mindspore.hal.contiguous_tensors_handle.combine_tensor_list_contiguous

mindspore.hal.contiguous_tensors_handle.combine_tensor_list_contiguous(tensor_list, enable_mem_align=True)[source]

Return a contiguous memory handle where contiguous memory has been requested and slicing functionality is provided.

Parameters

tensor_list (list[Tensor], tuple[Tensor]) – The tensor list to be stored.
enable_mem_align (bool, optional) – Whether to enable the memory alignment function. False is not supported. Default True .

Returns

ContiguousTensorsHandle, a manager with contiguous memory.

Examples

>>> import numpy as np
>>> import mindspore as ms
>>> from mindspore import Tensor
>>> from mindspore.hal.contiguous_tensors_handle import combine_tensor_list_contiguous
>>> x = Tensor(np.array([1, 2, 3]).astype(np.float32))
>>> y = Tensor(np.array([4, 5, 6]).astype(np.float32))
>>> handle = combine_tensor_list_contiguous([x, y], True)
>>> print(handle[0].shape)
[1]
>>> print(handle[1: 3].asnumpy())
[2, 3]
>>> print(output.slice_by_tensor_index(0, 1).asnumpy())
[1, 2, 3]