I have trained my model to classify CIFAR10 images using ResNets, Regularization and Data Augmentation in PyTorch, but the problem is I am not able to achieve 90% accuracy in five minutes as the trainer is claiming in video. I got 90% accuarcy in 18 minutes using exactly same parameters as trainer selected. What could be the possible reasons? Please help.
It depends on the platform and the GPU/CPU you are using to train your model.