Feature Fusion

Hi folks,
I have two different image data-set but related to a same class. I am going to use pre-trained net like resnet34 for both to detect features and then, concatenate those features into a classifier (like the following image). Please advice how can I do it in fastai?