Implementing Player-Zero training (1/2)

[Klazkin]

Explain the context of the issue, what is being addressed in detail.

The goal

• [x] Appending result (winner) post generation
• [x] BUG evaluation seems to be incorrect.
• [x] Network architecture needs to be adjusted to accommodate more actions.
• [x] BUG why are there INVALID_ACTIONs in generated decks?
• [x] BUG Account for BLOODDRAWING resulting in victory with SHACLES status.
• [x] Add new attributes (attack/defence) to unit generation code.
• [x] Make N action sampler for auto-generation.
  • Should the above condition be random range?

Currently serializes the root state and trains to predict it.

Time tracking

Time Estimate: 5 hours 0 minutes Time spent: 7 hours 35 minutes

Resources

...