When I say Program I mean Taco + Wavenet and all of the packages needed for them to work

PLAN

0. Install linux

1. Set up and learn the program

   0. Set up the program on baseline dataset (LJ dataset)
   1. Set up the program on Polish language dataset (either to find one online or make one myself by cutting up Librivox audiobooks)

2. Set up final dataset

   0. Download all of the "Miodowe Lata episodes"
   1. Find the cleanest audio in source and prepare the dataset (god it will be so tidious if It can't be automated)
   2. Train the program

3. Make a proof of concept funny video

   0. Pull out subtitles from a Miodowe Lata Kit-Show BEZ IWONY 18+ YTP
   1. Use the program to swap the IVONA (popular polish voice synthesiser) voices for the synthesised actor ones
   2. Clean up the documentation and publish results somewehere

4. Prepare the audio music for the YTPMV

   0. Prepare and clean up sounds from the sitcom to use as instruments
   1. Compose the music with the midi as a reference
   2. Mixing, mastering.

5. Prepare audio vocals for the YTPMV

   0. Prepare romaji and phonetic spelling of the song
   1. Put the prepared photetic spelling in the program
   2. Cut up the voices produced to the music

6. Prepare the video for the YTPMV

   0. Learn AE or Blender
   1. Problems to solve when I will get to them