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

Replaced std::vector-based interfaces with HostDeviceVector-based interfaces. (#3116)

Browse Source 
* Replaced std::vector-based interfaces with HostDeviceVector-based interfaces.

- replacement was performed in the learner, boosters, predictors,
  updaters, and objective functions
- only interfaces used in training were replaced;
  interfaces like PredictInstance() still use std::vector
- refactoring necessary for replacement of interfaces was also performed,
  such as using HostDeviceVector in prediction cache

* HostDeviceVector-based interfaces for custom objective function example plugin.
This commit is contained in:
Andrew V. Adinetz
2018-02-28 01:00:04 +01:00
committed by Rory Mitchell
parent 11bfa8584d
commit d5992dd881
38 changed files with 371 additions and 519 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
include/xgboost/gbm.h
View File

@@ -68,12 +68,9 @@ class GradientBooster {
* \param obj The objective function, optional, can be nullptr when use customized version
* the booster may change content of gpair
*/
virtual void DoBoost(DMatrix* p_fmat,
std::vector<bst_gpair>* in_gpair,
ObjFunction* obj = nullptr) = 0;
virtual void DoBoost(DMatrix* p_fmat,
HostDeviceVector<bst_gpair>* in_gpair,
ObjFunction* obj = nullptr);
ObjFunction* obj = nullptr) = 0;
/*!
* \brief generate predictions for given feature matrix
@@ -82,12 +79,9 @@ class GradientBooster {
* \param ntree_limit limit the number of trees used in prediction, when it equals 0, this means
* we do not limit number of trees, this parameter is only valid for gbtree, but not for gblinear
*/
virtual void PredictBatch(DMatrix* dmat,
std::vector<bst_float>* out_preds,
unsigned ntree_limit = 0) = 0;
virtual void PredictBatch(DMatrix* dmat,
HostDeviceVector<bst_float>* out_preds,
unsigned ntree_limit = 0);
unsigned ntree_limit = 0) = 0;
/*!
* \brief online prediction function, predict score for one instance at a time
* NOTE: use the batch prediction interface if possible, batch prediction is usually

10
include/xgboost/learner.h
View File

@@ -84,7 +84,7 @@ class Learner : public rabit::Serializable {
*/
virtual void BoostOneIter(int iter,
DMatrix* train,
std::vector<bst_gpair>* in_gpair) = 0;
HostDeviceVector<bst_gpair>* in_gpair) = 0;
/*!
* \brief evaluate the model for specific iteration using the configured metrics.
* \param iter iteration number
@@ -109,7 +109,7 @@ class Learner : public rabit::Serializable {
*/
virtual void Predict(DMatrix* data,
bool output_margin,
std::vector<bst_float> *out_preds,
HostDeviceVector<bst_float> *out_preds,
unsigned ntree_limit = 0,
bool pred_leaf = false,
bool pred_contribs = false,
@@ -169,7 +169,7 @@ class Learner : public rabit::Serializable {
*/
inline void Predict(const SparseBatch::Inst &inst,
bool output_margin,
std::vector<bst_float> *out_preds,
HostDeviceVector<bst_float> *out_preds,
unsigned ntree_limit = 0) const;
/*!
* \brief Create a new instance of learner.
@@ -192,9 +192,9 @@ class Learner : public rabit::Serializable {
// implementation of inline functions.
inline void Learner::Predict(const SparseBatch::Inst& inst,
bool output_margin,
std::vector<bst_float>* out_preds,
HostDeviceVector<bst_float>* out_preds,
unsigned ntree_limit) const {
gbm_->PredictInstance(inst, out_preds, ntree_limit);
gbm_->PredictInstance(inst, &out_preds->data_h(), ntree_limit);
if (!output_margin) {
obj_->PredTransform(out_preds);
}

12
include/xgboost/objective.h
View File

@@ -44,14 +44,10 @@ class ObjFunction {
* \param iteration current iteration number.
* \param out_gpair output of get gradient, saves gradient and second order gradient in
*/
virtual void GetGradient(const std::vector<bst_float>& preds,
const MetaInfo& info,
int iteration,
std::vector<bst_gpair>* out_gpair) = 0;
virtual void GetGradient(HostDeviceVector<bst_float>* preds,
const MetaInfo& info,
int iteration,
HostDeviceVector<bst_gpair>* out_gpair);
HostDeviceVector<bst_gpair>* out_gpair) = 0;
/*! \return the default evaluation metric for the objective */
virtual const char* DefaultEvalMetric() const = 0;
@@ -60,17 +56,13 @@ class ObjFunction {
* \brief transform prediction values, this is only called when Prediction is called
* \param io_preds prediction values, saves to this vector as well
*/
virtual void PredTransform(std::vector<bst_float> *io_preds) {}
virtual void PredTransform(HostDeviceVector<bst_float> *io_preds);
virtual void PredTransform(HostDeviceVector<bst_float> *io_preds) {}
/*!
* \brief transform prediction values, this is only called when Eval is called,
* usually it redirect to PredTransform
* \param io_preds prediction values, saves to this vector as well
*/
virtual void EvalTransform(std::vector<bst_float> *io_preds) {
this->PredTransform(io_preds);
}
virtual void EvalTransform(HostDeviceVector<bst_float> *io_preds) {
this->PredTransform(io_preds);
}

45
include/xgboost/predictor.h
View File

@@ -63,22 +63,6 @@ class Predictor {
* limit trees.
*/
virtual void PredictBatch(DMatrix* dmat, std::vector<bst_float>* out_preds,
const gbm::GBTreeModel& model, int tree_begin,
unsigned ntree_limit = 0) = 0;
/**
* \brief Generate batch predictions for a given feature matrix. May use
* cached predictions if available instead of calculating from scratch.
*
* \param [in,out] dmat Feature matrix.
* \param [in,out] out_preds The output preds.
* \param model The model to predict from.
* \param tree_begin The tree begin index.
* \param ntree_limit (Optional) The ntree limit. 0 means do not
* limit trees.
*/
virtual void PredictBatch(DMatrix* dmat, HostDeviceVector<bst_float>* out_preds,
const gbm::GBTreeModel& model, int tree_begin,
unsigned ntree_limit = 0) = 0;
@@ -186,41 +170,14 @@ class Predictor {
static Predictor* Create(std::string name);
protected:
/**
* \fn bool PredictFromCache(DMatrix* dmat, std::vector<bst_float>*
* out_preds, const gbm::GBTreeModel& model, unsigned ntree_limit = 0)
*
* \brief Attempt to predict from cache.
*
* \return True if it succeeds, false if it fails.
*/
bool PredictFromCache(DMatrix* dmat, std::vector<bst_float>* out_preds,
const gbm::GBTreeModel& model,
unsigned ntree_limit = 0);
/**
* \fn void Predictor::InitOutPredictions(const MetaInfo& info,
* std::vector<bst_float>* out_preds, const gbm::GBTreeModel& model) const;
*
* \brief Init out predictions according to base margin.
*
* \param info Dmatrix info possibly containing base margin.
* \param [in,out] out_preds The out preds.
* \param model The model.
*/
void InitOutPredictions(const MetaInfo& info,
std::vector<bst_float>* out_preds,
const gbm::GBTreeModel& model) const;
/**
* \struct PredictionCacheEntry
*
* \brief Contains pointer to input matrix and associated cached predictions.
*/
struct PredictionCacheEntry {
std::shared_ptr<DMatrix> data;
std::vector<bst_float> predictions;
HostDeviceVector<bst_float> predictions;
};
/**

9
include/xgboost/tree_updater.h
View File

@@ -40,12 +40,9 @@ class TreeUpdater {
* but maybe different random seeds, usually one tree is passed in at a time,
* there can be multiple trees when we train random forest style model
*/
virtual void Update(const std::vector<bst_gpair>& gpair,
DMatrix* data,
const std::vector<RegTree*>& trees) = 0;
virtual void Update(HostDeviceVector<bst_gpair>* gpair,
DMatrix* data,
const std::vector<RegTree*>& trees);
const std::vector<RegTree*>& trees) = 0;
/*!
* \brief determines whether updater has enough knowledge about a given dataset
@@ -58,11 +55,9 @@ class TreeUpdater {
* updated by the time this function returns.
*/
virtual bool UpdatePredictionCache(const DMatrix* data,
std::vector<bst_float>* out_preds) {
HostDeviceVector<bst_float>* out_preds) {
return false;
}
virtual bool UpdatePredictionCache(const DMatrix* data,
HostDeviceVector<bst_float>* out_preds);
/*!
* \brief Create a tree updater given name