This project explores parallelizing the forward pass of a simple fully connected neural network using OpenMP and CUDA.
Final report here: Final Report
This project explores parallelizing the forward pass of a simple fully connected neural network using OpenMP and CUDA.
Final report here: Final Report