

0.7

• Fixed cases, in which the layers descriptions have overlapped.

0.6

• Supporting dense layers as input.

0.5

• Changed build tool to flit, now available as package from PyPi under the name nnvisualizertf.

0.4

• Changed build configuration.

0.3

• Changed build tool to hachtling.

0.2

• Renaming to comply with the python naming conventions.

0.1

• Initial version.