oneflow.Tensor.is_offloaded

Tensor.is_offloaded()bool

Determine whether the tensor has been moved to CPU memory and the CUDA device memory has been released.