How can I train my computationally intensive deep learning model?
I’m trying to train a convolutional neural network on the FER 2013 dataset on Google Colaboratory. I did 50 epochs but the first epoch is taking about 2 hours to finish and then the runtime stops executing somewhere between the second epoch because it says that runtime execution was interrupted. I have spent the entire week trying to fix this and retrain my model but I am unable to. I even bought 100 compute units provided by Google Colab Pro but it is taking the same amount of time. Can someone help? I’m working on a really important project with a fast approaching deadline.