xgboost_lib-sys 3.0.5

/**
 * Copyright 2017-2024, XGBoost contributors
 */
#pragma once
#include <limits>   // for numeric_limits
#include <ostream>  // for ostream

#include "../data/batch_utils.h"   // for DftPrefetchBatches, StaticBatch
#include "gpu_hist/quantiser.cuh"  // for GradientQuantiser
#include "param.h"                 // for TrainParam
#include "xgboost/base.h"          // for bst_bin_t
#include "xgboost/task.h"          // for ObjInfo

namespace xgboost::tree {
struct GPUTrainingParam {
  // minimum amount of hessian(weight) allowed in a child
  float min_child_weight;
  // L2 regularization factor
  float reg_lambda;
  // L1 regularization factor
  float reg_alpha;
  // maximum delta update we can add in weight estimation
  // this parameter can be used to stabilize update
  // default=0 means no constraint on weight delta
  float max_delta_step;
  float learning_rate;
  uint32_t max_cat_to_onehot;
  bst_bin_t max_cat_threshold;

  GPUTrainingParam() = default;

  XGBOOST_DEVICE explicit GPUTrainingParam(const TrainParam& param)
      : min_child_weight(param.min_child_weight),
        reg_lambda(param.reg_lambda),
        reg_alpha(param.reg_alpha),
        max_delta_step(param.max_delta_step),
        learning_rate{param.learning_rate},
        max_cat_to_onehot{param.max_cat_to_onehot},
        max_cat_threshold{param.max_cat_threshold} {}
};

/**
 * @enum DefaultDirection node.cuh
 * @brief Default direction to be followed in case of missing values
 */
enum DefaultDirection {
  /** move to left child */
  kLeftDir = 0,
  /** move to right child */
  kRightDir
};

struct DeviceSplitCandidate {
  float loss_chg{-std::numeric_limits<float>::max()};
  DefaultDirection dir{kLeftDir};
  int findex {-1};
  float fvalue {0};
  // categorical split, either it's the split category for OHE or the threshold for partition-based
  // split.
  bst_cat_t thresh{-1};

  bool is_cat { false };

  GradientPairInt64 left_sum;
  GradientPairInt64 right_sum;

  XGBOOST_DEVICE DeviceSplitCandidate() {}  // NOLINT

  XGBOOST_DEVICE void Update(float loss_chg_in, DefaultDirection dir_in, float fvalue_in,
                             int findex_in, GradientPairInt64 left_sum_in,
                             GradientPairInt64 right_sum_in, bool cat,
                             const GPUTrainingParam& param, const GradientQuantiser& quantiser) {
    if (loss_chg_in > loss_chg &&
        quantiser.ToFloatingPoint(left_sum_in).GetHess() >= param.min_child_weight &&
        quantiser.ToFloatingPoint(right_sum_in).GetHess() >= param.min_child_weight) {
        loss_chg = loss_chg_in;
        dir = dir_in;
        fvalue = fvalue_in;
        is_cat = cat;
        left_sum = left_sum_in;
        right_sum = right_sum_in;
        findex = findex_in;
      }
  }

  /**
   * \brief Update for partition-based splits.
   */
  XGBOOST_DEVICE void UpdateCat(float loss_chg_in, DefaultDirection dir_in, bst_cat_t thresh_in,
                                bst_feature_t findex_in, GradientPairInt64 left_sum_in,
                                GradientPairInt64 right_sum_in, GPUTrainingParam const& param,
                                const GradientQuantiser& quantiser) {
      if (loss_chg_in > loss_chg &&
          quantiser.ToFloatingPoint(left_sum_in).GetHess() >= param.min_child_weight &&
          quantiser.ToFloatingPoint(right_sum_in).GetHess() >= param.min_child_weight) {
        loss_chg = loss_chg_in;
        dir = dir_in;
        fvalue = std::numeric_limits<float>::quiet_NaN();
        thresh = thresh_in;
        is_cat = true;
        left_sum = left_sum_in;
        right_sum = right_sum_in;
        findex = findex_in;
      }
  }

  [[nodiscard]] XGBOOST_DEVICE bool IsValid() const { return loss_chg > 0.0f; }

  friend std::ostream& operator<<(std::ostream& os, DeviceSplitCandidate const& c) {
    os << "loss_chg:" << c.loss_chg << ", "
       << "dir: " << c.dir << ", "
       << "findex: " << c.findex << ", "
       << "fvalue: " << c.fvalue << ", "
       << "thresh: " << c.thresh << ", "
       << "is_cat: " << c.is_cat << ", "
       << "left sum: " << c.left_sum << ", "
       << "right sum: " << c.right_sum << std::endl;
    return os;
  }
};

namespace cuda_impl {
inline BatchParam HistBatch(TrainParam const& param) {
  auto p = BatchParam{param.max_bin, TrainParam::DftSparseThreshold()};
  p.prefetch_copy = true;
  p.n_prefetch_batches = ::xgboost::cuda_impl::DftPrefetchBatches();
  return p;
}

inline BatchParam ApproxBatch(TrainParam const& p, common::Span<float const> hess,
                              ObjInfo const& task) {
  auto batch = BatchParam{p.max_bin, hess, !task.const_hess};
  batch.prefetch_copy = true;
  batch.n_prefetch_batches = ::xgboost::cuda_impl::DftPrefetchBatches();
  return batch;
}
}  // namespace cuda_impl

template <typename T>
struct SumCallbackOp {
  // Running prefix
  T running_total;
  // Constructor
  XGBOOST_DEVICE SumCallbackOp() : running_total(T()) {}
  XGBOOST_DEVICE T operator()(T block_aggregate) {
    T old_prefix = running_total;
    running_total += block_aggregate;
    return old_prefix;
  }
};
}  // namespace xgboost::tree