This project involved the implementation of a fully connected neural network
from the ground up, with the CIFAR10 dataset
utilized for model training. Key concepts learned throughout the process included linear regression
, forward propagation
, backward propagation
, and vectorization
. Additionally, an extra assignment involved the implementation of Convolutional Neural Networks
.
The main purpose of this project is understanding the importance of Vectorization and how it can drastically decrease the processing time for any neural network architecture.