I came across pytorch’s
__torch_function__ and now I was wondering if there is an equivalent in tensorflow.
For some context:
__torch_function__ is a method that allows all kinds of objects to be treated like torch.Tensors. For example if we call torch.stack([t1, t2, t3]) with t1, t2, t2 instances of
OurTensor class, this will still work, as long as the class
OurTensor implements the
__torch_function__() to tell torch how to handle our tensors.