Move LSTMTrainer from libtesseract to libtesseract_training

LSTMTrainer is only used for training, so the shared library for
Tesseract can be made smaller.

Signed-off-by: Stefan Weil <sw@weilnetz.de>

CMakeLists.txt

@@ -510,7 +510,6 @@ install(FILES
     src/lstm/input.h
     src/lstm/lstm.h
     src/lstm/lstmrecognizer.h
-    src/lstm/lstmtrainer.h
     src/lstm/maxpool.h
     src/lstm/networkbuilder.h
     src/lstm/network.h

src/lstm/Makefile.am

@@ -28,7 +28,7 @@ pkginclude_HEADERS =
 
 noinst_HEADERS = convolve.h ctc.h
 noinst_HEADERS += fullyconnected.h functions.h input.h
-noinst_HEADERS += lstm.h lstmrecognizer.h lstmtrainer.h maxpool.h
+noinst_HEADERS += lstm.h lstmrecognizer.h maxpool.h
 noinst_HEADERS += network.h networkbuilder.h networkio.h networkscratch.h
 noinst_HEADERS += parallel.h plumbing.h recodebeam.h reconfig.h reversed.h
 noinst_HEADERS += series.h static_shape.h stridemap.h
@@ -38,7 +38,7 @@ noinst_LTLIBRARIES = libtesseract_lstm.la
 
 libtesseract_lstm_la_SOURCES = \
         convolve.cpp ctc.cpp fullyconnected.cpp functions.cpp input.cpp \
-        lstm.cpp lstmrecognizer.cpp lstmtrainer.cpp maxpool.cpp \
+        lstm.cpp lstmrecognizer.cpp maxpool.cpp \
         networkbuilder.cpp network.cpp networkio.cpp \
         parallel.cpp plumbing.cpp recodebeam.cpp reconfig.cpp reversed.cpp \
         series.cpp stridemap.cpp tfnetwork.cpp weightmatrix.cpp

src/lstm/recodebeam.cpp

@@ -31,12 +31,6 @@
 
 namespace tesseract {
 
-// Clipping value for certainty inside Tesseract. Reflects the minimum value
-// of certainty that will be returned by ExtractBestPathAsUnicharIds.
-// Supposedly on a uniform scale that can be compared across languages and
-// engines.
-const float RecodeBeamSearch::kMinCertainty = -20.0f;
-
 // The beam width at each code position.
 const int RecodeBeamSearch::kBeamWidths[RecodedCharID::kMaxCodeLen + 1] = {
     5, 10, 16, 16, 16, 16, 16, 16, 16, 16,

src/lstm/recodebeam.h

@@ -219,7 +219,7 @@ class RecodeBeamSearch {
   // of certainty that will be returned by ExtractBestPathAsUnicharIds.
   // Supposedly on a uniform scale that can be compared across languages and
   // engines.
-  static const float kMinCertainty;
+  static constexpr float kMinCertainty = -20.0f;
   // Number of different code lengths for which we have a separate beam.
   static const int kNumLengths = RecodedCharID::kMaxCodeLen + 1;
   // Total number of beams: dawg/nodawg * number of NodeContinuation * number

src/training/CMakeLists.txt

@@ -179,6 +179,8 @@ set(unicharset_training_src
     lang_model_helpers.h
     lstmtester.cpp
     lstmtester.h
+    lstmtrainer.cpp
+    lstmtrainer.h
     normstrngs.cpp
     normstrngs.h
     unicharset_training_utils.cpp

src/training/Makefile.am

@@ -37,6 +37,7 @@ noinst_HEADERS = \
     lang_model_helpers.h \
     ligature_table.h \
     lstmtester.h \
+    lstmtrainer.h \
     normstrngs.h \
     pango_font_info.h \
     stringrenderer.h \
@@ -71,6 +72,7 @@ libtesseract_training_la_SOURCES = \
     lang_model_helpers.cpp \
     ligature_table.cpp \
     lstmtester.cpp \
+    lstmtrainer.cpp \
     normstrngs.cpp \
     pango_font_info.cpp \
     stringrenderer.cpp \

sw.cpp

@@ -112,6 +112,7 @@ void build(Solution &s)
         "src/training/icuerrorcode.h",
         "src/training/lang_model_helpers.*"_rr,
         "src/training/lstmtester.*"_rr,
+        "src/training/lstmtrainer.*"_rr,
         "src/training/normstrngs.*"_rr,
         "src/training/unicharset_training_utils.*"_rr,
         "src/training/validat.*"_rr;