Cast

Applicability

Product	Supported
Atlas 350 Accelerator Card	x
Atlas A3 training product/Atlas A3 inference product	√
Atlas A2 training product/Atlas A2 inference product	√
Atlas 200I/500 A2 inference product	x
Atlas inference product	√
Atlas training product	√

Function Usage

Converts an input tensor to the specified data type.

Prototype

const aclTensor *Cast(const aclTensor *self, op::DataType dstDtype, aclOpExecutor *executor)

Parameters

Parameter	Input/Output	Description
self	Input	Input tensor to be converted. The data type can be FLOAT16, FLOAT, DOUBLE, BFLOAT16, INT8, UINT8, INT16, UINT16, INT32, UINT32, INT64, UINT64, BOOL, COMPLEX64, or COMPLEX 128. The data format can be ND. NOTE: BFLOAT16 applies to the following product models: Atlas A2 training product/Atlas A2 inference product Atlas A3 training product/Atlas A3 inference product
dstDtype	Input	Destination type of the converted tensor. The data type can be FLOAT16, FLOAT, DOUBLE, BFLOAT16, INT8, UINT8, INT16, UINT16, INT32, UINT32, INT64, UINT64, BOOL, COMPLEX64, or COMPLEX 128. NOTE: BFLOAT16 applies to the following product models: Atlas A2 training product/Atlas A2 inference product Atlas A3 training product/Atlas A3 inference product
executor	Input	Operator executor, containing the operator computation process.

Returns

A tensor of dstDtype.

Constraints

None

Examples

// Standard format for creating an OpExecutor
auto uniqueExecutor = CREATE_EXECUTOR();

auto selfCasted = self;
// When self is of the Boolean type, use the Cast API to convert it to the uint8 type for integer computation.
if (self->GetDataType() == op::DataType::DT_BOOL) {
  selfCasted = l0op::Cast(self, op::DataType::DT_UINT8, uniqueExecutor.get());
  CHECK_RET(selfCasted != nullptr, ACLNN_ERR_PARAM_NULLPTR);
}

Parent topic: Basic Kernel Function APIs