2023-t1-finalproject-ann created by GitHub Classroom
This repository contains the implementation of a simple neural network in Rust, along with parallel versions and a comparison with a Python version using TensorFlow. The network is trained on the MNIST dataset for digit recognition.
Implemented from scratch in Rust: Gain a deep understanding of neural network architecture and training algorithms.
Simple and parallel versions: Compare the performance of different implementations.
Comparison with TensorFlow: See how the Rust implementation performs against a popular deep-learning library.
The simple version of the neural network achieved the fastest training time, followed by TensorFlow and the parallel versions.
Resources MNIST dataset: https://github.com/iamavieira/handwritten-digits-mnist
TensorFlow: https://www.tensorflow.org/