[object Object]

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

• Description: Uses the global mean value and standard deviation reciprocal as the operator input to perform BatchNorm on x. This operator is an element-level BatchNorm operation function used to normalize input data in specific scenarios. Compared with , aclnnBatchNormElemt may be adjusted for specific hardware or optimization requirements.

• Formula:

  $$y = \frac{(x-E[x])}{\sqrt{Var(x)+ ε}} * weight + bias$$

  The relationship between the standard deviation and variance is as follows:

  $$\frac{1}{S} = \frac{1}{\sqrt{Var(x) + eps}}$$

[object Object]

Each operator has calls. First, aclnnBatchNormElemtGetWorkspaceSize is called to obtain the input parameters and compute the required workspace size based on the process. Then, aclnnBatchNormElemt 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 types of [object Object], [object Object], [object Object], [object Object], [object Object], and [object Object] do not support 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 compute:
  • aclnnBatchNormElemt defaults to a non-deterministic implementation. You can call aclrtCtxSetSysParamOpt to enable deterministic compute.

[object Object]

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