hongshaorou/Tianshou
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

add render option for ddpg

Browse Source
This commit is contained in:
Dong Yan 2018-02-28 18:44:06 +08:00
parent 5ab2fa3b65
commit 528c4be93c
1 changed files with 6 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
examples/ddpg_example.py
View File

@ -5,6 +5,7 @@ import tensorflow as tf
import gym import gym
import numpy as np import numpy as np
import time import time
import argparse
# our lib imports here! It's ok to append path in examples # our lib imports here! It's ok to append path in examples
import sys import sys
@ -18,6 +19,9 @@ import tianshou.core.opt as opt
if __name__ == '__main__': if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument("--render", action="store_true", default=False)
args = parser.parse_args()
env = gym.make('Pendulum-v0') env = gym.make('Pendulum-v0')
observation_dim = env.observation_space.shape observation_dim = env.observation_space.shape
action_dim = env.action_space.shape action_dim = env.action_space.shape
@ -60,7 +64,8 @@ if __name__ == '__main__':
actor_train_op = actor_optimizer.apply_gradients(dpg_grads) actor_train_op = actor_optimizer.apply_gradients(dpg_grads)
### 3. define data collection ### 3. define data collection
data_collector = Batch(env, actor, [advantage_estimation.ddpg_return(actor, critic)], [actor, critic]) data_collector = Batch(env, actor, [advantage_estimation.ddpg_return(actor, critic)], [actor, critic],
render = args.render)
### 4. start training ### 4. start training
config = tf.ConfigProto() config = tf.ConfigProto()