wassname/DeepRL
1
0
Fork 0
mirror of https://github.com/wassname/DeepRL.git synced 2026-09-11 11:53:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

DDPG Pendulum

Browse Source
This commit is contained in:
Shangtong Zhang
2017-08-01 10:52:14 -06:00
parent 1be3b44999
commit 5116733f22
5 changed files with 69 additions and 97 deletions
Download Patch File Download Diff File Expand all files Collapse all files
network/continuous_action_network.py
+5 -3
View File
@@ -47,13 +47,15 @@ class DDPGActorNet(nn.Module, BasicNet):
def __init__(self,
state_dim,
action_dim,
output_gate,
action_gate,
action_scale,
gpu=False):
super(DDPGActorNet, self).__init__()
self.layer1 = nn.Linear(state_dim, 400)
self.layer2 = nn.Linear(400, 300)
self.layer3 = nn.Linear(300, action_dim)
self.output_gate = output_gate
self.action_gate = action_gate
self.action_scale = action_scale
BasicNet.__init__(self, None, False, False)
self.init_weights()
@@ -76,7 +78,7 @@ class DDPGActorNet(nn.Module, BasicNet):
x = F.relu(self.layer1(x))
x = F.relu(self.layer2(x))
x = self.layer3(x)
# x = self.output_gate(self.layer3(x))
x = self.action_scale * self.action_gate(x)
return x
def predict(self, x, to_numpy=True):