[object Object]

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

• Function: Performs backpropagation of .
• Formula:$$scalesD = outputSize[0] / inputSize[2]$$ $$scalesH = outputSize[1] / inputSize[3]$$ $$scalesW = outputSize[2] / inputSize[4]$$ $$gradInput(N, C, floor ( scalesD * ( D + 0.5 )), floor ( scalesH * ( H + 0.5 )), floor ( scalesW * ( W+ 0.5 ))) += gradOutput( N, C, D, H ,W)$$

[object Object]

Each operator has calls. First, aclnnUpsampleNearestExact3dBackwardGetWorkspaceSize is called to obtain the input parameters and compute the required workspace size based on the process. Then, aclnnUpsampleNearestExact3dBackward is called to perform computation.

[object Object]

• Parameters:

  [object Object]

• Returns:

  aclnnStatus: 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:

  aclnnStatus: status code. For details, see .

[object Object]

• The shape constraints of [object Object] and [object Object] are as follows:

  • The value of each dimension is less than or equal to 2^20.

  • The N and C axes of [object Object] must be the same as those of [object Object].

  • The memory usage must be less than 60 GB. The memory size can be calculated according to the following formula:

    $$N * C * (gradOut\_D * gradOut\_H * gradOut\_W + gradInput\_D * gradInput\_H * gradInput\_W + gradOut\_D * gradOut\_H * gradInput\_W + gradOut\_D * gradInput\_H * gradInput\_W) * sizeof(float) < 60 * 1024 * 1024 * 1024$$

    Where,

    • N indicates the N axis of the input and output.
    • C indicates the C axis of the input and output.

  • N * C * gradOut_D * gradOut_H < 2^31

  • gradInput_W * gradInput_H < 2^31

• If the data formats of gradOut and gradInput are not NCDHW or NDHWC, the input data is processed as NCDHW by default.
• The upscaling factor for input data must be less than or equal to 50. That is, $outputSize[0]/height H of output shape$, $outputSize[1]/width W of output shape$, and $outputSize[2]/depth D of output shape$ must be less than or equal to 50.
• Either outputSize or scalesD/scalesH/scalesW must be used.
  • If the value of scalesD, scalesH, or scalesW is less than or equal to 0, the value of outputSize is used.
  • If the values of scalesD, scalesH, and scalesW are greater than 0, the values of scalesD, scalesH, and scalesW are used, and $outputSize = [floor(inputSize\_D * scalesD), floor(inputSize\_H * scalesH), floor(inputSize\_W * scalesW)]$.
• Deterministic computing:
  • aclnnUpsampleNearestExact3dBackward defaults to a non-deterministic implementation. You can call aclrtCtxSetSysParamOpt to enable deterministic computing.

[object Object]

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