File size: 2,836 Bytes
a89d35f | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 | .. _dqn:
.. automodule:: stable_baselines3.dqn
DQN
===
`Deep Q Network (DQN) <https://arxiv.org/abs/1312.5602>`_ builds on `Fitted Q-Iteration (FQI) <http://ml.informatik.uni-freiburg.de/former/_media/publications/rieecml05.pdf>`_
and make use of different tricks to stabilize the learning with neural networks: it uses a replay buffer, a target network and gradient clipping.
.. rubric:: Available Policies
.. autosummary::
:nosignatures:
MlpPolicy
CnnPolicy
Notes
-----
- Original paper: https://arxiv.org/abs/1312.5602
- Further reference: https://www.nature.com/articles/nature14236
.. note::
This implementation provides only vanilla Deep Q-Learning and has no extensions such as Double-DQN, Dueling-DQN and Prioritized Experience Replay.
Can I use?
----------
- Recurrent policies: ❌
- Multi processing: ❌
- Gym spaces:
============= ====== ===========
Space Action Observation
============= ====== ===========
Discrete ✔ ✔
Box ❌ ✔
MultiDiscrete ❌ ✔
MultiBinary ❌ ✔
============= ====== ===========
Example
-------
.. code-block:: python
import gym
import numpy as np
from stable_baselines3 import DQN
from stable_baselines3.dqn import MlpPolicy
env = gym.make('CartPole-v0')
model = DQN(MlpPolicy, env, verbose=1)
model.learn(total_timesteps=10000, log_interval=4)
model.save("dqn_pendulum")
del model # remove to demonstrate saving and loading
model = DQN.load("dqn_pendulum")
obs = env.reset()
while True:
action, _states = model.predict(obs, deterministic=True)
obs, reward, done, info = env.step(action)
env.render()
if done:
obs = env.reset()
Results
-------
Atari Games
^^^^^^^^^^^
The complete learning curves are available in the `associated PR #110 <https://github.com/DLR-RM/stable-baselines3/pull/110>`_.
How to replicate the results?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Clone the `rl-zoo repo <https://github.com/DLR-RM/rl-baselines3-zoo>`_:
.. code-block:: bash
git clone https://github.com/DLR-RM/rl-baselines3-zoo
cd rl-baselines3-zoo/
Run the benchmark (replace ``$ENV_ID`` by the env id, for instance ``BreakoutNoFrameskip-v4``):
.. code-block:: bash
python train.py --algo dqn --env $ENV_ID --eval-episodes 10 --eval-freq 10000
Plot the results:
.. code-block:: bash
python scripts/all_plots.py -a dqn -e Pong Breakout -f logs/ -o logs/dqn_results
python scripts/plot_from_file.py -i logs/dqn_results.pkl -latex -l DQN
Parameters
----------
.. autoclass:: DQN
:members:
:inherited-members:
.. _dqn_policies:
DQN Policies
-------------
.. autoclass:: MlpPolicy
:members:
:inherited-members:
.. autoclass:: stable_baselines3.dqn.policies.DQNPolicy
:members:
:noindex:
.. autoclass:: CnnPolicy
:members:
|