Using a pretrained model in Pytorch

I am basically working on a problem where in I have to use a pretrained CNN to extract features from a given dataset and then feed this into a RNN to aggregate all these, I am pretty new to pytorch, can someone help me with me how to do this.