PyTorch implementation of Munchausen Reinforcement Learning based on DQN and SAC. Handles discrete and continuous action spaces

Last update: Mar 10, 2022

Overview

Exploring Munchausen Reinforcement Learning

This is the project repository of my team in the "Advanced Deep Learning for Robotics" course at TUM. Our project's topic is "Exploring Munchausen Reinforcement Learning" based on this paper.

For a detailed discussion, see the report and the final presentation.

Setup

Create a virtual environment.
Run pip3 install -r requirements.txt

Code Structure

This repository is structured as follows:

The directories M-DQN and M-SAC contain the implementations of the RL agents DQN and SAC extended with the Munchausen term, respectively.
The directories rl-baselines3-zoo contains a copy of this repository, where we included the implementations of M-DQN so that we can easily train and test the M-DQN agent on benchmark environments and also compare it to other classical agents. To do so, just follow the steps described in the original repository and insert M-DQN as the agent argument.
The directory particles-envcontains a modified version of this repository. The modified version contains code for a particles environment, where an agent wants to reach a goal, while avoiding obstacles. Besides, M-SAC agent is implemented and included in the code, so that it can be trained and compared to the classical SAC agent.
The directory action-gap contains implementation of callbacks for experiment manager of rl-baselines3-zoo which logs action-gap for tensorboard.

Softlearning is a reinforcement learning framework for training maximum entropy policies in continuous domains. Includes the official implementation of the Soft Actor-Critic algorithm.

Softlearning Softlearning is a deep reinforcement learning toolbox for training maximum entropy policies in continuous domains. The implementation is

997 Dec 30, 2022

DeepMetaHandles: Learning Deformation Meta-Handles of 3D Meshes with Biharmonic Coordinates

DeepMetaHandles (CVPR2021 Oral) [paper] [animations] DeepMetaHandles is a shape deformation technique. It learns a set of meta-handles for each given

73 Dec 15, 2022

A very short and easy implementation of Quantile Regression DQN

Quantile Regression DQN Quantile Regression DQN a Minimal Working Example, Distributional Reinforcement Learning with Quantile Regression (https://arx

80 Sep 17, 2022

A working implementation of the Categorical DQN (Distributional RL).

Categorical DQN. Implementation of the Categorical DQN as described in A distributional Perspective on Reinforcement Learning. Thanks to @tudor-berari

98 Sep 20, 2022

Official implementation of ACTION-Net: Multipath Excitation for Action Recognition (CVPR'21).

ACTION-Net Official implementation of ACTION-Net: Multipath Excitation for Action Recognition (CVPR'21). Getting Started EgoGesture data folder struct

171 Dec 26, 2022

[ICML 2020] Prediction-Guided Multi-Objective Reinforcement Learning for Continuous Robot Control

PG-MORL This repository contains the implementation for the paper Prediction-Guided Multi-Objective Reinforcement Learning for Continuous Robot Contro

65 Jan 7, 2023

This is the code of using DQN to play Sekiro .

Update for using DQN to play sekiro 2021.2.2（English Version） This is the code of using DQN to play Sekiro . I am very glad to tell that I have writen

144 Dec 25, 2022

Allows including an action inside another action (by preprocessing the Yaml file). This is how composite actions should have worked.

actions-includes Allows including an action inside another action (by preprocessing the Yaml file). Instead of using uses or run in your action step,

70 Nov 4, 2022

Human Action Controller - A human action controller running on different platforms.

Human Action Controller (HAC) Goal A human action controller running on different platforms. Fun Easy-to-use Accurate Anywhere Fun Examples Mouse Cont

27 Jul 20, 2022

Comments

how use pip3 in virtual env?

I creat conda virtual,but cannt use pip3 to install requirements. and cannt install pip3 in conda vitrtual.so what you method to use pip3 in conda env?

opened by wangchangquan 0

PyTorch implementation of Munchausen Reinforcement Learning based on DQN and SAC. Handles discrete and continuous action spaces

Related tags

Overview

Exploring Munchausen Reinforcement Learning

Setup

Code Structure

You might also like...

Softlearning is a reinforcement learning framework for training maximum entropy policies in continuous domains. Includes the official implementation of the Soft Actor-Critic algorithm.

DeepMetaHandles: Learning Deformation Meta-Handles of 3D Meshes with Biharmonic Coordinates

A very short and easy implementation of Quantile Regression DQN

A working implementation of the Categorical DQN (Distributional RL).

Official implementation of ACTION-Net: Multipath Excitation for Action Recognition (CVPR'21).

[ICML 2020] Prediction-Guided Multi-Objective Reinforcement Learning for Continuous Robot Control

This is the code of using DQN to play Sekiro .

Allows including an action inside another action (by preprocessing the Yaml file). This is how composite actions should have worked.

Human Action Controller - A human action controller running on different platforms.

Comments

how use pip3 in virtual env?

Owner

Mohamed Amine Ketata

Benchmark spaces - Benchmarks of how well different two dimensional spaces work for clustering algorithms

Quantile Regression DQN a Minimal Working Example, Distributional Reinforcement Learning with Quantile Regression

A clean and robust Pytorch implementation of PPO on continuous action space.

Official Pytorch Implementation of 'Learning Action Completeness from Points for Weakly-supervised Temporal Action Localization' (ICCV-21 Oral)

An algorithm that handles large-scale aerial photo co-registration, based on SURF, RANSAC and PyTorch autograd.

MLOps will help you to understand how to build a Continuous Integration and Continuous Delivery pipeline for an ML/AI project.

The official TensorFlow implementation of the paper Action Transformer: A Self-Attention Model for Short-Time Pose-Based Human Action Recognition

On the model-based stochastic value gradient for continuous reinforcement learning

Predicting path with preference based on user demonstration using Maximum Entropy Deep Inverse Reinforcement Learning in a continuous environment

An unopinionated replacement for PyTorch's Dataset and ImageFolder, that handles Tar archives