Pymarl

Visit Tool

PyMARL is an open-source Python framework for deep multi-agent reinforcement learning. It includes implementations of algorithms like QMIX, COMA, and VDN, facilitating research and development in cooperative multi-agent systems.

Claim this tool

2Views

At a glance

Pricing

Open Source

Free tier

Yes

API

Skill level

Technical

About

What is pymarl?

PyMARL is a Python-based, open-source framework developed by WhiRL for deep multi-agent reinforcement learning. It provides implementations of several prominent algorithms, including QMIX for monotonic value function factorisation, COMA for counterfactual multi-agent policy gradients, VDN for value-decomposition networks, IQL for independent Q-learning, and QTRAN for learning to factorize with transformation. The framework is built using PyTorch and integrates with SMAC (StarCraft Multi-Agent Challenge) as its environment, specifically using SC2.4.6.2.69232 for the results in the SMAC paper. PyMARL supports saving and loading trained models, as well as watching StarCraft II replays, making it a comprehensive tool for researchers and developers in the multi-agent RL domain.

Best used for

Ideal for professors and researchers who need to implement and experiment with deep multi-agent reinforcement learning algorithms, develop cooperative multi-agent systems, and replicate results from the StarCraft Multi-Agent Challenge. Especially valuable for those working with PyTorch and the SMAC environment.

Common actions

implement multi-agent RL

research reinforcement learning

develop cooperative AI

experiment with RL algorithms

deepfakelow-code/no-codeopen-sourcecollaborationworkflowsautomated workflowface swappinggithub copilot"AI Agents"

Capabilities

Key features

QMIX algorithm implementation
COMA algorithm implementation
VDN algorithm implementation
PyTorch-based framework
SMAC environment integration
Model saving and loading
StarCraft II replay watching

Target Audience

professor

Integrations

Not yet documented

Pricing & Plans

Open Source

Free

FAQs

What multi-agent reinforcement learning algorithms are included in PyMARL?

PyMARL includes implementations of several key algorithms such as QMIX for monotonic value function factorisation, COMA for counterfactual multi-agent policy gradients, VDN for value-decomposition networks, IQL for independent Q-learning, and QTRAN for learning to factorize with transformation.

Which environment does PyMARL use for its experiments?

PyMARL primarily uses the StarCraft Multi-Agent Challenge (SMAC) as its environment. The results presented in the SMAC paper specifically utilize StarCraft II version SC2.4.6.2.69232 for consistency and reproducibility.

Can I save and load trained models in PyMARL?

Yes, PyMARL supports saving and loading learnt models. You can enable model saving by setting `save_model = True` and adjust the frequency with `save_model_interval`. Models are saved in the 'models' folder within your results directory.

Is it possible to watch StarCraft II replays of trained models?

Yes, PyMARL allows you to save and watch StarCraft II replays. By setting the `save_replay` option and loading a model via `checkpoint_path`, the framework will generate .SC2Replay files. These can be viewed using the pysc2.bin.play command.

Trending

Subcategories trending in AI Agents & Automation

Chatbots & Conversational AI General-Purpose Agents Workflow Agents Personal Assistants RAG & Document AI Voice Agents

Trending

Also listed in

This tool also appears in

Research & Education › Academic Research Coding & Development › Open Source & Models Research & Education › Scientific Computing

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce