[object Object]

[object Object][object Object]undefined[object Object]

Description: Extracts the tensor $out$ from the input tensor $self$ based on the specified dimension $axes$, range $[starts, ends]$, and step $steps$. $starts[i]$ and $ends[i]$ can be set to values other than $[0, self.shape[axes[i]]]$. After the values are set, they are converted to valid values using the following formula (assume $self.shape[axes[i]] = N$):

$$starts[i] = \begin{cases} &0, & if\;starts[i] < -N \\ &N, & if\;starts[i] >= N\\ &(starts[i]+N) \% N, & otherwise \end{cases}$$ $$ends[i] = \begin{cases} &N, & if\; ends[i] >= N\\ &starts, & else\;if\; (ends[i]+N)\%N < starts[i] \\ &(ends[i]+N)\%N, & otherwise\\ \end{cases}$$

$out.shape$ is the same as $self.shape$ except on the [object Object] axis.

$$out.shape[axes[i]] = ⌊\frac{ends[i] - starts[i] + steps[i] - 1}{steps[i]}⌋$$ [object Object]

Each operator has calls. First, [object Object] is called to obtain the workspace size required for computation and the executor that contains the operator computation process. Then, [object Object] is called to perform computation.

• [object Object]
• [object Object]

[object Object]

• Parameters:

  • [object Object] (aclTensor*, computation input): aclTensor on the device, $self$ in the formula. The shape can have 1 to 8 dimensions. The data types of [object Object] and [object Object] are the same. are supported. The can be ND.

    • [object Object]Atlas training products[object Object]: The data type can be FLOAT, FLOAT16, INT32, INT64, INT8, UINT8, or BOOL.
    • [object Object]Atlas A2 training products/Atlas A2 inference products[object Object] and [object Object]Atlas A3 training products/Atlas A3 inference products[object Object]: The data type can be FLOAT, FLOAT16, INT32, INT64, INT8, UINT8, BOOL, or BFLOAT16.

  • [object Object] (aclIntArray*, computation input): aclIntArray on the host, $starts$ in the formula, indicating the start index of the slice position. [object Object], [object Object], [object Object], and [object Object] have the same number of elements.

  • [object Object] (aclIntArray*, computation input): aclIntArray on the host, $ends$ in the formula, indicating the end index of the slice position. [object Object], [object Object], [object Object], and [object Object] have the same number of elements.

  • [object Object] (aclIntArray*, computation input): aclIntArray on the host, $axes$ in the formula, indicating the axis for slicing. [object Object], [object Object], [object Object], and [object Object] have the same number of elements. The elements of [object Object] must be within the range [–self.dim(), self.dim() – 1], and the axes must be unique.

  • [object Object] (aclIntArray*, computation input): aclIntArray on the host, $steps$ in the formula, indicating the slice step. [object Object], [object Object], [object Object], and [object Object] have the same number of elements. The elements of [object Object] are all positive integers.

  • [object Object] (aclTensor*, computation output): aclTensor on the device, $out$ in the formula. The shape can have 0 to 8 dimensions and must meet the deduction rules in the function description. The data types of [object Object] and [object Object] are the same. are supported. The can be ND.

    • [object Object]Atlas training products[object Object]: The data type can be FLOAT, FLOAT16, INT32, INT64, INT8, UINT8, or BOOL.
    • [object Object]Atlas A2 training products/Atlas A2 inference products[object Object] and [object Object]Atlas A3 training products/Atlas A3 inference products[object Object]: The data type can be FLOAT, FLOAT16, INT32, INT64, INT8, UINT8, BOOL, or BFLOAT16.

  • [object Object] (uint64_t*, output): size of the workspace to be allocated on the device.

  • [object Object] (aclOpExecutor**, output): operator executor, containing the operator computation process.

• Returns:

  [object Object]: status code. For details, see .

[object Object]

• Parameters:

  • [object Object] (void*, input): address of the workspace to be allocated on the device.

  • [object Object] (uint64_t, input): size of the workspace to be allocated on the device, which is obtained by calling [object Object].

  • [object Object] (aclOpExecutor*, input): operator executor, containing the operator computation process.

  • [object Object] (aclrtStream, input): stream for executing the task.

• Returns:

  [object Object]: status code. For details, see .

[object Object]

• Deterministic compute:
  • [object Object] defaults to a deterministic implementation.

[object Object]

The following example is for reference only. For details, see .