[object Object]

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

Description: Performs backpropagation of .

[object Object]

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

[object Object]

• Parameters

  [object Object]

• 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

  • aclnnGroupNormSwishGrad defaults to a non-deterministic implementation. You can call aclrtCtxSetSysParamOpt to enable deterministic computation.

• Input shape restrictions:

  1. numGroups must be greater than 0.
  2. C must be exactly divided by group.
  3. The number of elements in dy is equal to $N * C * HxW$.
  4. The number of elements in mean is equal to $N * group$.
  5. The number of elements in rstd is equal to $N * group$.
  6. The number of elements in x is equal to $N * C * HxW$.
  7. The number of elements in gamma is equal to C.
  8. The number of elements in beta is equal to C.
  9. The ratio of C to group cannot exceed 4000.

[object Object]

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