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 β†’ gsurma β†’ Atari

gsurma / Atari

Licence: mit
AI research environment for the Atari 2600 games πŸ€–.

Programming Languages

python
139335 projects - #7 most used programming language
python2
120 projects

Labels

machine-learningartificial-intelligencereinforcement-learningaimldqngymrl

Projects that are alternatives of or similar to Atari

Ros2learn
ROS 2 enabled Machine Learning algorithms
Stars: ✭ 119 (-31.61%)
Mutual labels:  reinforcement-learning, ml, dqn, rl
Polyaxon
Machine Learning Platform for Kubernetes (MLOps tools for experimentation and automation)
Stars: ✭ 2,966 (+1604.6%)
Mutual labels:  artificial-intelligence, ai, reinforcement-learning, ml
Rlenv.directory
Explore and find reinforcement learning environments in a list of 150+ open source environments.
Stars: ✭ 79 (-54.6%)
Mutual labels:  gym, reinforcement-learning, rl
Torchrl
Pytorch Implementation of Reinforcement Learning Algorithms ( Soft Actor Critic(SAC)/ DDPG / TD3 /DQN / A2C/ PPO / TRPO)
Stars: ✭ 90 (-48.28%)
Mutual labels:  gym, reinforcement-learning, dqn
Doudizhu
AIζ–—εœ°δΈ»
Stars: ✭ 149 (-14.37%)
Mutual labels:  ai, reinforcement-learning, dqn
Trading Gym
A Trading environment base on Gym
Stars: ✭ 71 (-59.2%)
Mutual labels:  gym, reinforcement-learning, rl
Muzero General
MuZero
Stars: ✭ 1,187 (+582.18%)
Mutual labels:  gym, reinforcement-learning, rl
Rl Baselines3 Zoo
A collection of pre-trained RL agents using Stable Baselines3, training and hyperparameter optimization included.
Stars: ✭ 161 (-7.47%)
Mutual labels:  gym, reinforcement-learning, rl
Rl Baselines Zoo
A collection of 100+ pre-trained RL agents using Stable Baselines, training and hyperparameter optimization included.
Stars: ✭ 839 (+382.18%)
Mutual labels:  gym, reinforcement-learning, rl
Stable Baselines
Mirror of Stable-Baselines: a fork of OpenAI Baselines, implementations of reinforcement learning algorithms
Stars: ✭ 115 (-33.91%)
Mutual labels:  gym, reinforcement-learning, rl
Modelchimp
Experiment tracking for machine and deep learning projects
Stars: ✭ 121 (-30.46%)
Mutual labels:  artificial-intelligence, ai, ml
Pycm
Multi-class confusion matrix library in Python
Stars: ✭ 1,076 (+518.39%)
Mutual labels:  artificial-intelligence, ai, ml
Deep traffic
MIT DeepTraffic top 2% solution (75.01 mph) πŸš—.
Stars: ✭ 47 (-72.99%)
Mutual labels:  artificial-intelligence, reinforcement-learning, dqn
Caffe2
Caffe2 is a lightweight, modular, and scalable deep learning framework.
Stars: ✭ 8,409 (+4732.76%)
Mutual labels:  artificial-intelligence, ai, ml
Awesome Ai Books
Some awesome AI related books and pdfs for learning and downloading, also apply some playground models for learning
Stars: ✭ 855 (+391.38%)
Mutual labels:  artificial-intelligence, ai, reinforcement-learning
Dopamine
Dopamine is a research framework for fast prototyping of reinforcement learning algorithms.
Stars: ✭ 9,681 (+5463.79%)
Mutual labels:  ai, ml, rl
Image classifier
CNN image classifier implemented in Keras Notebook πŸ–ΌοΈ.
Stars: ✭ 139 (-20.11%)
Mutual labels:  artificial-intelligence, ai, ml
Pygame Learning Environment
PyGame Learning Environment (PLE) -- Reinforcement Learning Environment in Python.
Stars: ✭ 828 (+375.86%)
Mutual labels:  artificial-intelligence, ai, reinforcement-learning
Awesome Ai Ml Dl
Awesome Artificial Intelligence, Machine Learning and Deep Learning as we learn it. Study notes and a curated list of awesome resources of such topics.
Stars: ✭ 831 (+377.59%)
Mutual labels:  artificial-intelligence, ai, ml
Cartpole
OpenAI's cartpole env solver.
Stars: ✭ 107 (-38.51%)
Mutual labels:  ai, reinforcement-learning, dqn
View All Similar Projects βž”

Atari

Research Playground built on top of OpenAI's Atari Gym, prepared for implementing various Reinforcement Learning algorithms.

It can emulate any of the following games:

['Asterix', 'Asteroids', 'MsPacman', 'Kaboom', 'BankHeist', 'Kangaroo', 'Skiing', 'FishingDerby', 'Krull', 'Berzerk', 'Tutankham', 'Zaxxon', 'Venture', 'Riverraid', 'Centipede', 'Adventure', 'BeamRider', 'CrazyClimber', 'TimePilot', 'Carnival', 'Tennis', 'Seaquest', 'Bowling', 'SpaceInvaders', 'Freeway', 'YarsRevenge', 'RoadRunner', 'JourneyEscape', 'WizardOfWor', 'Gopher', 'Breakout', 'StarGunner', 'Atlantis', 'DoubleDunk', 'Hero', 'BattleZone', 'Solaris', 'UpNDown', 'Frostbite', 'KungFuMaster', 'Pooyan', 'Pitfall', 'MontezumaRevenge', 'PrivateEye', 'AirRaid', 'Amidar', 'Robotank', 'DemonAttack', 'Defender', 'NameThisGame', 'Phoenix', 'Gravitar', 'ElevatorAction', 'Pong', 'VideoPinball', 'IceHockey', 'Boxing', 'Assault', 'Alien', 'Qbert', 'Enduro', 'ChopperCommand', 'Jamesbond']

Check out corresponding Medium article: Atari - Reinforcement Learning in depth πŸ€– (Part 1: DDQN)

Purpose

The ultimate goal of this project is to implement and compare various RL approaches with atari games as a common denominator.

Usage

  1. Clone the repo.
  2. Go to the project's root folder.
  3. Install required packagespip install -r requirements.txt.
  4. Launch atari. I recommend starting with help command to see all available modes python atari.py --help.

DDQN

Hyperparameters

* GAMMA = 0.99
* MEMORY_SIZE = 900000
* BATCH_SIZE = 32
* TRAINING_FREQUENCY = 4
* TARGET_NETWORK_UPDATE_FREQUENCY = 40000
* MODEL_PERSISTENCE_UPDATE_FREQUENCY = 10000
* REPLAY_START_SIZE = 50000
* EXPLORATION_MAX = 1.0
* EXPLORATION_MIN = 0.1
* EXPLORATION_TEST = 0.02
* EXPLORATION_STEPS = 850000

Model Architecture

Deep Convolutional Neural Network by DeepMind

* Conv2D (None, 32, 20, 20)
* Conv2D (None, 64, 9, 9)
* Conv2D (None, 64, 7, 7)
* Flatten (None, 3136)
* Dense (None, 512)
* Dense (None, 4)

Trainable params: 1,686,180

Performance

After 5M of steps (~40h on Tesla K80 GPU or ~90h on 2.9 GHz Intel i7 Quad-Core CPU):

SpaceInvaders

Training:

Normalized score - each reward clipped to (-1, 1)

Testing:

Human average: ~372

DDQN average: ~479 (128%)

Breakout

Training:

Normalized score - each reward clipped to (-1, 1)

Testing:

Human average: ~28

DDQN average: ~62 (221%)

Genetic Evolution

Atlantis

Training:

Normalized score - each reward clipped to (-1, 1)

Testing:

Human average: ~29,000

GE average: 31,000 (106%)

Author

Greg (Grzegorz) Surma

PORTFOLIO

GITHUB

BLOG

Support via PayPal
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].