#42 run baselines in ci

torch_training/bla.py

@@ -157,18 +157,18 @@ def main(argv):
                 action_prob[action] += 1
                 action_dict.update({a: action})
             # Environment step
-        #
-        #     next_obs, all_rewards, done, _ = env.step(action_dict)
-        #     # print(all_rewards,action)
-        #     obs_original = next_obs.copy()
-        #     for a in range(env.get_num_agents()):
-        #         data, distance, agent_data = split_tree(tree=np.array(next_obs[a]),
-        #                                                 current_depth=0)
-        #         data = norm_obs_clip(data)
-        #         distance = norm_obs_clip(distance)
-        #         agent_data = np.clip(agent_data, -1, 1)
-        #         next_obs[a] = np.concatenate((np.concatenate((data, distance)), agent_data))
-        #     time_obs.append(next_obs)
+
+            next_obs, all_rewards, done, _ = env.step(action_dict)
+            # print(all_rewards,action)
+            obs_original = next_obs.copy()
+            for a in range(env.get_num_agents()):
+                data, distance, agent_data = split_tree(tree=np.array(next_obs[a]),
+                                                        current_depth=0)
+                data = norm_obs_clip(data)
+                distance = norm_obs_clip(distance)
+                agent_data = np.clip(agent_data, -1, 1)
+                next_obs[a] = np.concatenate((np.concatenate((data, distance)), agent_data))
+            time_obs.append(next_obs)
         #
         #     # Update replay buffer and train agent
         #     for a in range(env.get_num_agents()):