Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Extract Sketch Entry from hist maker. (#7503)

Browse Source 
* Extract Sketch Entry from hist maker.

* Add a new sketch container for sorted inputs.
* Optimize bin search.
This commit is contained in:
Jiaming Yuan
2021-12-18 05:36:56 +08:00
committed by GitHub
parent b4a1236cfc
commit 9ab73f737e
15 changed files with 393 additions and 217 deletions
Download Patch File Download Diff File Expand all files Collapse all files

87
src/tree/updater_basemaker-inl.h
View File

@@ -369,92 +369,7 @@ class BaseMaker: public TreeUpdater {
}
}
}
/*! \brief common helper data structure to build sketch */
struct SketchEntry {
/*! \brief total sum of amount to be met */
double sum_total;
/*! \brief statistics used in the sketch */
double rmin, wmin;
/*! \brief last seen feature value */
bst_float last_fvalue;
/*! \brief current size of sketch */
double next_goal;
// pointer to the sketch to put things in
common::WXQuantileSketch<bst_float, bst_float> *sketch;
// initialize the space
inline void Init(unsigned max_size) {
next_goal = -1.0f;
rmin = wmin = 0.0f;
sketch->temp.Reserve(max_size + 1);
sketch->temp.size = 0;
}
/*!
* \brief push a new element to sketch
* \param fvalue feature value, comes in sorted ascending order
* \param w weight
* \param max_size
*/
inline void Push(bst_float fvalue, bst_float w, unsigned max_size) {
if (next_goal == -1.0f) {
next_goal = 0.0f;
last_fvalue = fvalue;
wmin = w;
return;
}
if (last_fvalue != fvalue) {
double rmax = rmin + wmin;
if (rmax >= next_goal && sketch->temp.size != max_size) {
if (sketch->temp.size == 0 ||
last_fvalue > sketch->temp.data[sketch->temp.size-1].value) {
// push to sketch
sketch->temp.data[sketch->temp.size] =
common::WXQuantileSketch<bst_float, bst_float>::
Entry(static_cast<bst_float>(rmin),
static_cast<bst_float>(rmax),
static_cast<bst_float>(wmin), last_fvalue);
CHECK_LT(sketch->temp.size, max_size)
<< "invalid maximum size max_size=" << max_size
<< ", stemp.size" << sketch->temp.size;
++sketch->temp.size;
}
if (sketch->temp.size == max_size) {
next_goal = sum_total * 2.0f + 1e-5f;
} else {
next_goal = static_cast<bst_float>(sketch->temp.size * sum_total / max_size);
}
} else {
if (rmax >= next_goal) {
LOG(TRACKER) << "INFO: rmax=" << rmax
<< ", sum_total=" << sum_total
<< ", naxt_goal=" << next_goal
<< ", size=" << sketch->temp.size;
}
}
rmin = rmax;
wmin = w;
last_fvalue = fvalue;
} else {
wmin += w;
}
}
/*! \brief push final unfinished value to the sketch */
inline void Finalize(unsigned max_size) {
double rmax = rmin + wmin;
if (sketch->temp.size == 0 || last_fvalue > sketch->temp.data[sketch->temp.size-1].value) {
CHECK_LE(sketch->temp.size, max_size)
<< "Finalize: invalid maximum size, max_size=" << max_size
<< ", stemp.size=" << sketch->temp.size;
// push to sketch
sketch->temp.data[sketch->temp.size] =
common::WXQuantileSketch<bst_float, bst_float>::
Entry(static_cast<bst_float>(rmin),
static_cast<bst_float>(rmax),
static_cast<bst_float>(wmin), last_fvalue);
++sketch->temp.size;
}
sketch->PushTemp();
}
};
using SketchEntry = common::SortedQuantile;
/*! \brief training parameter of tree grower */
TrainParam param_;
/*! \brief queue of nodes to be expanded */