get_reference_grid() assumes single precision

[urlicht]

layer_util.get_reference_grid() assumes 32 bit single precision when generating the grid. More specifically, in the following line, the tensor is casted to 32 bit: https://github.com/DeepRegNet/DeepReg/blob/650a2f1a88ad3c6932be305d6a98a36e26feedc6/deepreg/model/layer_util.py#L49

This becomes problematic when trying to use the grid (e.g. warping) with half-precision / 16 bit tensor.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.