[object Object]

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

• Description: Performs backpropagation of .
• Formula:

$$out = gradOutput - sum(gradOutput) \cdot exp(output)$$ [object Object]

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

[object Object]

• Parameters:

  [object Object]
  • For [object Object]Atlas inference products[object Object] and [object Object]Atlas training products[object Object], the data types FLOAT16 and FLOAT32 are supported, and no additional space needs to be allocated. Other data types are supported by the automatic casting capability, but additional space needs to be allocated.

• 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 compute:
  • aclnnLogSoftmaxBackward defaults to a deterministic implementation.

[object Object]

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