GitPlanet

Cheap and reliable Node.js hosting starts at $3/month, and $1/month static HTML hosting

Created with love in Canada, visit hostnodejs.com today

Feel like to post an Ad? Learn Details
All Projects → mengf1 → Dher

mengf1 / Dher

DHER: Hindsight Experience Replay for Dynamic Goals (ICLR-2019)

Programming Languages

python
139335 projects - #7 most used programming language

Labels

reinforcement-learning

Projects that are alternatives of or similar to Dher

Fullstackmachinelearning
Mostly free resources for end-to-end machine learning engineering, including open courses from CalTech, Columbia, Berkeley, MIT, and Stanford (in alphabetical order).
Stars: ✭ 39 (-18.75%)
Mutual labels:  reinforcement-learning
Bullet3
Bullet Physics SDK: real-time collision detection and multi-physics simulation for VR, games, visual effects, robotics, machine learning etc.
Stars: ✭ 8,714 (+18054.17%)
Mutual labels:  reinforcement-learning
Biped trajectory optimization
Implementing trajectory optimization on bipedal system
Stars: ✭ 45 (-6.25%)
Mutual labels:  reinforcement-learning
Bolero
Behavior Optimization and Learning for Robots
Stars: ✭ 39 (-18.75%)
Mutual labels:  reinforcement-learning
Machine Learning From Scratch
Succinct Machine Learning algorithm implementations from scratch in Python, solving real-world problems (Notebooks and Book). Examples of Logistic Regression, Linear Regression, Decision Trees, K-means clustering, Sentiment Analysis, Recommender Systems, Neural Networks and Reinforcement Learning.
Stars: ✭ 42 (-12.5%)
Mutual labels:  reinforcement-learning
Gym Gridworlds
Gridworld environments for OpenAI gym.
Stars: ✭ 43 (-10.42%)
Mutual labels:  reinforcement-learning
Neural Architecture Search With Rl
Minimal Tensorflow implementation of the paper "Neural Architecture Search With Reinforcement Learning" presented at ICLR 2017
Stars: ✭ 37 (-22.92%)
Mutual labels:  reinforcement-learning
Reinforcement learning
Predict/log/learn/update reinforcement learning loop
Stars: ✭ 47 (-2.08%)
Mutual labels:  reinforcement-learning
Deep Q Learning
Minimal Deep Q Learning (DQN & DDQN) implementations in Keras
Stars: ✭ 1,013 (+2010.42%)
Mutual labels:  reinforcement-learning
Async Deeprl
Playing Atari games with TensorFlow implementation of Asynchronous Deep Q-Learning
Stars: ✭ 44 (-8.33%)
Mutual labels:  reinforcement-learning
Blocks
Blocks World -- Simulator, Code, and Models (Misra et al. EMNLP 2017)
Stars: ✭ 39 (-18.75%)
Mutual labels:  reinforcement-learning
Awesome Mlss
List of summer schools in machine learning + related fields across the globe
Stars: ✭ 1,001 (+1985.42%)
Mutual labels:  reinforcement-learning
Orbit
Open source collection of Reinforcement Learning Environments.
Stars: ✭ 44 (-8.33%)
Mutual labels:  reinforcement-learning
Deepbootcamp
Solved lab problems, slides and notes of the Deep Reinforcement Learning bootcamp 2017 held at UCBerkeley
Stars: ✭ 39 (-18.75%)
Mutual labels:  reinforcement-learning
Ml In Tf
Get started with Machine Learning in TensorFlow with a selection of good reads and implemented examples!
Stars: ✭ 45 (-6.25%)
Mutual labels:  reinforcement-learning
Deepqlearning.jl
Implementation of the Deep Q-learning algorithm to solve MDPs
Stars: ✭ 38 (-20.83%)
Mutual labels:  reinforcement-learning
Tf2.0 reinforcement learning
Basic reinforcement learning implementation with tensorflow version 2.0
Stars: ✭ 43 (-10.42%)
Mutual labels:  reinforcement-learning
Mujocounity
Reproducing MuJoCo benchmarks in a modern, commercial game /physics engine (Unity + PhysX).
Stars: ✭ 47 (-2.08%)
Mutual labels:  reinforcement-learning
Deep traffic
MIT DeepTraffic top 2% solution (75.01 mph) 🚗.
Stars: ✭ 47 (-2.08%)
Mutual labels:  reinforcement-learning
Deterministic Gail Pytorch
PyTorch implementation of Deterministic Generative Adversarial Imitation Learning (GAIL) for Off Policy learning
Stars: ✭ 44 (-8.33%)
Mutual labels:  reinforcement-learning
View All Similar Projects ➔

Dy Envs

There are dynamic goal environments. We modify the robotic manipulation environments created by OpenAI (Brockman et al., 2016) for our experiments.

Tasks

As shown in above figure, we assign certain rules to the goals so that they accordingly move in the environments while an agent is required to control the robotic arm's grippers to reach the goal that moves along a straight line (Dy-Reaching), to reach the goal that moves in a circle (Dy-Circling), or to push a block to the goal that moves along a straight line (Dy-Pushing).

NOTE: The first three tasks need mujoco. The fourth task does not need mujoco but pygame. It is cheaper to only install Dy-Snake.

How to install it

Our environments depend on openai gym. Please install gym (ver-0.10.9) at first.

cd dygym
python install -e .

Test new environments

cd dygym/test
python test_dyreach.py

DHER

Our algorithms depend on openai baselines. Please install baselines (ver-0.1.5) at first.

How to install it

cd dher
python install -e .

Use DDPG + DHER

cd dher/ddpg_dher/experiment
python train_dyreach.py

Use DQN + DHER

cd dher/dqn_dher/experiment
python train_dysnake.py

NOTE: In Dy-Snake, the first four digits of an observation indicate achieved goals and desired goals. Our implementation of DQN+DHER uses this trick.

Key idea - failed twice then success

Idea

Poster

This browser does not support PDFs. Please download the PDF to view it: Download PDF.

Citation

Please cite our ICLR paper if you use this repository in your publications:

@inproceedings{
fang2019dher,
title={{DHER}: Hindsight Experience Replay for Dynamic Goals},
author={Meng Fang and Cheng Zhou and Bei Shi and Boqing Gong and Jia Xu and Tong Zhang},
booktitle={International Conference on Learning Representations},
year={2019},
url={https://openreview.net/forum?id=Byf5-30qFX},
}

Licence

The MIT License

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].