set_as

Description

Sets a Tensor.

Prototype

set_as(value, dst_offset=0, src_offset=None)

Parameters

**Table 1** Parameter description
Parameter	Input/Output	Description
value	Input	Value to set. Supported data types: A Scalar. An Expr. A Tensor. Can have only one element. Otherwise, the first element of the Tensor is used. Must be in the following scope: Atlas training products: scope_ubuf Atlas inference product AI Core: scope_ubuf or scope_gm Atlas inference product Vector Core: scope_ubuf or scope_gm SoC (TsnsC): scope_ubuf or scope_gm Atlas 200I/500 A2 inference products: scope_ubuf or scope_gm
dst_offset	Input	Reserved and not recommended.
src_offset	Input	Reserved and not recommended.

Applicable Models

Atlas training products

Atlas inference product AI Core

Atlas inference product Vector Core

Atlas A2 training products/Atlas A2 inference products

Atlas 200I/500 A2 inference products

Restrictions

To set a Tensor to a Scalar value, they must be of the same data type.
The destination Tensor must be in the scope_ubuf scope.
Only the first element of the destination Tensor is changed.
Setting a Tensor of type int/uint to an Expr value is supported only when the Expr's Scalar is of type int or uint and the Expr's immediate is of type int or float.
For the Atlas training products, setting a Scalar of type float16 to an Expr value is not supported.

For the Atlas inference product AI Core, setting a Scalar of type float16 to an Expr value is not supported.

For the Atlas inference product Vector Core, setting a Scalar of type float16 to an Expr value is not supported.

For the Atlas A2 training products/Atlas A2 inference products, setting a Scalar of type float16 to an Expr value is not supported.

For the Atlas 200I/500 A2 inference products, setting a Scalar of type float16 to an Expr value is not supported.

Returns

None

Example

from tbe import tik
tik_instance = tik.Tik()
data_A = tik_instance.Tensor("float16", (128,), name="data_A", scope=tik.scope_ubuf)
data_B = tik_instance.Tensor("float16", (128,), name="data_A", scope=tik.scope_ubuf)
data_A[0].set_as(data_B[0:1])

Parent topic: Tensor Management