BatchNormalization

BatchNormalization - 1

Version

• name: BatchNormalization (GitHub)

• domain: main

• since_version: 1

• function:

• support_level: SupportType.COMMON

• shape inference: False

This version of the operator has been available since version 1.

Summary

Attributes

• consumed_inputs - INTS (required) : legacy optimization attribute.

• epsilon - FLOAT : The epsilon value to use to avoid division by zero, default is 1e-5f.

• is_test - INT : If set to nonzero, run spatial batch normalization in test mode, default is 0.

• momentum - FLOAT : Factor used in computing the running mean and variance.e.g., running_mean = running_mean * momentum + mean * (1 - momentum), default is 0.9f.

• spatial - INT : If true, compute the mean and variance across all spatial elements If false, compute the mean and variance across per feature.Default is 1.

Inputs

• X (heterogeneous) - T:

• scale (heterogeneous) - T:

• B (heterogeneous) - T:

• mean (heterogeneous) - T:

• var (heterogeneous) - T:

Outputs

Between 1 and 5 outputs.

• Y (heterogeneous) - T:

• mean (optional, heterogeneous) - T:

• var (optional, heterogeneous) - T:

• saved_mean (optional, heterogeneous) - T:

• saved_var (optional, heterogeneous) - T:

Type Constraints

• T in ( tensor(double), tensor(float), tensor(float16) ): Constrain input and output types to float tensors.