mirror/tesseract
1
0
Fork 0
mirror of https://github.com/tesseract-ocr/tesseract.git synced 2024-12-04 01:39:16 +08:00
Code Issues Actions 12 Packages Projects Releases Wiki Activity

Add separator and training_iteration to checkpoint name (#2752)

Browse Source 
* Add separator and training_iteration to checkpoint name
* specify modelname_N.NN_NN_NN.checkpoint for intermediate checkpoint
This commit is contained in:
Shreeshrii 2019-11-09 16:52:40 +05:30 committed by Stefan Weil
parent 944c1d9bff
commit 99dfa8a680
2 changed files with 6 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
doc/lstmeval.1.asc
View File

@ -8,11 +8,11 @@ lstmeval - Evaluation program for LSTM-based networks.
SYNOPSIS
--------
*lstmeval* --model 'lang.lstm|langtrain_checkpoint|pluscharsN.NNN_NN.checkpoint' [--traineddata lang/lang.traineddata] --eval_listfile 'lang.eval_files.txt' [--verbosity N] [--max_image_MB NNNN]
*lstmeval* --model 'lang.lstm|modelname_checkpoint|modelname_N.NN_NN_NN.checkpoint' [--traineddata lang/lang.traineddata] --eval_listfile 'lang.eval_files.txt' [--verbosity N] [--max_image_MB NNNN]
DESCRIPTION
-----------
lstmeval(1) evaluates LSTM-based networks. Either a recognition model or a training checkpoint can be given as input for evaluation along with a list of lstmf files. If evaluating a training checkpoint, '--traineddata' should also be specified.
lstmeval(1) evaluates LSTM-based networks. Either a recognition model or a training checkpoint can be given as input for evaluation along with a list of lstmf files. If evaluating a training checkpoint, '--traineddata' should also be specified. Intermediate training checkpoints can also be used.
OPTIONS
-------

6
src/training/lstmtrainer.cpp
View File

@ -910,11 +910,13 @@ void LSTMTrainer::SaveRecognitionDump(GenericVector<char>* data) const {
}
// Returns a suitable filename for a training dump, based on the model_base_,
// the iteration and the error rates.
// best_error_rate_, best_iteration_ and training_iteration_.
STRING LSTMTrainer::DumpFilename() const {
STRING filename;
filename.add_str_double(model_base_.c_str(), best_error_rate_);
filename += model_base_.c_str();
filename.add_str_double("_", best_error_rate_);
filename.add_str_int("_", best_iteration_);
filename.add_str_int("_", training_iteration_);
filename += ".checkpoint";
return filename;
}