tf.keras.layers.experimental.SyncBatchNormalization

View source on GitHub

Normalize and scale inputs or activations synchronously across replicas.

tf.keras.layers.experimental.SyncBatchNormalization(
    axis=-1, momentum=0.99, epsilon=0.001, center=True, scale=True,
    beta_initializer='zeros', gamma_initializer='ones',
    moving_mean_initializer='zeros', moving_variance_initializer='ones',
    beta_regularizer=None, gamma_regularizer=None, beta_constraint=None,
    gamma_constraint=None, renorm=False, renorm_clipping=None, renorm_momentum=0.99,
    trainable=True, adjustment=None, name=None, **kwargs
)

Applies batch normalization to activations of the previous layer at each batch by synchronizing the global batch statistics across all devices that are training the model. For specific details about batch normalization please refer to the tf.keras.layers.BatchNormalization layer docs.

If this layer is used when using tf.distribute strategy to train models across devices/workers, there will be an allreduce call to aggregate batch statistics across all replicas at every training step. Without tf.distribute strategy, this layer behaves as a regular tf.keras.layers.BatchNormalization layer.

Example usage:

strategy = tf.distribute.MirroredStrategy()

with strategy.scope():
  model = tf.keras.Sequential()
  model.add(tf.keras.layers.Dense(16))
  model.add(tf.keras.layers.experimental.SyncBatchNormalization())

axis Integer, the axis that should be normalized (typically the features axis). For instance, after a Conv2D layer with data_format="channels_first", set axis=1 in BatchNormalization.
momentum Momentum for the moving average.
epsilon Small float added to variance to avoid dividing by zero.
center If True, add offset of beta to normalized tensor. If False, beta is ignored.
scale If True, multiply by gamma. If False, gamma is not used. When the next layer is linear (also e.g. nn.relu), this can be disabled since the scaling will be done by the next layer.
beta_initializer Initializer for the beta weight.
gamma_initializer Initializer for the gamma weight.
moving_mean_initializer Initializer for the moving mean.
moving_variance_initializer Initializer for the moving variance.
beta_regularizer Optional regularizer for the beta weight.
gamma_regularizer Optional regularizer for the gamma weight.
beta_constraint Optional constraint for the beta weight.
gamma_constraint Optional constraint for the gamma weight.
renorm Whether to use Batch Renormalization (https://arxiv.org/abs/1702.03275). This adds extra variables during training. The inference is the same for either value of this parameter.
renorm_clipping A dictionary that may map keys 'rmax', 'rmin', 'dmax' to scalar Tensors used to clip the renorm correction. The correction (r, d) is used as corrected_value = normalized_value * r + d, with r clipped to [rmin, rmax], and d to [-dmax, dmax]. Missing rmax, rmin, dmax are set to inf, 0, inf, respectively.
renorm_momentum Momentum used to update the moving means and standard deviations with renorm. Unlike momentum, this affects training and should be neither too small (which would add noise) nor too large (which would give stale estimates). Note that momentum is still applied to get the means and variances for inference.
trainable Boolean, if True the variables will be marked as trainable.

Call arguments:

  • inputs: Input tensor (of any rank).
  • training: Python boolean indicating whether the layer should behave in training mode or in inference mode.
    • training=True: The layer will normalize its inputs using the mean and variance of the current batch of inputs.
    • training=False: The layer will normalize its inputs using the mean and variance of its moving statistics, learned during training.

Input shape:

Arbitrary. Use the keyword argument input_shape (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.

Output shape:

Same shape as input.