remove init-score parameter (#2776)

* remove related cpp codes

* removed more mentiones of init_score_filename params
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>

docs/Parameters.rst

@@ -591,24 +591,6 @@ IO Parameters
 
    -  **Note**: can be used only in CLI version
 
--  ``initscore_filename`` :raw-html:`<a id="initscore_filename" title="Permalink to this parameter" href="#initscore_filename">&#x1F517;&#xFE0E;</a>`, default = ``""``, type = string, aliases: ``init_score_filename``, ``init_score_file``, ``init_score``, ``input_init_score``
-
-   -  path of file with training initial scores
-
-   -  if ``""``, will use ``train_data_file`` + ``.init`` (if exists)
-
-   -  **Note**: works only in case of loading data directly from file
-
--  ``valid_data_initscores`` :raw-html:`<a id="valid_data_initscores" title="Permalink to this parameter" href="#valid_data_initscores">&#x1F517;&#xFE0E;</a>`, default = ``""``, type = string, aliases: ``valid_data_init_scores``, ``valid_init_score_file``, ``valid_init_score``
-
-   -  path(s) of file(s) with validation initial scores
-
-   -  if ``""``, will use ``valid_data_file`` + ``.init`` (if exists)
-
-   -  separate by ``,`` for multi-validation data
-
-   -  **Note**: works only in case of loading data directly from file
-
 -  ``pre_partition`` :raw-html:`<a id="pre_partition" title="Permalink to this parameter" href="#pre_partition">&#x1F517;&#xFE0E;</a>`, default = ``false``, type = bool, aliases: ``is_pre_partition``
 
    -  used for parallel learning (excluding the ``feature_parallel`` mode)
@@ -1077,11 +1059,9 @@ LightGBM supports continued training with initial scores. It uses an additional
 It means the initial score of the first data row is ``0.5``, second is ``-0.1``, and so on.
 The initial score file corresponds with data file line by line, and has per score per line.
 
-And if the name of data file is ``train.txt``, the initial score file should be named as ``train.txt.init`` and in the same folder as the data file.
+And if the name of data file is ``train.txt``, the initial score file should be named as ``train.txt.init`` and placed in the same folder as the data file.
 In this case, LightGBM will auto load initial score file if it exists.
 
-Otherwise, you should specify the path to the custom named file with initial scores by the ``initscore_filename`` `parameter <#initscore_filename>`__.
-
 Weight Data
 ~~~~~~~~~~~
 

examples/binary_classification/predict.conf

-
 task = predict
 
 data = binary.test

examples/binary_classification/train.conf

@@ -97,9 +97,6 @@ output_model = LightGBM_model.txt
 # output prediction file for predict task
 # output_result= prediction.txt
 
-# support continuous train from initial score file
-# input_init_score= init_score.txt
-
 
 # number of machines in parallel training, alias: num_machine
 num_machines = 1
@@ -110,5 +107,5 @@ local_listen_port = 12400
 # machines list file for parallel training, alias: mlist
 machine_list_file = mlist.txt
 
-# # force splits
+# force splits
 # forced_splits = forced_splits.json

examples/lambdarank/predict.conf

-
 task = predict
 
 data = rank.test

examples/lambdarank/train.conf

@@ -102,9 +102,6 @@ output_model = LightGBM_model.txt
 # output prediction file for predict task
 # output_result= prediction.txt
 
-# support continuous train from initial score file
-# input_init_score= init_score.txt
-
 
 # number of machines in parallel training, alias: num_machine
 num_machines = 1

examples/parallel_learning/predict.conf

-
 task = predict
 
 data = binary.test
 
 input_model= LightGBM_model.txt
-

examples/parallel_learning/train.conf

@@ -97,9 +97,6 @@ output_model = LightGBM_model.txt
 # output prediction file for predict task
 # output_result= prediction.txt
 
-# support continuous train from initial score file
-# input_init_score= init_score.txt
-
 
 # number of machines in parallel training, alias: num_machine
 num_machines = 2

examples/regression/predict.conf

-
 task = predict
 
 data = regression.test

examples/regression/train.conf

@@ -100,9 +100,6 @@ output_model = LightGBM_model.txt
 # output prediction file for predict task
 # output_result= prediction.txt
 
-# support continuous train from initial score file
-# input_init_score= init_score.txt
-
 
 # number of machines in parallel training, alias: num_machine
 num_machines = 1

examples/xendcg/train.conf

@@ -103,9 +103,6 @@ output_model = LightGBM_model.txt
 # output prediction file for predict task
 # output_result= prediction.txt
 
-# support continuous train from initial score file
-# input_init_score= init_score.txt
-
 
 # number of machines in parallel training, alias: num_machine
 num_machines = 1

include/LightGBM/config.h

@@ -549,20 +549,6 @@ struct Config {
   // desc = **Note**: can be used only in CLI version
   std::string output_result = "LightGBM_predict_result.txt";
 
-  // alias = init_score_filename, init_score_file, init_score, input_init_score
-  // desc = path of file with training initial scores
-  // desc = if ``""``, will use ``train_data_file`` + ``.init`` (if exists)
-  // desc = **Note**: works only in case of loading data directly from file
-  std::string initscore_filename = "";
-
-  // alias = valid_data_init_scores, valid_init_score_file, valid_init_score
-  // default = ""
-  // desc = path(s) of file(s) with validation initial scores
-  // desc = if ``""``, will use ``valid_data_file`` + ``.init`` (if exists)
-  // desc = separate by ``,`` for multi-validation data
-  // desc = **Note**: works only in case of loading data directly from file
-  std::vector<std::string> valid_data_initscores;
-
   // alias = is_pre_partition
   // desc = used for parallel learning (excluding the ``feature_parallel`` mode)
   // desc = ``true`` if training data are pre-partitioned, and different machines use different partitions

include/LightGBM/dataset.h

@@ -47,9 +47,8 @@ class Metadata {
   /*!
   * \brief Initialization will load query level informations, since it is need for sampling data
   * \param data_filename Filename of data
-  * \param init_score_filename Filename of initial score
   */
-  void Init(const char* data_filename, const char* initscore_file);
+  void Init(const char* data_filename);
   /*!
   * \brief init as subset
   * \param metadata Filename of data
@@ -213,7 +212,7 @@ class Metadata {
 
  private:
   /*! \brief Load initial scores from file */
-  void LoadInitialScore(const char* initscore_file);
+  void LoadInitialScore();
   /*! \brief Load wights from file */
   void LoadWeights();
   /*! \brief Load query boundaries from file */

include/LightGBM/dataset_loader.h

@@ -19,13 +19,13 @@ class DatasetLoader {
 
   LIGHTGBM_EXPORT ~DatasetLoader();
 
-  LIGHTGBM_EXPORT Dataset* LoadFromFile(const char* filename, const char* initscore_file, int rank, int num_machines);
+  LIGHTGBM_EXPORT Dataset* LoadFromFile(const char* filename, int rank, int num_machines);
 
-  LIGHTGBM_EXPORT Dataset* LoadFromFile(const char* filename, const char* initscore_file) {
-    return LoadFromFile(filename, initscore_file, 0, 1);
+  LIGHTGBM_EXPORT Dataset* LoadFromFile(const char* filename) {
+    return LoadFromFile(filename, 0, 1);
   }
 
-  LIGHTGBM_EXPORT Dataset* LoadFromFileAlignWithOtherDataset(const char* filename, const char* initscore_file, const Dataset* train_data);
+  LIGHTGBM_EXPORT Dataset* LoadFromFileAlignWithOtherDataset(const char* filename, const Dataset* train_data);
 
   LIGHTGBM_EXPORT Dataset* CostructFromSampleData(double** sample_values,
     int** sample_indices, int num_col, const int* num_per_col,

src/application/application.cpp

@@ -104,12 +104,10 @@ void Application::LoadData() {
   if (config_.is_parallel_find_bin) {
     // load data for parallel training
     train_data_.reset(dataset_loader.LoadFromFile(config_.data.c_str(),
-                                                  config_.initscore_filename.c_str(),
                                                   Network::rank(), Network::num_machines()));
   } else {
     // load data for single machine
-    train_data_.reset(dataset_loader.LoadFromFile(config_.data.c_str(), config_.initscore_filename.c_str(),
-                                                  0, 1));
+    train_data_.reset(dataset_loader.LoadFromFile(config_.data.c_str(), 0, 1));
   }
   // need save binary file
   if (config_.save_binary) {
@@ -136,7 +134,6 @@ void Application::LoadData() {
       auto new_dataset = std::unique_ptr<Dataset>(
         dataset_loader.LoadFromFileAlignWithOtherDataset(
           config_.valid[i].c_str(),
-          config_.valid_data_initscores[i].c_str(),
           train_data_.get()));
       valid_datas_.push_back(std::move(new_dataset));
       // need save binary file
@@ -228,8 +225,7 @@ void Application::Predict() {
     }
     DatasetLoader dataset_loader(config_, nullptr,
                                  config_.num_class, config_.data.c_str());
-    train_data_.reset(dataset_loader.LoadFromFile(config_.data.c_str(), config_.initscore_filename.c_str(),
-                                                  0, 1));
+    train_data_.reset(dataset_loader.LoadFromFile(config_.data.c_str(), 0, 1));
     train_metric_.clear();
     objective_fun_.reset(ObjectiveFunction::CreateObjectiveFunction(config_.objective,
                                                                     config_));

src/c_api.cpp

@@ -597,12 +597,12 @@ int LGBM_DatasetCreateFromFile(const char* filename,
   DatasetLoader loader(config, nullptr, 1, filename);
   if (reference == nullptr) {
     if (Network::num_machines() == 1) {
-      *out = loader.LoadFromFile(filename, "");
+      *out = loader.LoadFromFile(filename);
     } else {
-      *out = loader.LoadFromFile(filename, "", Network::rank(), Network::num_machines());
+      *out = loader.LoadFromFile(filename, Network::rank(), Network::num_machines());
     }
   } else {
-    *out = loader.LoadFromFileAlignWithOtherDataset(filename, "",
+    *out = loader.LoadFromFileAlignWithOtherDataset(filename,
                                                     reinterpret_cast<const Dataset*>(reference));
   }
   API_END();

src/io/config.cpp

@@ -210,12 +210,6 @@ void Config::Set(const std::unordered_map<std::string, std::string>& params) {
   // sort eval_at
   std::sort(eval_at.begin(), eval_at.end());
 
-  if (valid_data_initscores.size() == 0 && valid.size() > 0) {
-    valid_data_initscores = std::vector<std::string>(valid.size(), "");
-  }
-  CHECK(valid.size() == valid_data_initscores.size());
-
-  if (valid_data_initscores.empty()) {
   std::vector<std::string> new_valid;
   for (size_t i = 0; i < valid.size(); ++i) {
     if (valid[i] != data) {
@@ -226,7 +220,6 @@ void Config::Set(const std::unordered_map<std::string, std::string>& params) {
     }
   }
   valid = new_valid;
-  }
 
   // check for conflicts
   CheckParamConflict();

src/io/config_auto.cpp

@@ -110,13 +110,6 @@ const std::unordered_map<std::string, std::string>& Config::alias_table() {
   {"prediction_name", "output_result"},
   {"pred_name", "output_result"},
   {"name_pred", "output_result"},
-  {"init_score_filename", "initscore_filename"},
-  {"init_score_file", "initscore_filename"},
-  {"init_score", "initscore_filename"},
-  {"input_init_score", "initscore_filename"},
-  {"valid_data_init_scores", "valid_data_initscores"},
-  {"valid_init_score_file", "valid_data_initscores"},
-  {"valid_init_score", "valid_data_initscores"},
   {"is_pre_partition", "pre_partition"},
   {"is_enable_bundle", "enable_bundle"},
   {"bundle", "enable_bundle"},
@@ -242,8 +235,6 @@ const std::unordered_set<std::string>& Config::parameter_set() {
   "snapshot_freq",
   "input_model",
   "output_result",
-  "initscore_filename",
-  "valid_data_initscores",
   "pre_partition",
   "enable_bundle",
   "use_missing",
@@ -478,12 +469,6 @@ void Config::GetMembersFromString(const std::unordered_map<std::string, std::str
 
   GetString(params, "output_result", &output_result);
 
-  GetString(params, "initscore_filename", &initscore_filename);
-
-  if (GetString(params, "valid_data_initscores", &tmp_str)) {
-    valid_data_initscores = Common::Split(tmp_str.c_str(), ',');
-  }
-
   GetBool(params, "pre_partition", &pre_partition);
 
   GetBool(params, "enable_bundle", &enable_bundle);
@@ -668,8 +653,6 @@ std::string Config::SaveMembersToString() const {
   str_buf << "[snapshot_freq: " << snapshot_freq << "]\n";
   str_buf << "[input_model: " << input_model << "]\n";
   str_buf << "[output_result: " << output_result << "]\n";
-  str_buf << "[initscore_filename: " << initscore_filename << "]\n";
-  str_buf << "[valid_data_initscores: " << Common::Join(valid_data_initscores, ",") << "]\n";
   str_buf << "[pre_partition: " << pre_partition << "]\n";
   str_buf << "[enable_bundle: " << enable_bundle << "]\n";
   str_buf << "[use_missing: " << use_missing << "]\n";

src/io/dataset_loader.cpp

@@ -165,7 +165,7 @@ void DatasetLoader::SetHeader(const char* filename) {
   }
 }
 
-Dataset* DatasetLoader::LoadFromFile(const char* filename, const char* initscore_file, int rank, int num_machines) {
+Dataset* DatasetLoader::LoadFromFile(const char* filename, int rank, int num_machines) {
   // don't support query id in data file when training in parallel
   if (num_machines > 1 && !config_.pre_partition) {
     if (group_idx_ > 0) {
@@ -184,7 +184,7 @@ Dataset* DatasetLoader::LoadFromFile(const char* filename, const char* initscore
     }
     dataset->data_filename_ = filename;
     dataset->label_idx_ = label_idx_;
-    dataset->metadata_.Init(filename, initscore_file);
+    dataset->metadata_.Init(filename);
     if (!config_.two_round) {
       // read data to memory
       auto text_data = LoadTextDataToMemory(filename, dataset->metadata_, rank, num_machines, &num_global_data, &used_data_indices);
@@ -227,7 +227,7 @@ Dataset* DatasetLoader::LoadFromFile(const char* filename, const char* initscore
 
 
 
-Dataset* DatasetLoader::LoadFromFileAlignWithOtherDataset(const char* filename, const char* initscore_file, const Dataset* train_data) {
+Dataset* DatasetLoader::LoadFromFileAlignWithOtherDataset(const char* filename, const Dataset* train_data) {
   data_size_t num_global_data = 0;
   std::vector<data_size_t> used_data_indices;
   auto dataset = std::unique_ptr<Dataset>(new Dataset());
@@ -239,7 +239,7 @@ Dataset* DatasetLoader::LoadFromFileAlignWithOtherDataset(const char* filename,
     }
     dataset->data_filename_ = filename;
     dataset->label_idx_ = label_idx_;
-    dataset->metadata_.Init(filename, initscore_file);
+    dataset->metadata_.Init(filename);
     if (!config_.two_round) {
       // read data in memory
       auto text_data = LoadTextDataToMemory(filename, dataset->metadata_, 0, 1, &num_global_data, &used_data_indices);

src/io/metadata.cpp

@@ -20,13 +20,13 @@ Metadata::Metadata() {
   init_score_load_from_file_ = false;
 }
 
-void Metadata::Init(const char* data_filename, const char* initscore_file) {
+void Metadata::Init(const char* data_filename) {
   data_filename_ = data_filename;
   // for lambdarank, it needs query data for partition data in parallel learning
   LoadQueryBoundaries();
   LoadWeights();
   LoadQueryWeights();
-  LoadInitialScore(initscore_file);
+  LoadInitialScore();
 }
 
 Metadata::~Metadata() {
@@ -386,14 +386,12 @@ void Metadata::LoadWeights() {
   weight_load_from_file_ = true;
 }
 
-void Metadata::LoadInitialScore(const char* initscore_file) {
+void Metadata::LoadInitialScore() {
   num_init_score_ = 0;
-  std::string init_score_filename(initscore_file);
-  if (init_score_filename.size() <= 0) {
+  std::string init_score_filename(data_filename_);
   init_score_filename = std::string(data_filename_);
-    // default weight file name
+  // default init_score file name
   init_score_filename.append(".init");
-  }
   TextReader<size_t> reader(init_score_filename.c_str(), false);
   reader.ReadAllLines();
   if (reader.Lines().empty()) {