Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time


Testbed for deep reinforcement learning algorithms.

DQN playing Breakout   DQN playing Doom Health Gathering   DQN trying to play Doom Deathmatch


Reinforcement learning is a fundamental problem in artificial intelligence. In this setting, an agent interacts with an environment in order to maximize a reward. For example, we show our bot pixel screens of a game and want it to choose actions that result in a high score.

Mindpark is an environment for prototyping, testing, and comparing algorithms that do reinforcement learning. The library makes it easy to reuse part of behavior between algorithms, and monitor all kinds of metrics about your algorithms. It integrates well with TensorFlow, Theano, and other deep learning libraries, and with OpenAI's gym environments.

These are the algorithms that I implemented so far (feel free to contribute to this list):

Algorithm Publication Status
Deep Q-Network (DQN) Mnih et al. 2015 (PDF) Working consistently.
Double Deep Q-Network (DDQN) Hasselt, Guez, Silver. 2015 (PDF) Working consistently.
Asynchronous Advantage Actor-Critic (A3C) Mnih et al. 2016 (PDF) Partly working.
Reinforce Williams 1992 (PDF) Currently being tested.


To get started, clone the repository and install dependencies:

git clone && cd mindpark
sudo -H pip3 install .

An experiment compares between algorithms, hyper parameters, and environments. To start an experiment, run (-O turns on Python's optimizations):

python3 -O -m mindpark run definition/breakout.yaml

Videos and metrics are stored in a result directory, which is ~/experiment/mindpark/<timestamp>-breakout/ by default. You can plot statistics during or after the simulation by fuzzy matching an the folder name:

python3 -m mindpark stats breakout


Let's take a look at what the previous command creates.

Experiments consist of interleaved phases of training and evaluation. For example, an algorithm might use a lower exploration rate in favor of exploitation while being evaluated. Therefore, we display the metrics in two rows:

DQN statistics on Breakout

This illustrates the metrics after a few episodes of training, as you can see on the horizontal axes. This small example is good for explanation. But if you want to take a look, here are the metrics of a longer experiment.

Metric Description
score During the first 80 episodes of training (the time when I ran mindpark stats), the algorithm manages to get a score of 9, but usually get scores around 3 and 4. Below is the score during evaluation. It's lower because the algorithm hasn't learned much yet and performs worse than the random exploration done during training.
dqn/cost The training cost of the neural network. It starts at episode 10 which is when the training starts, before that, DQN builds up its replay memory. We don't train the neural network during evaluation, so that plot is empty.
epsilion_greedy/values That's the Q-values that the dqn behavior sends to epsilon_greedy to act greedily on. You can see that they evolve over time: Action 4 seems to be quite good. But that's only for a short run, so we shouldn't conclude too much.
epsilion_greedy/random A histogram whether the current action was chosen randomly or greedy wrt the predicted Q-values. During training, epsilon is annealed, so you see a shift in the distribution. During testing, epsilon is always 0.05, so not many random actions there.

The metric names are prefixed by the classes they come from. That's because algorithms are composed of reusable partial behaviors. See the Algorithms section for details.


Definitions are YAML files that contain all you need to run or reproduce an experiment:

epochs: 100
test_steps: 4e4
repeats: 5
  - Pong-v0
  - Breakout-v0
    name: LSTM-A3C (3 layers)
    type: A3C
    train_steps: 8e5
      network: lstm_three_layers
      initial_learning_rate: 2e-4
    name: DQN (Mnih et al. 2015)
    type: DQN
    train_steps: 2e5
    name: Random
    type: Random
    train_steps: 0

Each algorithm will be trained on each environment for the specified number of repeats. A simulation is divided into epochs that consist of a training and an evaluation phase.


To implement your own algorithm, subclass mindpark.core.Algorithm. Please refer to the existing algorithms for details, and ask if you have questions. Algorithms are composed of partial behaviors that can do preprocessing, exploration, learning, and more. To create a reusable chunk of behavior, subclass mindpark.core.Partial.

There are quite a few existing behaviors that you can import from mindpark.step and reuse in your algorithms. For more details, please look at the according Python files or open an issue. Current behaviors include: ActionMax, ActionSample, ClampReward, Delta, EpsilonGreedy, Experience, Filter, Grayscale, History, Identity, Maximum, Normalize, Random, RandomStart, Resize, Score, Skip, Subsample.


Mindpark is a Python 3 package, and there are no plans to support Python 2. Please install gym_doom manually.

sudo apt-get install -y gtk2.0-dev libsdl2-dev libfluidsynth-dev libopenal-dev libboost-all-dev
sudo -H python3 -c "import gym_pull; gym_pull.pull('')"

TensorFlow is only needed for the existing algorithms. You are free to use your libraries of choice to implement your own algorithms.


Your pull request is very welcome. I will set up a contributors file in that case, and you can choose if and how you want to be listed.

Please follow the existing code style, and run unit tests and the integration test after changes:

python3 test
python3 -m mindpark run definition/test.yaml -x


Feel free to reach out at or open an issue here on Github if you have any questions.