How to use `torch.arange` to create a list of tensors?

# Something like this:
#
import torch
starts = torch.tensor([2, 3, 5]).long()
ends = torch.tensor([10, 20, 32]).long()

a = torch.arange(starts, ends)
# a is a list full of tensors

As per the pytorch documentation, this is format for torch.arange torch.arange( *start=0* , *end* , *step=1* , *out=None* , *dtype=None* , *layout=torch.strided* , *device=None* , *requires_grad=False* ), where start,end and step are numbers.
So you cannot give a tensor as input to arrange. What exactly do you want to do with starts and ends?