[object Object]

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

• API description: Applies the nearest neighbor interpolation algorithm to upsample the input signal composed of several input channels. If the input shape is (N, C, D, H, W), the output shape is (N, C, outputSize[0], outputSize[1], outputSize[2]).
• Formula:$$d_{src} = min(floor((d_{dst} + 0.5) / scalesD), D - 1),scalesD = outputSize[0] / D$$ $$h_{src} = min(floor((h_{dst} + 0.5) / scalesH), H - 1),scalesH = outputSize[1] / H$$ $$w_{src} = min(floor((w_{dst} + 0.5) / scalesW), W - 1),scalesW = outputSize[2] / W$$ $$out(N, C, d_{dst},h_{dst}, w_{dst}) = self(N, C, d_{src},h_{src}, w_{src})$$

[object Object]

Each operator has calls. First, aclnnUpsampleNearestExact3dGetWorkspaceSize is called to obtain the input parameters and compute the required workspace size based on the process. Then, aclnnUpsampleNearestExact3d is called to perform computation.

[object Object]

• Parameters:

  [object Object]

  • [object Object]Atlas inference products[object Object]:

    The data types of [object Object] and [object Object] support only FLOAT32 and FLOAT16.

• 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]

• 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 $outputSize = [floor(self\_D * scalesD), floor(self\_H * scalesH), floor(self\_W * scalesW)]$.
• Deterministic computing:
  • aclnnUpsampleNearestExact3d defaults to a deterministic implementation.

[object Object]

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