diff --git a/nasim/agents/bruteforce_agent.py b/nasim/agents/bruteforce_agent.py
index caa5a377..76c8f2a3 100644
--- a/nasim/agents/bruteforce_agent.py
+++ b/nasim/agents/bruteforce_agent.py
@@ -72,7 +72,7 @@ def run_bruteforce_agent(env, step_limit=1e6, verbose=True):
                 act = next(act_iter)
                 cycle_complete = True
 
-        _, rew, done, env_step_limit_reached, _ = env.step(act)
+        _, rew, done, env_step_limit_reached, _ = env.step(int(act))
         total_reward += rew
 
         if cycle_complete and verbose:
diff --git a/nasim/agents/ql_replay_agent.py b/nasim/agents/ql_replay_agent.py
index 13414bd3..29ccba28 100644
--- a/nasim/agents/ql_replay_agent.py
+++ b/nasim/agents/ql_replay_agent.py
@@ -247,10 +247,9 @@ def run_train_episode(self, step_limit):
         episode_return = 0
 
         while not done and not env_step_limit_reached and steps < step_limit:
-            a = self.get_egreedy_action(o, self.get_epsilon())
-
+            a = self.get_egreedy_action(o[0], self.get_epsilon())
             next_o, r, done, env_step_limit_reached, _ = self.env.step(a)
-            self.replay.store(o, a, next_o, r, done)
+            self.replay.store(o[0], a, next_o, r, done)
             self.steps_done += 1
             mean_td_error, mean_v = self.optimize()
             self.logger.add_scalar(
@@ -287,7 +286,7 @@ def run_eval_episode(self,
             input("Initial state. Press enter to continue..")
 
         while not done and not env_step_limit_reached:
-            a = self.get_egreedy_action(o, eval_epsilon)
+            a = self.get_egreedy_action(o[0], eval_epsilon)
             next_o, r, done, env_step_limit_reached, _ = env.step(a)
             o = next_o
             episode_return += r
@@ -296,7 +295,7 @@ def run_eval_episode(self,
                 print("\n" + line_break)
                 print(f"Step {steps}")
                 print(line_break)
-                print(f"Action Performed = {env.action_space.get_action(a)}")
+                print(f"Action Performed = {env.action_space.get_action(int(a))}")
                 env.render(render_mode)
                 print(f"Reward = {r}")
                 print(f"Done = {done}")
diff --git a/nasim/agents/random_agent.py b/nasim/agents/random_agent.py
index e7fe7d1c..bc858946 100644
--- a/nasim/agents/random_agent.py
+++ b/nasim/agents/random_agent.py
@@ -35,7 +35,7 @@ def run_random_agent(env, step_limit=1e6, verbose=True):
 
     while not done and not env_step_limit_reached and t < step_limit:
         a = env.action_space.sample()
-        _, r, done, env_step_limit_reached, _ = env.step(a)
+        _, r, done, env_step_limit_reached, _ = env.step(int(a))
         total_reward += r
         if (t+1) % 100 == 0 and verbose:
             print(f"{t}: {total_reward}")