[object Object]

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

• Description:

  Computes the gradient of the BatchNorm operation during backpropagation and performs reduction operations on some intermediate results to optimize the compute efficiency. The calculation result is as follows:

  • Computes the gradient $\frac{\partial l}{\partial γ}$ of the loss function l with respect to the scaling weight γ.
  • Computes the gradient $\frac{\partial l}{\partial β}$ of the loss function l with respect to the offset β.
  • Deduces the intermediate quantities sumDy and sumDyXmu required for computing $\frac{\partial l}{\partial x_i}$, by using the deviation d[object Object]yi[object Object] of the loss function l with respect to the output y[object Object]i[object Object], where $\frac{\partial l}{\partial x_i}$ is the gradient of the loss function l with respect to the input x[object Object]i[object Object] of the corresponding layer.

• Formula:

  $$gradWeight = \frac{\partial l}{\partial γ} = \sum^m_{i=0} \frac{\partial l}{\partial y_i} \cdot \hat{(x_i)} = \frac{1}{{\sqrt{σ^2_B + eps}}} \cdot \sum^m_{i=0} \frac{\partial l} {\partial y_i} \cdot (x_i-μ_B)$$ $$gradBias = \frac{\partial l}{\partial β} = \sum^m_{i=0} \frac{\partial l}{\partial y_i}$$ $$sumDy = sum(l, y_i) = \displaystyle \sum^m_{i=0} \frac{\partial l}{\partial y_i}$$ $$sumDyXmu = sum(l, y_i, x_i, μ_B) = \displaystyle \sum^m_{i=0} \frac{\partial l}{\partial y_i} \cdot (x_i-μ_B)$$

[object Object]

Each operator has calls. First, aclnnBatchNormReduceBackwardGetWorkspaceSize is called to obtain the input parameters and compute the required workspace size based on the process. Then, aclnnBatchNormReduceBackward 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], [object Object], [object Object], [object Object], and [object Object] cannot be 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]

• If any input is an empty tensor, the output is an empty tensor.
• Deterministic compute:
  • aclnnBatchNormReduceBackward defaults to a deterministic implementation.

[object Object]

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