Kaixhin/ACER

Stars
250
Rank 161,479 (Top 4 %)
Language
Python
License
MIT License
Created over 7 years ago
Updated almost 2 years ago

Kaixhin/ACER

Kaixhin

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Actor-critic with experience replay

ACER

Actor-critic with experience replay (ACER) [1]. Uses batch off-policy updates to improve stability. Trust region updates can be enabled with --trust-region. Currently uses full trust region instead of "efficient" trust region (see issue #1).

Run with python main.py <options>. To run asynchronous advantage actor-critic (A3C) [2] (but with a Q-value head), use the --on-policy option.

Requirements

To install all dependencies with Anaconda run conda env create -f environment.yml and use source activate acer to activate the environment.

Results

Acknowledgements

@ikostrikov for pytorch-a3c
@apaszke for Reinforcement Learning (DQN) tutorial
@pfnet for ChainerRL

References

[1] Sample Efficient Actor-Critic with Experience Replay
[2] Asynchronous Methods for Deep Reinforcement Learning

Rainbow

Rainbow: Combining Improvements in Deep Reinforcement Learning

grokking-pytorch

The Hitchiker's Guide to PyTorch

dockerfiles

Compilation of Dockerfiles with automated builds enabled on the Docker Registry

Autoencoders

Torch implementations of various types of autoencoders

PlaNet

Deep Planning Network: Control from pixels by latent planning with learned dynamics

imitation-learning

Imitation learning algorithms

Atari

Persistent advantage learning dueling double DQN for the Arcade Learning Environment

FGLab

Future Gadget Laboratory

spinning-up-basic

Basic versions of agents from Spinning Up in Deep RL written in PyTorch

FCN-semantic-segmentation

Fully convolutional networks for semantic segmentation

NoisyNet-A3C

Noisy Networks for Exploration

nninit

Weight initialisation schemes for Torch7 neural network modules

rlenvs

Reinforcement learning environments for Torch7

FGMachine

Future Gadget Machine

malmo-challenge

Malmo Collaborative AI Challenge - Team Pig Catcher

torch-pastalog

A Torch interface for pastalog - simple, realtime visualization of neural network training performance

GUDRL

Generalised UDRL

Dist-A3C

Distributed A3C

EC

Episodic Control

human-level-control

Presentation on Human-Level Control Through Deep Reinforcement Learning

Easy21

Reinforcement Learning Assignment: Easy21

end-to-end

Presentation on End-to-End Training of Deep Visuomotor Policies

docker-torch-mega

Docker image for Torch with CUDA support + extra Torch libraries

cuda-workshop

SARCOS

ML models trained on the SARCOS dataset

IncSFA

Incremental Slow Feature Analysis

sybilsystem

MATLAB Deep Learning Library

MCAC

Minimal Criterion Artist Collective

GlassMate

Team Inforaptor's project for IC Hack '14

bakapunk

A tool for finding similar songs in your music library