remove the useless training.

src/boosting/gbdt.cpp

@@ -161,6 +161,7 @@ void GBDT::ResetTrainingData(const BoostingConfig* config, const Dataset* train_
   if (train_data_ != nullptr) {
     // reset config for tree learner
     tree_learner_->ResetConfig(&new_config->tree_config);
+    is_class_end_ = std::vector<bool>(num_class_, false);
   }
   gbdt_config_.reset(new_config.release());
 }
@@ -336,31 +337,35 @@ bool GBDT::TrainOneIter(const score_t* gradient, const score_t* hessian, bool is
     sub_gradient_time += std::chrono::steady_clock::now() - start_time;
 #endif
   }
-  bool shouldContinue = false;
+  bool should_continue = false;
   for (int curr_class = 0; curr_class < num_class_; ++curr_class) {
 #ifdef TIMETAG
     start_time = std::chrono::steady_clock::now();
 #endif
-    // train a new tree
-    std::unique_ptr<Tree> new_tree(tree_learner_->Train(gradient + curr_class * num_data_, hessian + curr_class * num_data_));
+    std::unique_ptr<Tree> new_tree(new Tree(2));
+    if (!is_class_end_[curr_class]) {
+      // train a new tree
+      new_tree.reset(tree_learner_->Train(gradient + curr_class * num_data_, hessian + curr_class * num_data_));
+    }
 #ifdef TIMETAG
     tree_time += std::chrono::steady_clock::now() - start_time;
 #endif
 
     if (new_tree->num_leaves() > 1) {
-      shouldContinue = true;
+      should_continue = true;
+      // shrinkage by learning rate
+      new_tree->Shrinkage(shrinkage_rate_);
+      // update score
+      UpdateScore(new_tree.get(), curr_class);
+      UpdateScoreOutOfBag(new_tree.get(), curr_class);
+    } else {
+      is_class_end_[curr_class] = true;
     }
 
-    // shrinkage by learning rate
-    new_tree->Shrinkage(shrinkage_rate_);
-    // update score
-    UpdateScore(new_tree.get(), curr_class);
-    UpdateScoreOutOfBag(new_tree.get(), curr_class);
-
     // add model
     models_.push_back(std::move(new_tree));
   }
-  if (!shouldContinue) {
+  if (!should_continue) {
     Log::Warning("Stopped training because there are no more leaves that meet the split requirements.");
     for (int curr_class = 0; curr_class < num_class_; ++curr_class) {
       models_.pop_back();
@@ -392,6 +397,7 @@ void GBDT::RollbackOneIter() {
   for (int curr_class = 0; curr_class < num_class_; ++curr_class) {
     models_.pop_back();
   }
+  is_class_end_ = std::vector<bool>(num_class_, false);
   --iter_;
 }
 

src/boosting/gbdt.h

@@ -344,6 +344,7 @@ protected:
   std::vector<data_size_t> right_write_pos_buf_;
   std::unique_ptr<Dataset> tmp_subset_;
   bool is_use_subset_;
+  std::vector<bool> is_class_end_;
 };
 
 }  // namespace LightGBM