Use bst_float consistently throughout (#1824)

* Fix various typos * Add override to functions that are overridden gcc gives warnings about functions that are being overridden by not being marked as oveirridden. This fixes it. * Use bst_float consistently Use bst_float for all the variables that involve weight, leaf value, gradient, hessian, gain, loss_chg, predictions, base_margin, feature values. In some cases, when due to additions and so on the value can take a larger value, double is used. This ensures that type conversions are minimal and reduces loss of precision.
2016-11-30 23:32:10 +05:30
parent da2556f58a
commit 6f16f0ef58
50 changed files with 392 additions and 389 deletions
--- a/include/xgboost/base.h
+++ b/include/xgboost/base.h
@@ -42,7 +42,7 @@
 /*! \brief namespace of xgboo st*/
 namespace xgboost {
 /*!
- * \brief unsigned interger type used in boost,
+ * \brief unsigned integer type used in boost,
 *  used for feature index and row index.
 */
 typedef uint32_t bst_uint;
@@ -62,7 +62,7 @@ struct bst_gpair {
 };

 /*! \brief small eps gap for minimum split decision. */
-const float rt_eps = 1e-6f;
+const bst_float rt_eps = 1e-6f;

 /*! \brief define unsigned long for openmp loop */
 typedef dmlc::omp_ulong omp_ulong;
--- a/include/xgboost/c_api.h
+++ b/include/xgboost/c_api.h
@@ -23,9 +23,10 @@ XGB_EXTERN_C {
 #define XGB_DLL XGB_EXTERN_C
 #endif

-// manually define unsign long
+// manually define unsigned long
 typedef uint64_t bst_ulong;  // NOLINT(*)

+
 /*! \brief handle to DMatrix */
 typedef void *DMatrixHandle;
 /*! \brief handle to Booster */
@@ -86,11 +87,11 @@ XGB_EXTERN_C typedef int XGBCallbackDataIterNext(
 * \brief get string message of the last error
 *
 *  all function in this file will return 0 when success
- *  and -1 when an error occured,
+ *  and -1 when an error occurred,
 *  XGBGetLastError can be called to retrieve the error
 *
- *  this function is threadsafe and can be called by different thread
- * \return const char* error inforomation
+ *  this function is thread safe and can be called by different thread
+ * \return const char* error information
 */
 XGB_DLL const char *XGBGetLastError();

@@ -124,7 +125,7 @@ XGB_DLL int XGDMatrixCreateFromDataIter(
 * \param indptr pointer to row headers
 * \param indices findex
 * \param data fvalue
- * \param nindptr number of rows in the matix + 1
+ * \param nindptr number of rows in the matrix + 1
 * \param nelem number of nonzero elements in the matrix
 * \param num_col number of columns; when it's set to 0, then guess from data
 * \param out created dmatrix
@@ -143,7 +144,7 @@ XGB_DLL int XGDMatrixCreateFromCSREx(const size_t* indptr,
 * \param indptr pointer to row headers
 * \param indices findex
 * \param data fvalue
- * \param nindptr number of rows in the matix + 1
+ * \param nindptr number of rows in the matrix + 1
 * \param nelem number of nonzero elements in the matrix
 * \param out created dmatrix
 * \return 0 when success, -1 when failure happens
@@ -159,7 +160,7 @@ XGB_DLL int XGDMatrixCreateFromCSR(const bst_ulong *indptr,
 * \param col_ptr pointer to col headers
 * \param indices findex
 * \param data fvalue
- * \param nindptr number of rows in the matix + 1
+ * \param nindptr number of rows in the matrix + 1
 * \param nelem number of nonzero elements in the matrix
 * \param num_row number of rows; when it's set to 0, then guess from data
 * \param out created dmatrix
@@ -178,7 +179,7 @@ XGB_DLL int XGDMatrixCreateFromCSCEx(const size_t* col_ptr,
 * \param col_ptr pointer to col headers
 * \param indices findex
 * \param data fvalue
- * \param nindptr number of rows in the matix + 1
+ * \param nindptr number of rows in the matrix + 1
 * \param nelem number of nonzero elements in the matrix
 * \param out created dmatrix
 * \return 0 when success, -1 when failure happens
@@ -201,7 +202,7 @@ XGB_DLL int XGDMatrixCreateFromCSC(const bst_ulong *col_ptr,
 XGB_DLL int XGDMatrixCreateFromMat(const float *data,
                                   bst_ulong nrow,
                                   bst_ulong ncol,
-                                   float  missing,
+                                   float missing,
                                   DMatrixHandle *out);
 /*!
 * \brief create a new dmatrix from sliced content of existing matrix
--- a/include/xgboost/data.h
+++ b/include/xgboost/data.h
@@ -65,7 +65,7 @@ struct MetaInfo {
   * \param i Instance index.
   * \return The weight.
   */
-  inline float GetWeight(size_t i) const {
+  inline bst_float GetWeight(size_t i) const {
    return weights.size() != 0 ?  weights[i] : 1.0f;
  }
  /*!
@@ -253,7 +253,7 @@ class DMatrix {
   * \brief check if column access is supported, if not, initialize column access.
   * \param enabled whether certain feature should be included in column access.
   * \param subsample subsample ratio when generating column access.
-   * \param max_row_perbatch auxilary information, maximum row used in each column batch.
+   * \param max_row_perbatch auxiliary information, maximum row used in each column batch.
   *         this is a hint information that can be ignored by the implementation.
   * \return Number of column blocks in the column access.
   */
@@ -304,7 +304,7 @@ class DMatrix {
  static DMatrix* Create(std::unique_ptr<DataSource>&& source,
                         const std::string& cache_prefix = "");
  /*!
-   * \brief Create a DMatrix by loaidng data from parser.
+   * \brief Create a DMatrix by loading data from parser.
   *  Parser can later be deleted after the DMatrix i created.
   * \param parser The input data parser
   * \param cache_prefix The path to prefix of temporary cache file of the DMatrix when used in external memory mode.
--- a/include/xgboost/gbm.h
+++ b/include/xgboost/gbm.h
@@ -78,7 +78,7 @@ class GradientBooster {
   *    we do not limit number of trees, this parameter is only valid for gbtree, but not for gblinear
   */
  virtual void Predict(DMatrix* dmat,
-                       std::vector<float>* out_preds,
+                       std::vector<bst_float>* out_preds,
                       unsigned ntree_limit = 0) = 0;
  /*!
   * \brief online prediction function, predict score for one instance at a time
@@ -93,7 +93,7 @@ class GradientBooster {
   * \sa Predict
   */
  virtual void Predict(const SparseBatch::Inst& inst,
-                       std::vector<float>* out_preds,
+                       std::vector<bst_float>* out_preds,
                       unsigned ntree_limit = 0,
                       unsigned root_index = 0) = 0;
  /*!
@@ -105,7 +105,7 @@ class GradientBooster {
   *    we do not limit number of trees, this parameter is only valid for gbtree, but not for gblinear
   */
  virtual void PredictLeaf(DMatrix* dmat,
-                           std::vector<float>* out_preds,
+                           std::vector<bst_float>* out_preds,
                           unsigned ntree_limit = 0) = 0;
  /*!
   * \brief dump the model in the requested format
@@ -127,7 +127,7 @@ class GradientBooster {
  static GradientBooster* Create(
      const std::string& name,
      const std::vector<std::shared_ptr<DMatrix> >& cache_mats,
-      float base_margin);
+      bst_float base_margin);
 };

 // implementing configure.
@@ -144,7 +144,7 @@ struct GradientBoosterReg
    : public dmlc::FunctionRegEntryBase<
  GradientBoosterReg,
  std::function<GradientBooster* (const std::vector<std::shared_ptr<DMatrix> > &cached_mats,
-                                  float base_margin)> > {
+                                  bst_float base_margin)> > {
 };

 /*!
--- a/include/xgboost/learner.h
+++ b/include/xgboost/learner.h
@@ -106,7 +106,7 @@ class Learner : public rabit::Serializable {
   */
  virtual void Predict(DMatrix* data,
                       bool output_margin,
-                       std::vector<float> *out_preds,
+                       std::vector<bst_float> *out_preds,
                       unsigned ntree_limit = 0,
                       bool pred_leaf = false) const = 0;
  /*!
@@ -162,7 +162,7 @@ class Learner : public rabit::Serializable {
   */
  inline void Predict(const SparseBatch::Inst &inst,
                      bool output_margin,
-                      std::vector<float> *out_preds,
+                      std::vector<bst_float> *out_preds,
                      unsigned ntree_limit = 0) const;
  /*!
   * \brief Create a new instance of learner.
@@ -185,7 +185,7 @@ class Learner : public rabit::Serializable {
 // implementation of inline functions.
 inline void Learner::Predict(const SparseBatch::Inst& inst,
                             bool output_margin,
-                             std::vector<float>* out_preds,
+                             std::vector<bst_float>* out_preds,
                             unsigned ntree_limit) const {
  gbm_->Predict(inst, out_preds, ntree_limit);
  if (out_preds->size() == 1) {
--- a/include/xgboost/metric.h
+++ b/include/xgboost/metric.h
@@ -29,9 +29,9 @@ class Metric {
   *        the average statistics across all the node,
   *        this is only supported by some metrics
   */
-  virtual float Eval(const std::vector<float>& preds,
-                     const MetaInfo& info,
-                     bool distributed) const = 0;
+  virtual bst_float Eval(const std::vector<bst_float>& preds,
+                         const MetaInfo& info,
+                         bool distributed) const = 0;
  /*! \return name of metric */
  virtual const char* Name() const = 0;
  /*! \brief virtual destructor */
--- a/include/xgboost/objective.h
+++ b/include/xgboost/objective.h
@@ -41,7 +41,7 @@ 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<float>& preds,
+  virtual void GetGradient(const std::vector<bst_float>& preds,
                           const MetaInfo& info,
                           int iteration,
                           std::vector<bst_gpair>* out_gpair) = 0;
@@ -52,13 +52,13 @@ 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<float> *io_preds) {}
+  virtual void PredTransform(std::vector<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<float> *io_preds) {
+  virtual void EvalTransform(std::vector<bst_float> *io_preds) {
    this->PredTransform(io_preds);
  }
  /*!
@@ -67,7 +67,7 @@ class ObjFunction {
   * used by gradient boosting
   * \return transformed value
   */
-  virtual float ProbToMargin(float base_score) const {
+  virtual bst_float ProbToMargin(bst_float base_score) const {
    return base_score;
  }
  /*!
--- a/include/xgboost/tree_model.h
+++ b/include/xgboost/tree_model.h
@@ -106,7 +106,7 @@ class TreeModel {
      return cleft_ == -1;
    }
    /*! \return get leaf value of leaf node */
-    inline float leaf_value() const {
+    inline bst_float leaf_value() const {
      return (this->info_).leaf_value;
    }
    /*! \return get split condition of the node */
@@ -154,7 +154,7 @@ class TreeModel {
     * \param right right index, could be used to store
     *        additional information
     */
-    inline void set_leaf(float value, int right = -1) {
+    inline void set_leaf(bst_float value, int right = -1) {
      (this->info_).leaf_value = value;
      this->cleft_ = -1;
      this->cright_ = right;
@@ -171,7 +171,7 @@ class TreeModel {
     *        we have split condition
     */
    union Info{
-      float leaf_value;
+      bst_float leaf_value;
      TSplitCond split_cond;
    };
    // pointer to parent, highest bit is used to
@@ -230,7 +230,7 @@ class TreeModel {
   * \param rid node id of the node
   * \param value new leaf value
   */
-  inline void ChangeToLeaf(int rid, float value) {
+  inline void ChangeToLeaf(int rid, bst_float value) {
    CHECK(nodes[nodes[rid].cleft() ].is_leaf());
    CHECK(nodes[nodes[rid].cright()].is_leaf());
    this->DeleteNode(nodes[rid].cleft());
@@ -242,7 +242,7 @@ class TreeModel {
   * \param rid node id of the node
   * \param value new leaf value
   */
-  inline void CollapseToLeaf(int rid, float value) {
+  inline void CollapseToLeaf(int rid, bst_float value) {
    if (nodes[rid].is_leaf()) return;
    if (!nodes[nodes[rid].cleft() ].is_leaf()) {
      CollapseToLeaf(nodes[rid].cleft(), 0.0f);
@@ -338,7 +338,7 @@ class TreeModel {
  }
  /*!
   * \brief add child nodes to node
-   * \param nid node id to add childs
+   * \param nid node id to add children to
   */
  inline void AddChilds(int nid) {
    int pleft  = this->AllocNode();
@@ -398,11 +398,11 @@ class TreeModel {
 /*! \brief node statistics used in regression tree */
 struct RTreeNodeStat {
  /*! \brief loss change caused by current split */
-  float loss_chg;
+  bst_float loss_chg;
  /*! \brief sum of hessian values, used to measure coverage of data */
-  float sum_hess;
+  bst_float sum_hess;
  /*! \brief weight of current node */
-  float base_weight;
+  bst_float base_weight;
  /*! \brief number of child that is leaf node known up to now */
  int leaf_child_cnt;
 };
@@ -426,12 +426,12 @@ class RegTree: public TreeModel<bst_float, RTreeNodeStat> {
    inline void Init(size_t size);
    /*!
     * \brief fill the vector with sparse vector
-     * \param inst The sparse instance to fil.
+     * \param inst The sparse instance to fill.
     */
    inline void Fill(const RowBatch::Inst& inst);
    /*!
     * \brief drop the trace after fill, must be called after fill.
-     * \param inst The sparse instanc to drop.
+     * \param inst The sparse instance to drop.
     */
    inline void Drop(const RowBatch::Inst& inst);
    /*!
@@ -439,7 +439,7 @@ class RegTree: public TreeModel<bst_float, RTreeNodeStat> {
     * \param i feature index.
     * \return the i-th feature value
     */
-    inline float fvalue(size_t i) const;
+    inline bst_float fvalue(size_t i) const;
    /*!
     * \brief check whether i-th entry is missing
     * \param i feature index.
@@ -453,7 +453,7 @@ class RegTree: public TreeModel<bst_float, RTreeNodeStat> {
     *  when flag == -1, this indicate the value is missing
     */
    union Entry {
-      float fvalue;
+      bst_float fvalue;
      int flag;
    };
    std::vector<Entry> data;
@@ -471,14 +471,14 @@ class RegTree: public TreeModel<bst_float, RTreeNodeStat> {
   * \param root_id starting root index of the instance
   * \return the leaf index of the given feature
   */
-  inline float Predict(const FVec& feat, unsigned root_id = 0) const;
+  inline bst_float Predict(const FVec& feat, unsigned root_id = 0) const;
  /*!
   * \brief get next position of the tree given current pid
   * \param pid Current node id.
   * \param fvalue feature value if not missing.
   * \param is_unknown Whether current required feature is missing.
   */
-  inline int GetNext(int pid, float fvalue, bool is_unknown) const;
+  inline int GetNext(int pid, bst_float fvalue, bool is_unknown) const;
  /*!
   * \brief dump the model in the requested format as a text string
   * \param fmap feature map that may help give interpretations of feature
@@ -513,7 +513,7 @@ inline void RegTree::FVec::Drop(const RowBatch::Inst& inst) {
  }
 }

-inline float RegTree::FVec::fvalue(size_t i) const {
+inline bst_float RegTree::FVec::fvalue(size_t i) const {
  return data[i].fvalue;
 }

@@ -530,14 +530,14 @@ inline int RegTree::GetLeafIndex(const RegTree::FVec& feat, unsigned root_id) co
  return pid;
 }

-inline float RegTree::Predict(const RegTree::FVec& feat, unsigned root_id) const {
+inline bst_float RegTree::Predict(const RegTree::FVec& feat, unsigned root_id) const {
  int pid = this->GetLeafIndex(feat, root_id);
  return (*this)[pid].leaf_value();
 }

 /*! \brief get next position of the tree given current pid */
-inline int RegTree::GetNext(int pid, float fvalue, bool is_unknown) const {
-  float split_value = (*this)[pid].split_cond();
+inline int RegTree::GetNext(int pid, bst_float fvalue, bool is_unknown) const {
+  bst_float split_value = (*this)[pid].split_cond();
  if (is_unknown) {
    return (*this)[pid].cdefault();
  } else {