spatial_cross_attention

deeplay.blocks.conv.conv2d.spatial_cross_attention(block: ~deeplay.blocks.conv.conv2d.Conv2dBlock, to_channel_last: bool = False, normalization: ~deeplay.external.layer.Layer | ~typing.Type[~torch.nn.modules.module.Module] = <class 'torch.nn.modules.normalization.LayerNorm'>, condition_name: str = 'condition')