pr-shukla / maddpg-keras

Implementation Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm in keras

• 👋 Hi, I’m Prashant Shukla
• 👀 I’m ML Engineer with 3 year experience
• 🌱 I’m interested in Cognitive Science and AI
• 💞️ My area of expertise is Statistics, NLP, Reinforcement Learning, Graph ML, Computer Vision
• 📫 You can reach me on my Email: [email protected]

maddpg-keras's Issues

[BUG] Update previous state to new state

• While updating train.py for flexible number of agents, updating of prev_state to new_state was deleted by mistake
• Update code to restore updating of prev_state

[ENH] Add flexible number of agents feature for training

• Currently this implementation supports only 3 agents during training, with this enhancement it can support any number of agents decided by user during training

[DOC] Update Readme with new updates

[ENH] Add requirements.txt

[ENH] Update directory structure

• Seperate train and predict files in different folders

[DOC] Add docstring and algorithm logic in comments

• Add logic of codes in comments, to enable understanding of implementation for new user
• Add docstrings to each functions