tesseract::LSTMTester Class Reference

#include <lstmtester.h>

Public Member Functions
	LSTMTester (int64_t max_memory)
bool	LoadAllEvalData (const STRING &filenames_file)
bool	LoadAllEvalData (const GenericVector< STRING > &filenames)
STRING	RunEvalAsync (int iteration, const double *training_errors, const TessdataManager &model_mgr, int training_stage)
STRING	RunEvalSync (int iteration, const double *training_errors, const TessdataManager &model_mgr, int training_stage, int verbosity)

Detailed Description

Definition at line 28 of file lstmtester.h.

Constructor & Destructor Documentation

LSTMTester()

tesseract::LSTMTester::LSTMTester

(

int64_t

max_memory

)

Definition at line 25 of file lstmtester.cpp.

    : test_data_(max_memory) {}

Member Function Documentation

LoadAllEvalData() [1/2]

bool tesseract::LSTMTester::LoadAllEvalData

(

const GenericVector< STRING > &

filenames

)

Definition at line 44 of file lstmtester.cpp.

                                                                       {
  test_data_.Clear();
  bool result = test_data_.LoadDocuments(filenames, CS_SEQUENTIAL, nullptr);
  total_pages_ = test_data_.TotalPages();
  return result;
}

LoadAllEvalData() [2/2]

bool tesseract::LSTMTester::LoadAllEvalData

(

const STRING &

filenames_file

)

Definition at line 31 of file lstmtester.cpp.

                                                             {
  GenericVector<STRING> filenames;
  if (!LoadFileLinesToStrings(filenames_file.c_str(), &filenames)) {
    tprintf("Failed to load list of eval filenames from %s\n",
            filenames_file.c_str());
    return false;
  }
  return LoadAllEvalData(filenames);
}

RunEvalAsync()

STRING tesseract::LSTMTester::RunEvalAsync	(	int	iteration,
		const double *	training_errors,
		const TessdataManager &	model_mgr,
		int	training_stage
	)

Definition at line 53 of file lstmtester.cpp.

                                                    {
  STRING result;
  if (total_pages_ == 0) {
    result.add_str_int("No test data at iteration ", iteration);
    return result;
  }
  if (!LockIfNotRunning()) {
    result.add_str_int("Previous test incomplete, skipping test at iteration ",
                       iteration);
    return result;
  }
  // Save the args.
  STRING prev_result = test_result_;
  test_result_ = "";
  if (training_errors != nullptr) {
    test_iteration_ = iteration;
    test_training_errors_ = training_errors;
    test_model_mgr_ = model_mgr;
    test_training_stage_ = training_stage;
    std::thread t(&LSTMTester::ThreadFunc, this);
    t.detach();
  } else {
    UnlockRunning();
  }
  return prev_result;
}

RunEvalSync()

STRING tesseract::LSTMTester::RunEvalSync	(	int	iteration,
		const double *	training_errors,
		const TessdataManager &	model_mgr,
		int	training_stage,
		int	verbosity
	)

Definition at line 84 of file lstmtester.cpp.

                                                                  {
  LSTMTrainer trainer;
  trainer.InitCharSet(model_mgr);
  TFile fp;
  if (!model_mgr.GetComponent(TESSDATA_LSTM, &fp) ||
      !trainer.DeSerialize(&model_mgr, &fp)) {
    return "Deserialize failed";
  }
  int eval_iteration = 0;
  double char_error = 0.0;
  double word_error = 0.0;
  int error_count = 0;
  while (error_count < total_pages_) {
    const ImageData* trainingdata = test_data_.GetPageBySerial(eval_iteration);
    trainer.SetIteration(++eval_iteration);
    NetworkIO fwd_outputs, targets;
    Trainability result =
        trainer.PrepareForBackward(trainingdata, &fwd_outputs, &targets);
    if (result != UNENCODABLE) {
      char_error += trainer.NewSingleError(tesseract::ET_CHAR_ERROR);
      word_error += trainer.NewSingleError(tesseract::ET_WORD_RECERR);
      ++error_count;
      if (verbosity > 1 || (verbosity > 0 && result != PERFECT)) {
        tprintf("Truth:%s\n", trainingdata->transcription().c_str());
        GenericVector<int> ocr_labels;
        GenericVector<int> xcoords;
        trainer.LabelsFromOutputs(fwd_outputs, &ocr_labels, &xcoords);
        STRING ocr_text = trainer.DecodeLabels(ocr_labels);
        tprintf("OCR  :%s\n", ocr_text.c_str());
      }
    }
  }
  char_error *= 100.0 / total_pages_;
  word_error *= 100.0 / total_pages_;
  STRING result;
  result.add_str_int("At iteration ", iteration);
  result.add_str_int(", stage ", training_stage);
  result.add_str_double(", Eval Char error rate=", char_error);
  result.add_str_double(", Word error rate=", word_error);
  return result;
}

---

The documentation for this class was generated from the following files:

• /tesseract/src/training/lstmtester.h
• /tesseract/src/training/lstmtester.cpp