[BUG] FlopsProfiler upsample flops compute bug

Describe the bug
the upsample flops compute code :
`def _upsample_flops_compute(*args, **kwargs):

scale_factor = kwargs.get('scale_factor', None)
if scale_factor is None and len(args) > 2:
    scale_factor = args[2]
assert scale_factor is not None, "either size or scale_factor should be defined"

flops = input.numel()
if isinstance(scale_factor, tuple) and len(scale_factor) == len(input):
    flops *= int(_prod(scale_factor))
else:
    flops *= scale_factor**len(input)
return flops, 0`

len(scale_factor) == len(input) -> len(scale_factor) == len(input.size())
flops *= scale_factor ** len(input) -> flops *= scale_factor ** (len(input.size())-1)

upsample flops compute in torchstat: https://github.com/Swall0w/torchstat/blob/master/torchstat/compute_flops.py#L83

Locate the _upsample_flops_compute entry point used by FlopsProfiler, then compare its argument and dimensionality handling with torchstat's torchstat/compute_flops.py at the referenced section. Verify the calculation for tuple and scalar scale_factor cases, including spatial dimensions, and add or run the relevant profiler checks if available.