Modified demo config and run demo script (#83)

* Add removed q_dot scaling factor

* Refactored cubic function

* Separate config and add run file.

* Resolve flake8 issue

* Resolve flake8 issue

* Remove whitespace before ( issue

scripts/algorithms/common/abstract/agent.py

@@ -80,7 +80,7 @@ class AbstractAgent(object):
         path = os.path.join("./save/" + save_name + "_ep_" + str(n_episode) + ".pt")
         torch.save(params, path)
 
-        print ("[INFO] Saved the model and optimizer to", path)
+        print("[INFO] Saved the model and optimizer to", path)
 
     @abstractmethod
     def write_log(self, *args):
@@ -109,7 +109,7 @@ class AbstractAgent(object):
                 score += reward
                 step += 1
 
-            print (
+            print(
                 "[INFO] episode %d\tstep: %d\ttotal score: %d"
                 % (i_episode, step, score)
             )

scripts/algorithms/fd/sac_agent.py

@@ -201,7 +201,7 @@ class Agent(SACAgent):
     def pretrain(self):
         """Pretraining steps."""
         pretrain_loss = list()
-        print ("[INFO] Pre-Train %d steps." % self.hyper_params["PRETRAIN_STEP"])
+        print("[INFO] Pre-Train %d steps." % self.hyper_params["PRETRAIN_STEP"])
         for i_step in range(1, self.hyper_params["PRETRAIN_STEP"] + 1):
             loss = self.update_model()
             pretrain_loss.append(loss)  # for logging

scripts/algorithms/fd/td3_agent.py

@@ -180,7 +180,7 @@ class Agent(TD3Agent):
     def pretrain(self):
         """Pretraining steps."""
         pretrain_loss = list()
-        print ("[INFO] Pre-Train %d steps." % self.hyper_params["PRETRAIN_STEP"])
+        print("[INFO] Pre-Train %d steps." % self.hyper_params["PRETRAIN_STEP"])
         for i_step in range(1, self.hyper_params["PRETRAIN_STEP"] + 1):
             loss = self.update_model()
             pretrain_loss.append(loss)  # for logging

scripts/algorithms/sac/agent.py

@@ -225,7 +225,7 @@ class Agent(AbstractAgent):
     def load_params(self, path):
         """Load model and optimizer parameters."""
         if not os.path.exists(path):
-            print ("[ERROR] the input path does not exist. ->", path)
+            print("[ERROR] the input path does not exist. ->", path)
             return
 
         params = torch.load(path)
@@ -242,7 +242,7 @@ class Agent(AbstractAgent):
         if self.hyper_params["AUTO_ENTROPY_TUNING"]:
             self.alpha_optimizer.load_state_dict(params["alpha_optim"])
 
-        print ("[INFO] loaded the model and optimizer from", path)
+        print("[INFO] loaded the model and optimizer from", path)
 
     def save_params(self, n_episode):
         """Save model and optimizer parameters."""
@@ -267,7 +267,7 @@ class Agent(AbstractAgent):
         """Write log about loss and score"""
         total_loss = loss.sum()
 
-        print (
+        print(
             "[INFO] episode %d, episode_step %d, total step %d, total score: %d\n"
             "total loss: %.3f actor_loss: %.3f qf_1_loss: %.3f qf_2_loss: %.3f "
             "vf_loss: %.3f alpha_loss: %.3f\n"

scripts/algorithms/td3/agent.py

@@ -177,7 +177,7 @@ class Agent(AbstractAgent):
     def load_params(self, path):
         """Load model and optimizer parameters."""
         if not os.path.exists(path):
-            print ("[ERROR] the input path does not exist. ->", path)
+            print("[ERROR] the input path does not exist. ->", path)
             return
 
         params = torch.load(path)
@@ -189,7 +189,7 @@ class Agent(AbstractAgent):
         self.critic2_target.load_state_dict(params["critic2_target_state_dict"])
         self.actor_optim.load_state_dict(params["actor_optim_state_dict"])
         self.critic_optim.load_state_dict(params["critic_optim_state_dict"])
-        print ("[INFO] loaded the model and optimizer from", path)
+        print("[INFO] loaded the model and optimizer from", path)
 
     def save_params(self, n_episode):
         """Save model and optimizer parameters."""
@@ -210,7 +210,7 @@ class Agent(AbstractAgent):
         """Write log about loss and score"""
         total_loss = loss.sum()
 
-        print (
+        print(
             "[INFO] total_steps: %d episode: %d total score: %d, total loss: %f\n"
             "actor_loss: %.3f critic1_loss: %.3f critic2_loss: %.3f\n"
             % (self.total_steps, i, score, total_loss, loss[0], loss[1], loss[2])