test env update

rl_agent/policy.py

@@ -10,7 +10,7 @@ class Policy(nn.Module):
 
         self.layer1 = nn.Linear(input_channels, 2 * input_channels)
         self.tanh1   = nn.Tanh()
-        self.layer2 = nn.linear(2 * input_channels, 1)
+        self.layer2 = nn.Linear(2 * input_channels, 1)
         self.tanh2 = nn.Tanh()
 
     def forward(self, state):

rl_agent/test_env.py

@@ -4,6 +4,7 @@ from utils import myOptimizer
 
 import pandas as pd
 import numpy as np
+import torch
 
 if __name__ == "__main__":
 
@@ -22,9 +23,9 @@ if __name__ == "__main__":
     second_momentum = 0.0
     transaction_cost = 0.0001
     adaptation_rate = 0.01
-    state_size = 9
+    state_size = 8
 
-    agent = Policy(input_channels=state_size)
+    agent = Policy(input_channels=state_size).float()
     optimizer = myOptimizer(learning_rate, first_momentum, second_momentum, adaptation_rate, transaction_cost)
 
     
@@ -36,10 +37,12 @@ if __name__ == "__main__":
 
     env = Environment(train, history=history)
     observation = env.reset()
-    for _ in range(9, 12):
 
+    for _ in range(9, 12):
+        print(type(observation))
+        observation = torch.as_tensor(observation)
         action = agent(observation)
-        observation, reward, _ = env.step(action)
+        observation, reward, _ = env.step(action.data.numpy())
 
     print(env.profits)