Construct a lighter and smaller Residual Network (ResNet) architecture.Train your network to perform image classication by minimizng the cross-entropy between the network’s prediction
and the CIFAR-10 targets.
As the training proceeds, plot the loss value and the classication accuracy on training and validation set,
and choose the iteration with the least error.
All applicants go through a series of tests that check their level of English and knowledge of formatting styles. The applicant is also required to present a sample of writing to the Evaluation Department. If you wish to find out more about the procedure, check out the whole process.