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

important change to regrank interface, need some more test

Browse Source
This commit is contained in:
tqchen 2014-05-03 14:20:27 -07:00
parent 2305ea7af7
commit aab1b0e7b3
8 changed files with 223 additions and 145 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
booster/xgboost_data.h
View File

@ -321,6 +321,8 @@ namespace xgboost{
fi.Read(&col_access, sizeof(int));
if (col_access != 0){
FMatrixS::LoadBinary(fi, col_ptr_, col_data_);
}else{
this->InitData();
}
}
/*!

9
python/Makefile
View File

@ -1,19 +1,18 @@
export CC = gcc
export CXX = g++
export CFLAGS = -Wall -O3 -msse2 -Wno-unknown-pragmas -fopenmp
export CFLAGS = -Wall -msse2 -Wno-unknown-pragmas -fopenmp
# specify tensor path
SLIB = xgboostpy.so
OBJ = xgboost_python.o
SLIB = libxgboostpy.so
.PHONY: clean all
all: $(SLIB)
export LDFLAGS= -pthread -lm
xgboostpy.so: xgboost_python.cpp ../regrank/*.h ../booster/*.h ../booster/*/*.hpp ../booster/*.hpp
libxgboostpy.so: xgboost_python.cpp ../regrank/*.h ../booster/*.h ../booster/*/*.hpp ../booster/*.hpp
$(SLIB) :
$(CXX) $(CFLAGS) $(LDFLAGS) -shared -o $@ $(filter %.cpp %.o %.c, $^)
$(CXX) $(CFLAGS) -fPIC $(LDFLAGS) -shared -o $@ $(filter %.cpp %.o %.c, $^)
$(BIN) :
$(CXX) $(CFLAGS) $(LDFLAGS) -o $@ $(filter %.cpp %.o %.c, $^)

15
python/xgboost.py
View File

@ -2,7 +2,7 @@
import ctypes
# load in xgboost library
#xglib = ctypes.cdll.LoadLibrary('./libxgboostpy.so')
xglib = ctypes.cdll.LoadLibrary('./libxgboostpy.so')
# entry type of sparse matrix
class REntry(ctypes.Structure):
@ -10,6 +10,13 @@ class REntry(ctypes.Structure):
class DMatrix:
def __init__(fname = None):
self.__handle = xglib.
def __init__(self,fname = None):
self.__handle = xglib.XGDMatrixCreate();
if fname != None:
xglib.XGDMatrixLoad(self.__handle, ctypes.c_char_p(fname), 0)
def __del__(self):
xglib.XGDMatrixFree(self.__handle)
dmata = DMatrix('xx.buffer')

47
python/xgboost_python.cpp
View File

@ -1,12 +1,41 @@
#include "xgboost_python.h"
#include "../regrank/xgboost_regrank.h"
#include "../regrank/xgboost_regrank_data.h"
void* XGDMatrixCreate(void){
return NULL;
}
void XGDMatrixFree(void *handle){
}
void XGDMatrixLoad(void *handle, const char *fname){
}
void XGDMatrixSaveBinary( void *handle, const char *fname ){
}
namespace xgboost{
namespace python{
class DMatrix: public regrank::DMatrix{
public:
// whether column is initialized
bool init_col_;
public:
DMatrix(void){
init_col_ = false;
}
~DMatrix(void){}
public:
inline void Load(const char *fname, bool silent){
this->CacheLoad(fname, silent);
init_col_ = this->data.HaveColAccess();
}
};
};
};
using namespace xgboost::python;
extern "C"{
void* XGDMatrixCreate(void){
return new DMatrix();
}
void XGDMatrixFree(void *handle){
delete static_cast<DMatrix*>(handle);
}
void XGDMatrixLoad(void *handle, const char *fname, int silent){
static_cast<DMatrix*>(handle)->Load(fname, silent!=0);
}
void XGDMatrixSaveBinary(void *handle, const char *fname, int silent){
static_cast<DMatrix*>(handle)->SaveBinary(fname, silent!=0);
}
};

76
python/xgboost_python.h
View File

@ -7,37 +7,51 @@
* use c style interface
*/
#include "../booster/xgboost_data.h"
/*! \brief type of row entry */
typedef xgboost::booster::FMatrixS::REntry XGEntry;
/*!
* \brief create a data matrix
* \return a new data matrix
*/
void* XGDMatrixCreate(void);
/*!
* \brief free space in data matrix
*/
void XGDMatrixFree(void *handle);
/*!
* \brief load a data matrix from text file or buffer(if exists)
* \param handle a instance of data matrix
* \param fname file name
*/
void XGDMatrixLoad(void *handle, const char *fname);
/*!
* \brief load a data matrix into binary file
* \param handle a instance of data matrix
* \param fname file name
*/
void XGDMatrixSaveBinary( void *handle, const char *fname );
/*!
* \brief add row
* \param handle a instance of data matrix
* \param fname file name
* \return a new data matrix
*/
//void XGDMatrixPush( void *handle, const std::pair<int,> );
extern "C"{
/*! \brief type of row entry */
typedef xgboost::booster::FMatrixS::REntry XGEntry;
/*!
* \brief create a data matrix
* \return a new data matrix
*/
void* XGDMatrixCreate(void);
/*!
* \brief free space in data matrix
*/
void XGDMatrixFree(void *handle);
/*!
* \brief load a data matrix from text file or buffer(if exists)
* \param handle a instance of data matrix
* \param fname file name
* \param silent print statistics when loading
*/
void XGDMatrixLoad(void *handle, const char *fname, int silent);
/*!
* \brief load a data matrix into binary file
* \param handle a instance of data matrix
* \param fname file name
* \param silent print statistics when saving
*/
void XGDMatrixSaveBinary(void *handle, const char *fname, int silent);
/*!
* \brief add row
* \param handle a instance of data matrix
* \param fname file name
* \return a new data matrix
*/
void XGDMatrixPush(void *handle, const XGEntry *data, int len);
/*!
* \brief create a booster
*/
void* XGBoostCreate(void);
/*!
* \brief create a booster
*/
void* XGBoost(void);
};
#endif

187
regrank/xgboost_regrank.h
View File

@ -28,40 +28,36 @@ namespace xgboost{
name_obj_ = "reg";
}
/*!
* \brief a regression booter associated with training and evaluating data
* \param train pointer to the training data
* \param evals array of evaluating data
* \param evname name of evaluation data, used print statistics
*/
RegRankBoostLearner(const DMatrix *train,
const std::vector<DMatrix *> &evals,
const std::vector<std::string> &evname){
* \brief a regression booter associated with training and evaluating data
* \param mats array of pointers to matrix whose prediction result need to be cached
*/
RegRankBoostLearner(const std::vector<const DMatrix *> mats){
silent = 0;
this->SetData(train, evals, evname);
}
obj_ = NULL;
name_obj_ = "reg";
this->SetCacheData(mats);
}
/*!
* \brief associate regression booster with training and evaluating data
* \param train pointer to the training data
* \param evals array of evaluating data
* \param evname name of evaluation data, used print statistics
*/
inline void SetData(const DMatrix *train,
const std::vector<DMatrix *> &evals,
const std::vector<std::string> &evname){
this->train_ = train;
this->evals_ = evals;
this->evname_ = evname;
* \brief add internal cache space for mat, this can speedup prediction for matrix,
* please cache prediction for training and eval data
* warning: if the model is loaded from file from some previous training history
* set cache data must be called with exactly SAME
* data matrices to continue training otherwise it will cause error
* \param mats array of pointers to matrix whose prediction result need to be cached
*/
inline void SetCacheData(const std::vector<const DMatrix *> mats){
// estimate feature bound
int num_feature = (int)(train->data.NumCol());
int num_feature = 0;
// assign buffer index
unsigned buffer_size = static_cast<unsigned>(train->Size());
unsigned buffer_size = 0;
for (size_t i = 0; i < evals.size(); ++i){
buffer_size += static_cast<unsigned>(evals[i]->Size());
num_feature = std::max(num_feature, (int)(evals[i]->data.NumCol()));
utils::Assert( cache_.size() == 0, "can only call cache data once" );
for( size_t i = 0; i < mats.size(); ++i ){
cache_.push_back( CacheEntry( mats[i], buffer_size ) );
buffer_size += static_cast<unsigned>(mats[i]->Size());
num_feature = std::max(num_feature, (int)(mats[i]->data.NumCol()));
}
char str_temp[25];
if (num_feature > mparam.num_feature){
mparam.num_feature = num_feature;
@ -74,15 +70,13 @@ namespace xgboost{
if (!silent){
printf("buffer_size=%u\n", buffer_size);
}
// set eval_preds tmp sapce
this->eval_preds_.resize(evals.size(), std::vector<float>());
}
/*!
* \brief set parameters from outside
* \param name name of the parameter
* \param val value of the parameter
*/
* \brief set parameters from outside
* \param name name of the parameter
* \param val value of the parameter
*/
inline void SetParam(const char *name, const char *val){
if (!strcmp(name, "silent")) silent = atoi(val);
if (!strcmp(name, "eval_metric")) evaluator_.AddEval(val);
@ -104,8 +98,8 @@ namespace xgboost{
evaluator_.AddEval( obj_->DefaultEvalMetric() );
}
/*!
* \brief initialize the current data storage for model, if the model is used first time, call this function
*/
* \brief initialize the current data storage for model, if the model is used first time, call this function
*/
inline void InitModel(void){
base_gbm.InitModel();
mparam.AdjustBase();
@ -147,74 +141,66 @@ namespace xgboost{
* \brief update the model for one iteration
* \param iteration iteration number
*/
inline void UpdateOneIter(int iter){
this->PredictBuffer(preds_, *train_, 0);
obj_->GetGradient(preds_, train_->info, base_gbm.NumBoosters(), grad_, hess_);
inline void UpdateOneIter(int iter, const DMatrix &train){
this->PredictRaw(preds_, train);
obj_->GetGradient(preds_, train.info, base_gbm.NumBoosters(), grad_, hess_);
// do boost
std::vector<unsigned> root_index;
base_gbm.DoBoost(grad_, hess_, train_->data, root_index);
base_gbm.DoBoost(grad_, hess_, train.data, root_index);
}
/*!
* \brief evaluate the model for specific iteration
* \param iter iteration number
* \param evals datas i want to evaluate
* \param evname name of each dataset
* \param fo file to output log
*/
inline void EvalOneIter(int iter, FILE *fo = stderr){
inline void EvalOneIter(int iter,
const std::vector<const DMatrix*> &evals,
const std::vector<std::string> &evname,
FILE *fo=stderr ){
fprintf(fo, "[%d]", iter);
int buffer_offset = static_cast<int>(train_->Size());
for (size_t i = 0; i < evals_.size(); ++i){
std::vector<float> &preds = this->eval_preds_[i];
this->PredictBuffer(preds, *evals_[i], buffer_offset);
obj_->PredTransform(preds);
evaluator_.Eval(fo, evname_[i].c_str(), preds, evals_[i]->info);
buffer_offset += static_cast<int>(evals_[i]->Size());
for (size_t i = 0; i < evals.size(); ++i){
this->PredictRaw(preds_, *evals[i]);
obj_->PredTransform(preds_);
evaluator_.Eval(fo, evname[i].c_str(), preds_, evals[i]->info);
}
fprintf(fo, "\n");
fflush(fo);
}
/*! \brief get prediction, without buffering */
inline void Predict(std::vector<float> &preds, const DMatrix &data){
preds.resize(data.Size());
const unsigned ndata = static_cast<unsigned>(data.Size());
#pragma omp parallel for schedule( static )
for (unsigned j = 0; j < ndata; ++j){
preds[j] = mparam.base_score + base_gbm.Predict(data.data, j, -1);
}
this->PredictRaw(preds,data);
obj_->PredTransform( preds );
}
public:
/*!
* \brief interactive update
* \param action action type
* \parma train training data
*/
inline void UpdateInteract(std::string action){
this->InteractPredict(preds_, *train_, 0);
int buffer_offset = static_cast<int>(train_->Size());
for (size_t i = 0; i < evals_.size(); ++i){
std::vector<float> &preds = this->eval_preds_[i];
this->InteractPredict(preds, *evals_[i], buffer_offset);
buffer_offset += static_cast<int>(evals_[i]->Size());
inline void UpdateInteract(std::string action, const DMatrix& train){
for(size_t i = 0; i < cache_.size(); ++i){
this->InteractPredict(preds_, *cache_[i].mat_);
}
if (action == "remove"){
base_gbm.DelteBooster(); return;
}
obj_->GetGradient(preds_, train_->info, base_gbm.NumBoosters(), grad_, hess_);
obj_->GetGradient(preds_, train.info, base_gbm.NumBoosters(), grad_, hess_);
std::vector<unsigned> root_index;
base_gbm.DoBoost(grad_, hess_, train_->data, root_index);
base_gbm.DoBoost(grad_, hess_, train.data, root_index);
this->InteractRePredict(*train_, 0);
buffer_offset = static_cast<int>(train_->Size());
for (size_t i = 0; i < evals_.size(); ++i){
this->InteractRePredict(*evals_[i], buffer_offset);
buffer_offset += static_cast<int>(evals_[i]->Size());
for(size_t i = 0; i < cache_.size(); ++i){
this->InteractRePredict(*cache_[i].mat_);
}
}
private:
/*! \brief get the transformed predictions, given data */
inline void InteractPredict(std::vector<float> &preds, const DMatrix &data, unsigned buffer_offset){
inline void InteractPredict(std::vector<float> &preds, const DMatrix &data){
int buffer_offset = this->FindBufferOffset(data);
utils::Assert( buffer_offset >=0, "interact mode must cache training data" );
preds.resize(data.Size());
const unsigned ndata = static_cast<unsigned>(data.Size());
#pragma omp parallel for schedule( static )
@ -224,7 +210,9 @@ namespace xgboost{
obj_->PredTransform( preds );
}
/*! \brief repredict trial */
inline void InteractRePredict(const DMatrix &data, unsigned buffer_offset){
inline void InteractRePredict(const DMatrix &data){
int buffer_offset = this->FindBufferOffset(data);
utils::Assert( buffer_offset >=0, "interact mode must cache training data" );
const unsigned ndata = static_cast<unsigned>(data.Size());
#pragma omp parallel for schedule( static )
for (unsigned j = 0; j < ndata; ++j){
@ -232,13 +220,24 @@ namespace xgboost{
}
}
private:
/*! \brief get the transformed predictions, given data */
inline void PredictBuffer(std::vector<float> &preds, const DMatrix &data, unsigned buffer_offset){
/*! \brief get un-transformed prediction*/
inline void PredictRaw(std::vector<float> &preds, const DMatrix &data){
this->PredictBuffer(preds, data, this->FindBufferOffset(data) );
}
/*! \brief get the un-transformed predictions, given data */
inline void PredictBuffer(std::vector<float> &preds, const DMatrix &data, int buffer_offset){
preds.resize(data.Size());
const unsigned ndata = static_cast<unsigned>(data.Size());
#pragma omp parallel for schedule( static )
for (unsigned j = 0; j < ndata; ++j){
preds[j] = mparam.base_score + base_gbm.Predict(data.data, j, buffer_offset + j);
if( buffer_offset >= 0 ){
#pragma omp parallel for schedule( static )
for (unsigned j = 0; j < ndata; ++j){
preds[j] = mparam.base_score + base_gbm.Predict(data.data, j, buffer_offset + j);
}
}else
#pragma omp parallel for schedule( static )
for (unsigned j = 0; j < ndata; ++j){
preds[j] = mparam.base_score + base_gbm.Predict(data.data, j, -1);
}{
}
}
private:
@ -260,10 +259,10 @@ namespace xgboost{
memset(reserved, 0, sizeof(reserved));
}
/*!
* \brief set parameters from outside
* \param name name of the parameter
* \param val value of the parameter
*/
* \brief set parameters from outside
* \param name name of the parameter
* \param val value of the parameter
*/
inline void SetParam(const char *name, const char *val){
if (!strcmp("base_score", name)) base_score = (float)atof(val);
if (!strcmp("loss_type", name)) loss_type = atoi(val);
@ -279,15 +278,28 @@ namespace xgboost{
}
}
};
private:
struct CacheEntry{
const DMatrix *mat_;
int buffer_offset_;
CacheEntry(const DMatrix *mat, int buffer_offset)
:mat_(mat), buffer_offset_(buffer_offset){}
};
/*! \brief the entries indicates that we have internal prediction cache */
std::vector<CacheEntry> cache_;
private:
// find internal bufer offset for certain matrix, if not exist, return -1
inline int FindBufferOffset(const DMatrix &mat){
for(size_t i = 0; i < cache_.size(); ++i){
if( cache_[i].mat_ == &mat ) return cache_[i].buffer_offset_;
}
return -1;
}
private:
int silent;
EvalSet evaluator_;
booster::GBMBase base_gbm;
ModelParam mparam;
const DMatrix *train_;
std::vector<DMatrix *> evals_;
std::vector<std::string> evname_;
std::vector<unsigned> buffer_index_;
ModelParam mparam;
// objective fnction
IObjFunction *obj_;
// name of objective function
@ -295,7 +307,6 @@ namespace xgboost{
std::vector< std::pair<std::string, std::string> > cfg_;
private:
std::vector<float> grad_, hess_, preds_;
std::vector< std::vector<float> > eval_preds_;
};
}
};

2
regrank/xgboost_regrank_data.h
View File

@ -116,8 +116,6 @@ namespace xgboost{
}
}
fs.Close();
// initialize column support as well
data.InitData();
if (!silent){
printf("%ux%u matrix with %lu entries is loaded from %s\n",

30
regrank/xgboost_regrank_main.cpp
View File

@ -62,6 +62,7 @@ namespace xgboost{
if (!strcmp("seed", name)) random::Seed(atoi(val));
if (!strcmp("num_round", name)) num_round = atoi(val);
if (!strcmp("save_period", name)) save_period = atoi(val);
if (!strcmp("eval_train", name)) eval_train = atoi(val);
if (!strcmp("task", name)) task = val;
if (!strcmp("data", name)) train_path = val;
if (!strcmp("test:data", name)) test_path = val;
@ -92,6 +93,7 @@ namespace xgboost{
use_buffer = 1;
num_round = 10;
save_period = 0;
eval_train = 0;
dump_model_stats = 0;
task = "train";
model_in = "NULL";
@ -122,9 +124,22 @@ namespace xgboost{
for (size_t i = 0; i < eval_data_names.size(); ++i){
deval.push_back(new DMatrix());
deval.back()->CacheLoad(eval_data_paths[i].c_str(), silent != 0, use_buffer != 0);
devalall.push_back(deval.back());
}
std::vector<const DMatrix *> dcache(1, &data);
for( size_t i = 0; i < deval.size(); ++ i){
dcache.push_back( deval[i] );
}
// set cache data to be all training and evaluation data
learner.SetCacheData(dcache);
// add training set to evaluation set if needed
if( eval_train != 0 ){
devalall.push_back( &data );
eval_data_names.push_back( std::string("train") );
}
}
learner.SetData(&data, deval, eval_data_names);
}
inline void InitLearner(void){
cfg.BeforeFirst();
@ -148,8 +163,8 @@ namespace xgboost{
for (int i = 0; i < num_round; ++i){
elapsed = (unsigned long)(time(NULL) - start);
if (!silent) printf("boosting round %d, %lu sec elapsed\n", i, elapsed);
learner.UpdateOneIter(i);
learner.EvalOneIter(i);
learner.UpdateOneIter(i, data);
learner.EvalOneIter(i, devalall, eval_data_names);
if (save_period != 0 && (i + 1) % save_period == 0){
this->SaveModel(i);
}
@ -169,7 +184,7 @@ namespace xgboost{
}
}
inline void TaskEval(void){
learner.EvalOneIter(0);
learner.EvalOneIter(0, devalall, eval_data_names);
}
inline void TaskInteractive(void){
const time_t start = time(NULL);
@ -179,7 +194,7 @@ namespace xgboost{
cfg_batch.BeforeFirst();
while (cfg_batch.Next()){
if (!strcmp(cfg_batch.name(), "run")){
learner.UpdateInteract(interact_action);
learner.UpdateInteract(interact_action, data);
batch_action += 1;
}
else{
@ -188,7 +203,7 @@ namespace xgboost{
}
if (batch_action == 0){
learner.UpdateInteract(interact_action);
learner.UpdateInteract(interact_action, data);
}
utils::Assert(model_out != "NULL", "interactive mode must specify model_out");
this->SaveModel(model_out.c_str());
@ -235,6 +250,8 @@ namespace xgboost{
int silent;
/* \brief whether use auto binary buffer */
int use_buffer;
/* \brief whether evaluate training statistics */
int eval_train;
/* \brief number of boosting iterations */
int num_round;
/* \brief the period to save the model, 0 means only save the final round model */
@ -272,6 +289,7 @@ namespace xgboost{
private:
DMatrix data;
std::vector<DMatrix*> deval;
std::vector<const DMatrix*> devalall;
utils::FeatMap fmap;
RegRankBoostLearner learner;
};