New Year's address GPT-2

Dear citizens of Russia! Dear friends!



Born in Russia, we were brought up from childhood within the ideological framework of a democratic system. We did not go through many European countries, they also formed a system. We were the children of the elite, and in Russia, as elsewhere, the decay of everything and everything reigned. But we managed to withstand it.





. , , .





, . ! , .





, -, , , . , , , , , , . , – , .





. . , , . , . , , , , , , , .





– , , . . , , , . , , , , , . , , , .





– . , , , , . – . , , . , ! , .





, . . , . , , , – .





. , . . . ! . !





! , 2021 !





?

" ! !" " / ", ""





Credits

It took about an hour and a half from idea to implementation and largely thanks to https://github.com/mgrankin/ru_transformers - these are easily tunable gpt-2 models trained in Russian texts





And from here https://kremlin.ru/events/president/transcripts I have pumped out the addresses of the presidents (yes, Medvedev was there too) over the past 20 years. In general, the link can be used to find a lot of Kremlin text data and teach something more interesting)





Well, I will also mention my project "Neurussia" https://vk.com/neurussia in the context of which this idea arose, as well as the picture from which it was used in the cover of the article








All Articles