ML and AI for Games

This repository is a collection of Python code and Jupyter Notebooks showcasing various projects related to card games and artificial intelligence. It explores techniques for:

• Card Detection with Neural Networks: Train neural networks to identify cards in images, enabling computer vision applications for card games.
• Webcam-based Card Reading: Utilize your webcam to capture and recognize cards in real-time, setting the stage for interactive card game experiences.
• Game Coding Examples: Dive into code examples for different card games, providing a foundation for building your own or implementing variations.
• Reinforcement Learning for AI Players: Train AI models using reinforcement learning to play card games strategically, creating intelligent opponents for your games.

Games

The Great Dalmuti

The Great Dalmuti is a card game where players compete to be the first to get rid of all their cards. The game uses a special deck where cards are numbered 1-13, with lower numbers appearing more frequently (1 appears once, 2 appears twice, up to 12 appearing 12 times, and 13 appearing twice as jokers).

Game Rules:

• Players are ranked based on their performance in the previous round
• At the start of each round, the highest-ranked player (Greater Dalmuti) trades two cards with the lowest-ranked player (Greater Peon)
• If there are more than 3 players, the second-highest player (Lesser Dalmuti) trades one card with the second-lowest player (Lesser Peon)
• Players take turns playing cards, trying to match or beat the previous play
• The first player to get rid of all their cards wins the round
• Rankings are updated after each round based on finish order

How to Play:

Install the package and use the dalmuti command to play:

Command Options:

• -n, --num-players: Number of players (default: 4, one of which is human)
• -p, --players: Player definitions in format name:type (can be used multiple times)
  • Types: human, cpu, or cpu_easy
  • Example: -p "Alice:human" -p "Bob:cpu"
• -q, --quiet: Hide game state after each turn (game state is shown by default)
• -g, --num-games: Number of games to play (default: 10)

# Play with default settings (4 players, 1 human + 3 CPU, 10 games)
dalmuti

# Play with a specific number of players
dalmuti -n 6

# Customize players with names and types
dalmuti -p "Alice:human" -p "Bob:cpu" -p "Charlie:cpu_easy" -p "Diana:cpu"

# Hide game state (game state is shown by default)
dalmuti -q

# Play a single game
dalmuti -g 1

# Combine options: play 5 games with 6 players (game state shown by default)
dalmuti -n 6 -g 5