mindspore.ops.BNTrainingReduce

class mindspore.ops.BNTrainingReduce(*args, **kwargs)[source]

For the BatchNorm operation this operator updates the moving averages for training and is used in conjunction with BNTrainingUpdate.

Inputs:

  • x (Tensor) - A 4-D Tensor with float16 or float32 data type. Tensor of shape \((N, C, A, B)\).

Outputs:

  • sum (Tensor) - A 1-D Tensor with float32 data type. Tensor of shape \((C,)\).

  • square_sum (Tensor) - A 1-D Tensor with float16 or float32 data type. Tensor of shape \((C,)\).

Raises

  • TypeError – If x is not a Tensor.

  • TypeError – If dtype of x is neither float16 nor float32.

Supported Platforms:

Ascend

Examples

>>> x = Tensor(np.ones([128, 3, 32, 3]), mindspore.float32)
>>> bn_training_reduce = ops.BNTrainingReduce()
>>> output = bn_training_reduce(x)
>>> print(output)
(Tensor(shape=[3], dtype=Float32, value=
[ 1.22880000e+04, 1.22880000e+04, 1.22880000e+04]), Tensor(shape=[3], dtype=Float32, value=
[ 1.22880000e+04, 1.22880000e+04, 1.22880000e+04]))