BoostedTreesMakeStatsSummary

public final class BoostedTreesMakeStatsSummary

Makes the summary of accumulated stats for the batch.

The summary stats contains gradients and hessians accumulated into the corresponding node and bucket for each example.

Public Methods

Output<Float>
asOutput()
Returns the symbolic handle of a tensor.
static BoostedTreesMakeStatsSummary
create(Scope scope, Operand<Integer> nodeIds, Operand<Float> gradients, Operand<Float> hessians, Iterable<Operand<Integer>> bucketizedFeaturesList, Long maxSplits, Long numBuckets)
Factory method to create a class wrapping a new BoostedTreesMakeStatsSummary operation.
Output<Float>
statsSummary()
output Rank 4 Tensor (shape=[#features, #splits, #buckets, 2]) containing accumulated stats put into the corresponding node and bucket.

Inherited Methods

Public Methods

public Output<Float> asOutput ()

Returns the symbolic handle of a tensor.

Inputs to TensorFlow operations are outputs of another TensorFlow operation. This method is used to obtain a symbolic handle that represents the computation of the input.

public static BoostedTreesMakeStatsSummary create (Scope scope, Operand<Integer> nodeIds, Operand<Float> gradients, Operand<Float> hessians, Iterable<Operand<Integer>> bucketizedFeaturesList, Long maxSplits, Long numBuckets)

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

Parameters
scope current scope
nodeIds int32 Rank 1 Tensor containing node ids, which each example falls into for the requested layer.
gradients float32; Rank 2 Tensor (shape=[#examples, 1]) for gradients.
hessians float32; Rank 2 Tensor (shape=[#examples, 1]) for hessians.
bucketizedFeaturesList int32 list of Rank 1 Tensors, each containing the bucketized feature (for each feature column).
maxSplits int; the maximum number of splits possible in the whole tree.
numBuckets int; equals to the maximum possible value of bucketized feature.
Returns
  • a new instance of BoostedTreesMakeStatsSummary

public Output<Float> statsSummary ()

output Rank 4 Tensor (shape=[#features, #splits, #buckets, 2]) containing accumulated stats put into the corresponding node and bucket. The first index of 4th dimension refers to gradients, and the second to hessians.