Update low-level functional transforms with value_range

Following https://github.com/pytorch/vision/pull/5500#discussion_r816599553 we may want to update low-level functional transforms with value_range argument to avoid implicit hard-coded max range definition:

• 255 for uint8
• 1.0 for float

Today this is done for

• adjust_hue https://github.com/pytorch/vision/blob/95d418970e6dbf2e4d928a204c4e620da7bccdc0/torchvision/transforms/functional_tensor.py#L210-L211

• all ops using _blend: https://github.com/pytorch/vision/blob/95d418970e6dbf2e4d928a204c4e620da7bccdc0/torchvision/transforms/functional_tensor.py#L259-L262

We can introduce new argument value_range and use it explicitly for these ops. In general we can think of value_range as a tuple (min, max) which would cover majority of imagery where channels ranges are similar. There could be however other type of images (e.g. think of non-RGB color spaces or particular imagery) where value ranges could vary per channel, thus we may need to represent value_range as a list of 2-tuples: [(min_1, max_1), (min_2, max_2), …]

cc @bjuncek @datumbox

[vfdev-5]

I think the reason is that torch script does not support global. We could do however something like:

def convert_image_dtype(image: torch.Tensor, dtype: torch.dtype = torch.float) -> torch.Tensor:

    # ...
    _max_values: Dict[torch.dtype, int] = {
        torch.uint8: 255,
        torch.int8: 127,
        torch.int16: int(2 ** 15),
        torch.int32: int(2 ** 31),
        torch.int64: int(2 ** 63),
        ...
    }

    input_max = _max_values[image.dtype]

@datumbox do you think it worth a change ?

EDIT: updated 256 -> 255

[datumbox]

I think we should close this for now.