Combine AdaBoost and ResNet in pytorch

I want to combine Adaboost with ResNet. To get better hyperparameters. My goal is to remove the final layer of ResNet, which is a FC 512, and to bring it to Adaboost, and to use AdBoost instead of the decision tree of the FC network to repeat it until the convergence, and then get the final output of the network.
Can you help me to do this?

Did you end up completing this? I am curious how you did it and whether it worked.