Chord in jazz solo assignment

I play piano (of coz not professional) and thus wanna know about the details of the model.

Typically, a musical piece consist of

  1. chord
  2. pitch
  3. Time signatures 4/4 3/4

If we use one hot vector to represent a single note, let say we use typical 88 keys pianos (A0,B0, C1 to C8). Can I assume the one hot vector representation represent a single node in 88 keys?
If so, I wonder why when we feed into the LSTM, we use one hot vector rather than “3-4” hot vector". Many time we may add chord with a note to make the music more colorful. So, typically, there are 3 - 4 notes playing at the same time, so should the model feed in
3-4 hot vector

[0 0 0 0 1 0 0 0 0 0 0 0 0 ] Model representation, one hot vector
[0 0 0 0 1 0 1 0 1 0 0 0 0 ] My thinking, 3-4 hot vector

OR I think in a wrong way, chord was taken completely in another perspective in the model?

And more importantly, we need to know the time signature so that we could know when it’s suitable to put chord on a note, normally on the the first note in a bar. Is that learnt by the model or some preprocessing is required as well?

You are thinking in terms of music notation concepts.

This method learns from music performance.