During this project, I aimed to visualize a Convolutional Neural Network (CNN) using Processing, a highly effective language for visualization.
The network architecture comprises three convolutional layers and two fully connected layers. For the convolutional layers, I configured a kernel size of 3 by 3, padding of 1, and a stride of 2.
I trained the network using PyTorch and saved the weights and biases. Subsequently, I employed the parsed parameters in the respective Conv2D and MLP layers.
I created a visualization of an MLP during the early stages of my master's degree three years ago (https://youtu.be/WQYCK1YpsjE). However, due to various commitments, such as completing my degree, pursuing a doctorate, and managing other responsibilities, I could not develop the CNN version as intended. I apologize for the delay. :(
For those interested in exploring further, the source codes are readily accessible on GitHub. I hope this visualization enhances your understanding of CNN operations.