Neural Plot

A Library for visualizing Neural Networks of the TensorFlow/Keras models.

Visit library page - https://github.com/Rajsoni03/neuralplot

This Library is working on Matplotlib.

This works well for smaller models but takes longer for larger models.

Currently Under Development

Try this out (Interactive plot is not supported in colab.)

Single Layer Graph Plotting (Development Notebooks/1-Neural Net.ipynb)

for i, c in enumerate(['r', 'g', 'b']):
    x,y,z = mat3dl(32, 32, 3, (i,i+1))
    ax.scatter(x, z, y, c=c, marker='o')
plt.show()

Multilayer Layer Graph Plotting (Development Notebooks/2-Neural Net.ipynb)

layers = [(32, 32, 3, 'rgb', '.'),(16, 16, 5, 'r', '+'),(20,20,3,'y', 's'),(5,5,5),(10,1,1,'b', '^')]
neuralplot(layers)

2D Network Plot (Development Notebooks/3-Neural Net.ipynb)

layers = [(1, 32, 1, 'b','o'),(1, 32, 1, 'r', '.'),(1, 8, 1, 'g', '.'),(1, 12, 1, 'g', '.'),(1, 4, 1, 'm', 's')]
neuralplot(layers)

Plot Simple 3D Network (Development Notebooks/3-Neural Net.ipynb)

Plot Complex 3D Networks (Development Notebooks/3-Neural Net.ipynb)

layers = [(4, 4, 3, 'rgb','s'),(2, 2, 2, 'y', 'o'),(3, 3, 3, 'g', '.'),(5, 5, 1, 'b', '.'),(2, 1, 1, 'b', '^')]
neuralplot(layers)

layers = [(8, 8, 1, 'r','o'),(4, 4, 2, 'g', '.'),(2, 2, 3, 'b', '.'),(5, 5, 2, 'y', '.'),(2, 1, 1, 'c', 's')]
neuralplot(layers)

TensorFlow/Keras Models to 3D Networks Graph (Example Notebook.ipynb)

from neuralplot import neuralplot
neuralplot(model=model, grid=True, connection=True, linewidth=0.1)

neuralplot(model=model, grid=False, connection=True, linewidth=0.1)