Fine Tuning all Fonts List

[Ibr]

Hi,

for engtrain and engeval they almost have the same command but for eval you specify the font using the argument --font-list, while in train you define the fonts in language-specifics.sh , I ran both command and I noticed that they produce the same results files, except in engtrain case there will be lstmf file for each font in language-specific.sh.

but now there is another question, in the fine tuning the command is:

training/lstmtraining --model_output ~/tesstutorial/impact_from_small/impact \
  --continue_from ~/tesstutorial/engoutput/base_checkpoint \
  --train_listfile ~/tesstutorial/engeval/eng.training_files.txt \
  --max_iterations 1200

and for evaluating its also uses:

training/lstmeval --model ~/tesstutorial/impact_from_small/impact_checkpoint \
  --eval_listfile ~/tesstutorial/engeval/eng.training_files.txt

for me it seems its a typo at the fine tuning command since the argument --eval_listfile  ~/tesstutorial/engeval

so in both cases the commands contains engeval, my question, since I don't want to run the fine tuning many times, so can I use the training files that exists in the engtrain ? instead to create files for the engeval for every font and then run the fine tune against its training files.

Thanks