tf.tpu.experimental.embedding.PositionalWeightedCombiner

Note: This API is new and only available via pip install tf-nightly.

Positional weighted combiner for TPU embeddings on SparseCore.

View aliases

Compat aliases for migration

See Migration guide for more details.

tf.compat.v1.tpu.experimental.embedding.PositionalWeightedCombiner

tf.tpu.experimental.embedding.PositionalWeightedCombiner(
    max_valency: int,
    initializer: init_ops_v2.Initializer,
    custom_optimizer: tf.tpu.experimental.embedding.CustomOptimizer
) -> Any

This class specifies a positional weighted sum combiner over the TPU embedding lookup results. It also supports a custom optimizer computation to be performed on the positional weights.

Args
max_valency	The valency cap of the corresponding feature. This value is also used as the number of positional weights.
initializer	The initializer for the positional weights.
custom_optimizer	CustomOptimizer encapsulating the custom optimizer computation, learning rate, slot names, slot initializers, and hyperparameters.

Attributes
hyperparameters

Methods

__eq__

View source

__eq__(
    other: Any
) -> Union[Any, bool]

Return self==value.