change the get_screen method so that it does not fill the gpu memory (#415)

Co-authored-by: holly1238 <77758406+holly1238@users.noreply.github.com>

Author: andreistirb

intermediate_source/reinforcement_q_learning.py

@@ -224,6 +224,7 @@ def conv2d_size_out(size, kernel_size = 5, stride = 2):
     # Called with either one element to determine next action, or a batch
     # during optimization. Returns tensor([[left0exp,right0exp]...]).
     def forward(self, x):
+        x = x.to(device)
         x = F.relu(self.bn1(self.conv1(x)))
         x = F.relu(self.bn2(self.conv2(x)))
         x = F.relu(self.bn3(self.conv3(x)))
@@ -273,7 +274,7 @@ def get_screen():
     screen = np.ascontiguousarray(screen, dtype=np.float32) / 255
     screen = torch.from_numpy(screen)
     # Resize, and add a batch dimension (BCHW)
-    return resize(screen).unsqueeze(0).to(device)
+    return resize(screen).unsqueeze(0)
 
 
 env.reset()