torch.Tensor.is_pinned

Tensor.is_pinned()

Returns true if this tensor resides in pinned memory. By default, the device pinned memory on will be the current accelerator.