- We are gonna challenge a two-layer fully-conncted network to optimize some hyperparameters so that we receive near-perfect score on MNIST in 3 epochs
- Additionally we optimize the CUDA operations so we get the most out of our GPU and reach this result as soon as possible