RetrieveTPUEmbeddingADAMParametersGradAccumDebug

public final class RetrieveTPUEmbeddingADAMParametersGradAccumDebug

Retrieve ADAM embedding parameters with debug support.

An op that retrieves optimization parameters from embedding to host memory. Must be preceded by a ConfigureTPUEmbeddingHost op that sets up the correct embedding table configuration. For example, this op is used to retrieve updated parameters before saving a checkpoint.

Nested Classes

class RetrieveTPUEmbeddingADAMParametersGradAccumDebug.Options Optional attributes for RetrieveTPUEmbeddingADAMParametersGradAccumDebug

Public Methods

static RetrieveTPUEmbeddingADAMParametersGradAccumDebug.Options
config (String config)
static RetrieveTPUEmbeddingADAMParametersGradAccumDebug
create ( Scope scope, Long numShards, Long shardId, Options... options)
Factory method to create a class wrapping a new RetrieveTPUEmbeddingADAMParametersGradAccumDebug operation.
Output <Float>
gradientAccumulators ()
Parameter gradient_accumulators updated by the ADAM optimization algorithm.
Output <Float>
momenta ()
Parameter momenta updated by the ADAM optimization algorithm.
Output <Float>
parameters ()
Parameter parameters updated by the ADAM optimization algorithm.
static RetrieveTPUEmbeddingADAMParametersGradAccumDebug.Options
tableId (Long tableId)
static RetrieveTPUEmbeddingADAMParametersGradAccumDebug.Options
tableName (String tableName)
Output <Float>
velocities ()
Parameter velocities updated by the ADAM optimization algorithm.

Inherited Methods

Public Methods

public static RetrieveTPUEmbeddingADAMParametersGradAccumDebug create ( Scope scope, Long numShards, Long shardId, Options... options)

Factory method to create a class wrapping a new RetrieveTPUEmbeddingADAMParametersGradAccumDebug operation.

Parameters
scope current scope
options carries optional attributes values
Returns
  • a new instance of RetrieveTPUEmbeddingADAMParametersGradAccumDebug

public Output <Float> gradientAccumulators ()

Parameter gradient_accumulators updated by the ADAM optimization algorithm.

public Output <Float> momenta ()

Parameter momenta updated by the ADAM optimization algorithm.

public Output <Float> parameters ()

Parameter parameters updated by the ADAM optimization algorithm.

public static RetrieveTPUEmbeddingADAMParametersGradAccumDebug.Options tableName (String tableName)

public Output <Float> velocities ()

Parameter velocities updated by the ADAM optimization algorithm.