added an option to save the output vectors

Summary:
added an option to save the output vectors too. -saveOutput 1 will
save the output vectors along with the input ones. Saves to a file
called [something].output

Reviewed By: ajoulin

Differential Revision: D3923702

fbshipit-source-id: c2fb1cd0f6d51fcd9ed95fd54ee7e44c5e6fe861

src/args.cc

@@ -36,6 +36,7 @@ Args::Args() {
   label = "__label__";
   verbose = 2;
   pretrainedVectors = "";
+  saveOutput = 0;
 }
 
 void Args::parseArgs(int argc, char** argv) {
@@ -113,6 +114,8 @@ void Args::parseArgs(int argc, char** argv) {
       verbose = atoi(argv[ai + 1]);
     } else if (strcmp(argv[ai], "-pretrainedVectors") == 0) {
       pretrainedVectors = std::string(argv[ai + 1]);
+    } else if (strcmp(argv[ai], "-saveOutput") == 0) {
+      saveOutput = atoi(argv[ai + 1]);
     } else {
       std::cout << "Unknown argument: " << argv[ai] << std::endl;
       printHelp();
@@ -158,6 +161,7 @@ void Args::printHelp() {
     << "  -label              labels prefix [" << label << "]\n"
     << "  -verbose            verbosity level [" << verbose << "]\n"
     << "  -pretrainedVectors  pretrained word vectors for supervised learning []"
+    << "  -saveOutput         whether output params should be saved [" << saveOutput << "]\n"
     << std::endl;
 }
 

src/args.h

@@ -44,6 +44,7 @@ class Args {
     std::string label;
     int verbose;
     std::string pretrainedVectors;
+    int saveOutput;
 
     void parseArgs(int, char**);
     void printHelp();

src/fasttext.cc

@@ -47,6 +47,23 @@ void FastText::saveVectors() {
   ofs.close();
 }
 
+void FastText::saveOutput() {
+  std::ofstream ofs(args_->output + ".output");
+  if (!ofs.is_open()) {
+    std::cout << "Error opening file for saving vectors." << std::endl;
+    exit(EXIT_FAILURE);
+  }
+  ofs << dict_->nwords() << " " << args_->dim << std::endl;
+  Vector vec(args_->dim);
+  for (int32_t i = 0; i < dict_->nwords(); i++) {
+    std::string word = dict_->getWord(i);
+    vec.zero();
+    vec.addRow(*output_, i);
+    ofs << word << " " << vec << std::endl;
+  }
+  ofs.close();
+}
+
 void FastText::saveModel() {
   std::ofstream ofs(args_->output + ".bin", std::ofstream::binary);
   if (!ofs.is_open()) {
@@ -365,6 +382,9 @@ void FastText::train(std::shared_ptr<Args> args) {
   saveModel();
   if (args_->model != model_name::sup) {
     saveVectors();
+    if (args_->saveOutput > 0) {
+      saveOutput();
+    }
   }
 }
 

src/fasttext.h

@@ -38,6 +38,7 @@ class FastText {
   public:
     void getVector(Vector&, const std::string&);
     void saveVectors();
+    void saveOutput();
     void saveModel();
     void loadModel(const std::string&);
     void loadModel(std::istream&);