[object Object]

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

• Description: Obtains the coordinate data of the input x based on indices.
• Formula:$$out = self[indices]$$

[object Object]

Each operator has calls. First, aclnnIndexGetWorkspaceSize is called to obtain the input parameters and compute the required workspace size based on the process. Then, aclnnIndex is called to perform computation.

[object Object]

• Parameters

  [object Object]
  • [object Object]Atlas training products[object Object] and [object Object]Atlas inference products[object Object]: The data type cannot be BFLOAT16.

• Returns

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

  The first-phase API implements input parameter verification. The following errors may be thrown.

  [object Object]

[object Object]

• Parameters

  [object Object]

• Returns

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

[object Object]

• Deterministic computation:
  • aclnnIndex defaults to a deterministic implementation.
• If self is a non-zero-dimensional tensor, the number of tensors in indices must be less than or equal to the number of dimensions of self. If self is a zero-dimensional tensor, indices can contain only one tensor.
• The shapes of tensors in indices must be the same or meet the broadcast relationship, and the values in each tensor cannot exceed the size of the corresponding dimension in self. Otherwise, unpredictable behavior may occur, such as out-of-bounds address.
• If indices is of the BOOL type, the shape of each tensor in indices must be the same as that of the corresponding dimension in self, and the shapes of the new tensors generated by filtering each tensor using its own Boolean index must meet the broadcast relationship.

[object Object]

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