[object Object]

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

• Description: Performs backpropagation of . The sigmoid gradient input is computed based on the sigmoid backpropagation gradient and forward output.
• Formula:$$out = {\frac{1} {1+{e}^{-input}}}$$ $$grad\_input = grad\_output * \sigma(x) * (1 - \sigma(x))$$

$\sigma(x)$ is the forward output of the sigmoid function, and $\sigma(x)*(1-\sigma(x))$ is the derivative of the sigmoid function.

[object Object]

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

[object Object]

• Parameters:

  [object Object]
  • [object Object]Atlas inference products[object Object] and [object Object]Atlas training products[object Object]: The data type can be FLOAT, FLOAT16, DOUBLE, COMPLEX64, or COMPLEX128.

• 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:
  • aclnnSigmoidBackward defaults to a deterministic implementation.

[object Object]

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