Issues

• Allow PPO to train in multi-agent games (more than just 1 opponent)

  agentsenhancementppotraining
  roeey777/Splendor-AI#51 · Filter by author roeey777 opened on Dec 24, 2024

• DQN - add a DQN/DDQN/Rainbow agent

  agentsdqnenhancement
  roeey777/Splendor-AI#27 · Filter by author roeey777 opened on Oct 12, 2024

• Engine - support agent nicknames to shorten execution arguments

  enhancement
  roeey777/Splendor-AI#12 · Filter by author roeey777 opened on Sep 30, 2024

• Genetic - use different representations (maybe ensambles or polynomial regression or even a MLP)

  agentsenhancementgenetic-algorithm
  roeey777/Splendor-AI#11 · Filter by author roeey777 opened on Sep 30, 2024

• PPO2 - add a PPO2 agent

  agentsenhancementppo
  roeey777/Splendor-AI#7 · Filter by author roeey777 opened on Sep 30, 2024

• PPO - log metrices using tensorboard instead/alongside of csv file

  agentsenhancementloggingppotraining
  roeey777/Splendor-AI#4 · Filter by author roeey777 opened on Sep 30, 2024
  1

• PPO - Use multi-processing & vectorized environments to speed-up the learning process

  agentsenhancementppotraining
  roeey777/Splendor-AI#2 · Filter by author roeey777 opened on Sep 30, 2024