ConvertToTensor

Description

Converts host-side data of various types to host-side aclTensor objects.

Prototype

aclTensor *ConvertToTensor(const aclIntArray *value, DataType dataType)

aclTensor *ConvertToTensor(const aclBoolArray *value, DataType dataType)

aclTensor *ConvertToTensor(const aclFloatArray *value, DataType dataType)

aclTensor *ConvertToTensor(const aclFp16Array *value, DataType dataType)

aclTensor *ConvertToTensor(const aclBf16Array *value, DataType dataType)

aclTensor *ConvertToTensor(const aclScalar *value, DataType dataType)

template<typename T> 
aclTensor *ConvertToTensor(const T *value, uint64_t size, DataType dataType)

Parameters

Returns

Converted tensor on the host.

Constraints

The input parameter pointer must not be null.

Data type restrictions on parameters of each API:

• API 1: ConvertToTensor(const aclIntArray *value, DataType dataType)

  Parameter	Restriction
  value	aclIntArray, int64_t data type
  dataType	DT_FLOAT, DT_FLOAT16, DT_BF16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8, DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, and DT_DOUBLE

• API 2: ConvertToTensor(const aclBoolArray *value, DataType dataType)

  Parameter	Restriction
  value	aclBoolArray, bool data type
  dataType	DT_FLOAT, DT_FLOAT16, DT_BF16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8, DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, and DT_DOUBLE

• API 3: ConvertToTensor(const aclFloatArray *value, DataType dataType)

  Parameter	Restriction
  value	aclFloatArray, float data type.
  dataType	DT_FLOAT, DT_FLOAT16, DT_BF16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8, DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, and DT_DOUBLE

• API 4: ConvertToTensor(const aclFp16Array *value, DataType dataType)

  Parameter	Restriction
  value	aclFp16Array, fp16_t data type.
  dataType	DT_FLOAT, DT_FLOAT16, DT_BF16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8, DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, and DT_DOUBLE

• API 5: ConvertToTensor(const aclBf16Array *value, DataType dataType)

  Parameter	Restriction
  value	aclBf16Array, bfloat16 data type
  dataType	DT_FLOAT, DT_FLOAT16, DT_BF16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8, DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, and DT_DOUBLE

• API 6: ConvertToTensor(const T *value, uint64_t size, DataType dataType)

  Parameter	Restriction
  value	The template parameter T can be int64_t, uint64_t, int32_t, uint32_t, int8_t, uint8_t, int16_t, uint16_t, float, double, bool, char, op::bfloat16, or op::fp16_t.
  size	A uint64_t value that indicates the number of elements
  dataType	DT_FLOAT, DT_FLOAT16, DT_BF16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8, DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, and DT_DOUBLE

• API 7: ConvertToTensor(const aclScalar *value, DataType dataType)

  Parameter	Restriction
  value	aclScalar: DT_BOOL, DT_INT8, DT_UINT8, DT_INT16, DT_UINT16, DT_INT32, DT_UINT32, DT_INT64, DT_UINT64, DT_FLOAT, DT_DOUBLE, DT_FLOAT16, DT_BF16, DT_COMPLEX64, DT_COMPLEX128, DT_FLOAT8_E5M2, DT_FLOAT8_E4M3FN, DT_FLOAT8_E8M0, DT_FLOAT6_E3M2, DT_FLOAT6_E2M3, DT_FLOAT4_E2M1, DT_FLOAT4_E1M2, and DT_HIFLOAT8
  dataType	DT_BOOL, DT_INT8, DT_UINT8, DT_INT16, DT_UINT16, DT_INT32, DT_UINT32, DT_INT64, DT_UINT64, DT_FLOAT, DT_DOUBLE, DT_FLOAT16, DT_BF16, DT_COMPLEX64, DT_COMPLEX128, DT_FLOAT8_E5M2, DT_FLOAT8_E4M3FN, DT_FLOAT8_E8M0, DT_FLOAT6_E3M2, DT_FLOAT6_E2M3, DT_FLOAT4_E2M1, DT_FLOAT4_E1M2, and DT_HIFLOAT8

Examples

1
2
3
4
5
6
7

// Convert an aclScalar and an int64_t to tensors on the host.
void Func(aclOpExecutor *executor) {
    int64_t val = 5;
    aclScalar *scalar = executor->AllocScalar(val);
    aclTensor *tensor = executor->ConvertToTensor(scalar, DT_INT64);
    tensor = executor->ConvertToTensor(&val, 1, DT_INT64);
}