TransposeOperation

Applicable Products

Hardware Model	Supported
Atlas 350 accelerator cards	√
Atlas A3 inference products/Atlas A3 training products	√
Atlas A2 training products/Atlas A2 inference products	√
Atlas training products	x
Atlas inference products	√
Atlas 200I/500 A2 inference products	x

Description

Rearranges the input tensors and transposes them in multiple dimensions.

Figure 1 TransposeOperation

Definition

struct TransposeParam {
    SVector<int32_t> perm;
    uint8_t rsv[8] = {0};
};

Parameters

Member	Type	Default Value	Description
perm	SVector<int32_t>	-	Rearrangement result of the input dimensions. For example, if the shape of the input tensor is [a,b,c] and that of perm is [0,2,1], the first and second dimensions are exchanged, and the shape of the output tensor is [a,c,b]. Ensure that the input of perm is correct and has the same dimension as input x.
rsv[8]	uint8_t	{0}	Reserved

Member

Type

Default Value

Description

perm

SVector<int32_t>

Rearrangement result of the input dimensions.

For example, if the shape of the input tensor is [a,b,c] and that of perm is [0,2,1], the first and second dimensions are exchanged, and the shape of the output tensor is [a,c,b].

Ensure that the input of perm is correct and has the same dimension as input x.

rsv[8]

uint8_t

{0}

Reserved

Input

Parameter	Dimension	Data Type	Format
x	[dim_0, dim_1, ..., dim_n]	float16/bf16/int64/int8/int32	ND

Output

Parameter	Dimension	Data Type	Format
output	The dimension is determined by the parameter.	float16/bf16/int64/int8/int32 The data type is the same as that of x.	ND

Parameter

Dimension

Data Type

Format

output

The dimension is determined by the parameter.

float16/bf16/int64/int8/int32

The data type is the same as that of x.

Constraints

Only the float16 data type is supported on Atlas 200I/500 A2 inference products.

Parent topic: atb/infer_op_params.h