Skip to content

Latest commit

 

History

History
13 lines (9 loc) · 556 Bytes

File metadata and controls

13 lines (9 loc) · 556 Bytes

VisualNet

This is a visual simulation of a working Deep Neural Network (Multilayer Perceptron to be precise).
The goal of this simulation is to demonstrate an almost 100% GPU usage:

The computation and rendering both will be covered by the same GPU and data will stay inside VRAM.

It's broken. The scope is big and I am lazy..

Started out this project with Java, then moved to Go.
The low level libraries are too slow and painful to compile in windows, I think I regret switching to Go.

Here is a screenshot: visualnet.png