This project is using Blackjack-v0 env from OpenAI Gym.
There are two actions in the space action of this environment, 0 (stand) or 1 (hit).
The state or observation of the env is a tuple
>>> (sum_hand(self.player), self.dealer[0], usable_ace(self.player))- sum_hand(player): the agent's current sum hand
- dealer[0]: value of dealers first card, the second card is not shown
- usable_ace(player): if agent is using ace