Closed 0Over0 closed 1 week ago
Напиши мне в тг. Так будет быстрее. @bropines
Я отправил приглашение. Я раньше никогда телеграмом не пользовался
@dmMaze Google translator died, I also get an error.
[ERROR ] base:translate:165 - This translator seems to messed up the translation which resulted in inconsistent translated line count.
Set concate_text to False or change textblk_break in the source code may solve the problem.
Traceback (most recent call last):
File "G:\Ballon-translator-portable-main\ui\module_manager.py", line 486, in run
self.job()
File "G:\Ballon-translator-portable-main\ui\module_manager.py", line 310, in <lambda>
self.job = lambda : self._blktrans_pipeline(blk_list, tgt_img, mode, blk_ids, tgt_mask)
File "G:\Ballon-translator-portable-main\ui\module_manager.py", line 322, in _blktrans_pipeline
self.translate_thread.module.translate_textblk_lst(blk_list)
File "G:\Ballon-translator-portable-main\modules\translators\base.py", line 205, in translate_textblk_lst
_translations = self.translate(text_list)
File "G:\Ballon-translator-portable-main\modules\translators\base.py", line 163, in translate
assert len(text_trans) == len(text)
AssertionError
@bropines It's not dead:
base:translate:165 - This translator seems to messed up the translation which resulted in inconsistent translated line count. Set concate_text to False or change textblk_break in the source code may solve the problem.
It's not dead:
Well, it’s different for me. In 2 mangas the effect is the same, in the third everything is fine. He doesn't give out the data clearly. I can add logs and see what went wrong
@bropines It's not dead:
base:translate:165 - This translator seems to messed up the translation which resulted in inconsistent translated line count. Set concate_text to False or change textblk_break in the source code may solve the problem.
Короче. Решил я потестить что и как. Это была какая-то вчерашная приколюха от гугла. Потому что сейчас он работает на любом кол-ве текста без нареканий. Либо гугл сломали фронтенд, либо мы просто вовремя пришли на их баги. Сейчас я прогнал с десяток комиксов, все великолепно работает (правда долго)
Briefly speaking. I decided to test what and how. It was some kind of yesterday's prank from Google. Because now it works on any amount of text without any complaints. Either Google broke the frontend, or we just caught their bugs in time. Now I've run about a dozen comics, everything works great (though it takes a long time)
For example, out of 100 pages, 12 may not be translated. Moreover, if I try to translate these pages separately, the result will be the same. There are no translation problems on other scanners. (If anything, I speak Russian)