[object Object]

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

• Description: For each dimension in the given dimension [object Object], if all elements in the dimension of the input tensor evaluate to [object Object], [object Object] is returned. Otherwise, [object Object] is returned. If [object Object] is [object Object], the output Tensor maintains the same number of dimensions as the input. Otherwise, the dimensions specified by [object Object] are reduced, and the output tensor has [object Object] fewer dimensions than the input.
• Given an input tensor with shape $(A \times B \times C \times D)$ and [object Object] set to [0, 2]. If [object Object] is [object Object], the output tensor shape will be $(B \times D)$, reducing the input by two dimensions. If [object Object] is [object Object], the output tensor shape will be $(1 \times B \times 1 \times D)$, maintaining the same dimensionality as the input.

[object Object]

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

[object Object]

• Parameters:

  [object Object]
  • [object Object]Atlas inference products[object Object], [object Object]Atlas training products[object Object], and [object Object]Atlas 200I/500 A2 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:
  • [object Object] defaults to a deterministic implementation.

[object Object]

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