I am working on adversarial attacks. Some of the attacks I am interested in are written using PyTorch and some are based on TensorFlow. Thus I need to have the trained network in both PyTorch and TensorFlow and they must be exactly similar to have a fair comparison.
How can I have a PyTorch network in TensorFlow while I can do differentiating to generate adversary.
Thanks for your help in advance
You would have to reimplement the model in the desired framework, as I don’t think there is a way to use TF models in PyTorch and vice versa.
Thanks for your response.