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

simplify

Browse Source
This commit is contained in:
tqchen 2014-11-18 10:52:18 -08:00
parent 5de0a2cdc0
commit ce7ecadf5e
1 changed files with 22 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

50
src/tree/updater_histmaker-inl.hpp
View File

@ -377,8 +377,7 @@ class CQHistMaker: public HistMaker<TStats> {
for (size_t i = 0; i < sketchs.size(); ++i) { for (size_t i = 0; i < sketchs.size(); ++i) {
sketchs[i].Init(info.num_row, this->param.sketch_eps); sketchs[i].Init(info.num_row, this->param.sketch_eps);
} }
std::vector< std::vector<SketchEntry> > stemp; thread_temp.resize(this->get_nthread());
stemp.resize(this->get_nthread());
// start accumulating statistics // start accumulating statistics
utils::IIterator<ColBatch> *iter = p_fmat->ColIterator(); utils::IIterator<ColBatch> *iter = p_fmat->ColIterator();
@ -390,7 +389,7 @@ class CQHistMaker: public HistMaker<TStats> {
#pragma omp parallel for schedule(dynamic, 1) #pragma omp parallel for schedule(dynamic, 1)
for (bst_omp_uint i = 0; i < nsize; ++i) { for (bst_omp_uint i = 0; i < nsize; ++i) {
this->MakeSketch(gpair, batch[i], tree, batch.col_index[i], this->MakeSketch(gpair, batch[i], tree, batch.col_index[i],
&stemp[omp_get_thread_num()]); &thread_temp[omp_get_thread_num()]);
} }
} }
// setup maximum size // setup maximum size
@ -460,7 +459,6 @@ class CQHistMaker: public HistMaker<TStats> {
* \param max_size * \param max_size
*/ */
inline void Push(bst_float fvalue, bst_float w, unsigned max_size) { inline void Push(bst_float fvalue, bst_float w, unsigned max_size) {
if (w == 0.0f) return;
if (wmin == 0.0f) { if (wmin == 0.0f) {
last_fvalue = fvalue; last_fvalue = fvalue;
wmin = w; wmin = w;
@ -520,56 +518,52 @@ class CQHistMaker: public HistMaker<TStats> {
if (c.length == 0) return; if (c.length == 0) return;
// initialize sbuilder for use // initialize sbuilder for use
std::vector<SketchEntry> &sbuilder = *p_temp; std::vector<SketchEntry> &sbuilder = *p_temp;
sbuilder.resize(this->qexpand.size()); sbuilder.resize(tree.param.num_nodes);
for (size_t i = 0; i < sbuilder.size(); ++i) { for (size_t i = 0; i < this->qexpand.size(); ++i) {
sbuilder[i].sum_total = 0.0f; const unsigned nid = this->qexpand[i];
sbuilder[i].sketch = &sketchs[i * tree.param.num_feature + fid]; const unsigned wid = this->node2workindex[nid];
sbuilder[nid].sum_total = 0.0f;
sbuilder[nid].sketch = &sketchs[wid * tree.param.num_feature + fid];
} }
// second pass, build the sketch
for (bst_uint j = 0; j < c.length; ++j) {
const bst_uint ridx = c[j].index;
const int nid = this->position[ridx];
if (nid >= 0) {
const int wid = this->node2workindex[nid];
sbuilder[wid].sketch->Push(c[j].fvalue, gpair[ridx].hess);
}
}
return;
// first pass, get sum of weight, TODO, optimization to skip first pass // first pass, get sum of weight, TODO, optimization to skip first pass
for (bst_uint j = 0; j < c.length; ++j) { for (bst_uint j = 0; j < c.length; ++j) {
const bst_uint ridx = c[j].index; const bst_uint ridx = c[j].index;
const int nid = this->position[ridx]; const int nid = this->position[ridx];
if (nid >= 0) { if (nid >= 0) {
const int wid = this->node2workindex[nid]; sbuilder[nid].sum_total += gpair[ridx].hess;
sbuilder[wid].sum_total += gpair[ridx].hess;
} }
} }
// if only one value, no need to do second pass // if only one value, no need to do second pass
if (c[0].fvalue == c[c.length-1].fvalue) { if (c[0].fvalue == c[c.length-1].fvalue) {
for (size_t wid = 0; wid < this->qexpand.size(); ++wid) { for (size_t i = 0; i < this->qexpand.size(); ++i) {
sbuilder[wid].sketch->Push(c[0].fvalue, sbuilder[wid].sum_total); const int nid = this->qexpand[i];
sbuilder[nid].sketch->Push(c[0].fvalue, sbuilder[nid].sum_total);
} }
return; return;
} }
// two pass scan // two pass scan
unsigned max_size = static_cast<unsigned>(this->param.sketch_ratio / this->param.sketch_eps); unsigned max_size = static_cast<unsigned>(this->param.sketch_ratio / this->param.sketch_eps);
for (size_t wid = 0; wid < sbuilder.size(); ++wid) { for (size_t i = 0; i < this->qexpand.size(); ++i) {
sbuilder[wid].Init(max_size); const int nid = this->qexpand[i];
sbuilder[nid].Init(max_size);
} }
// second pass, build the sketch // second pass, build the sketch
for (bst_uint j = 0; j < c.length; ++j) { for (bst_uint j = 0; j < c.length; ++j) {
const bst_uint ridx = c[j].index; const bst_uint ridx = c[j].index;
const int nid = this->position[ridx]; const int nid = this->position[ridx];
if (nid >= 0) { if (nid >= 0) {
const int wid = this->node2workindex[nid]; sbuilder[nid].Push(c[j].fvalue, gpair[ridx].hess, max_size);
sbuilder[wid].Push(c[j].fvalue, gpair[ridx].hess, max_size);
} }
} }
for (size_t wid = 0; wid < sbuilder.size(); ++wid) { for (size_t i = 0; i < this->qexpand.size(); ++i) {
sbuilder[wid].Finalize(max_size); const int nid = this->qexpand[i];
sbuilder[nid].Finalize(max_size);
} }
} }
typedef utils::WXQuantileSketch<bst_float, bst_float> WXQSketch; typedef utils::WXQuantileSketch<bst_float, bst_float> WXQSketch;
// thread temp data
std::vector< std::vector<SketchEntry> > thread_temp;
// summary array // summary array
WXQSketch::SummaryArray summary_array; WXQSketch::SummaryArray summary_array;
// reducer for summary // reducer for summary