Lucy | real time deep learning visualization | neural network parameters visualization in real time

Lucy visualizes the parameters of neural networks in real time. As the neural network trains, its parameters are captured and streamed through a flask API towards the visualization system.
For Lucy to be able to adapt to the number of parameters of the network, subsampling is used to reduce the number of points. The resulting parameters are separated by layers and packaged in visual grids. The vertical coordinate of the representation is the parameter value.

In this real time deep learning visualization, the Mnist dataset is used together with a network composed of 4 main layers, two convolutional layers and two fully connected layers.

Real time neural network training visualization
======================================
Lucy is a neural network parameters visualization system, a neural network visualization tool that allows you to capture deep learning weights visualization scenarios. Navigate a neural network visualized by Lucy and explore deep learning parameters visualization productions.

This is the first version of Lucy. New updates will be posted in the future.
Created by Javier Ideami

CONNECT:
- Subscribe to this YouTube channel

[varvaravetrova]

I absolutely love this! Going to show tomorrow to my students!