tonydeep / naf-tensorflow

"Continuous Deep Q-Learning with Model-based Acceleration" in TensorFlow

Normalized Advantage Functions (NAF) in TensorFlow

TensorFlow implementation of Continuous Deep q-Learning with Model-based Acceleration.

Requirements

• Python 2.7
• gym
• TensorFlow 0.9+

Usage

First, install prerequisites with:

$ pip install tqdm gym[all]

To train a model for an environment with a continuous action space:

$ python main.py --env=Pendulum-v0 --is_train=True
$ python main.py --env=Pendulum-v0 --is_train=True --display=True

To test and record the screens with gym:

$ python main.py --env=Pendulum-v0 --is_train=False
$ python main.py --env=Pendulum-v0 --is_train=False --display=True

Results

Training details of Pendulum-v0 with different hyperparameters.

$ python main.py --env=Pendulum-v0 # dark green
$ python main.py --env=Pendulum-v0 --action_fn=tanh # light green
$ python main.py --env=Pendulum-v0 --use_batch_norm=True # yellow
$ python main.py --env=Pendulum-v0 --use_seperate_networks=True # green

References

• rllab
• keras implementation

Author

Taehoon Kim / @carpedm20