diff --git a/__pycache__/model.cpython-38.pyc b/__pycache__/model.cpython-38.pyc
new file mode 100644
index 0000000..96b075c
Binary files /dev/null and b/__pycache__/model.cpython-38.pyc differ
diff --git a/__pycache__/replay_memory.cpython-38.pyc b/__pycache__/replay_memory.cpython-38.pyc
new file mode 100644
index 0000000..4b28683
Binary files /dev/null and b/__pycache__/replay_memory.cpython-38.pyc differ
diff --git a/__pycache__/sac.cpython-38.pyc b/__pycache__/sac.cpython-38.pyc
new file mode 100644
index 0000000..659207e
Binary files /dev/null and b/__pycache__/sac.cpython-38.pyc differ
diff --git a/__pycache__/utils.cpython-38.pyc b/__pycache__/utils.cpython-38.pyc
new file mode 100644
index 0000000..1bf2124
Binary files /dev/null and b/__pycache__/utils.cpython-38.pyc differ
diff --git a/main.py b/main.py
index a3047c5..562e3d9 100644
--- a/main.py
+++ b/main.py
@@ -9,7 +9,7 @@ from tensorboardX import SummaryWriter
 from replay_memory import ReplayMemory
 
 parser = argparse.ArgumentParser(description='PyTorch Soft Actor-Critic Args')
-parser.add_argument('--env-name', default="HalfCheetah-v2",
+parser.add_argument('--env-name', default="HalfCheetahBulletEnv-v0",
                     help='Mujoco Gym environment (default: HalfCheetah-v2)')
 parser.add_argument('--policy', default="Gaussian",
                     help='Policy Type: Gaussian | Deterministic (default: Gaussian)')
diff --git a/runs/2020-06-06_00-16-03_SAC_HalfCheetahBulletEnv-v0_Gaussian_/events.out.tfevents.1591382763.pranjal-Lenovo-ideapad-730S-13IWL b/runs/2020-06-06_00-16-03_SAC_HalfCheetahBulletEnv-v0_Gaussian_/events.out.tfevents.1591382763.pranjal-Lenovo-ideapad-730S-13IWL
new file mode 100644
index 0000000..e69de29
diff --git a/runs/2020-06-06_00-16-31_SAC_HalfCheetahBulletEnv-v0_Gaussian_/events.out.tfevents.1591382791.pranjal-Lenovo-ideapad-730S-13IWL b/runs/2020-06-06_00-16-31_SAC_HalfCheetahBulletEnv-v0_Gaussian_/events.out.tfevents.1591382791.pranjal-Lenovo-ideapad-730S-13IWL
new file mode 100644
index 0000000..e69de29
diff --git a/runs/2020-06-06_00-17-41_SAC_HalfCheetahBulletEnv-v0_Gaussian_/events.out.tfevents.1591382861.pranjal-Lenovo-ideapad-730S-13IWL b/runs/2020-06-06_00-17-41_SAC_HalfCheetahBulletEnv-v0_Gaussian_/events.out.tfevents.1591382861.pranjal-Lenovo-ideapad-730S-13IWL
new file mode 100644
index 0000000..c331439
Binary files /dev/null and b/runs/2020-06-06_00-17-41_SAC_HalfCheetahBulletEnv-v0_Gaussian_/events.out.tfevents.1591382861.pranjal-Lenovo-ideapad-730S-13IWL differ
diff --git a/sac.py b/sac.py
index 0ff0b45..aef23f0 100644
--- a/sac.py
+++ b/sac.py
@@ -67,6 +67,14 @@ class SAC(object):
         qf1, qf2 = self.critic(state_batch, action_batch)  # Two Q-functions to mitigate positive bias in the policy improvement step
         qf1_loss = F.mse_loss(qf1, next_q_value)  # JQ = 𝔼(st,at)~D[0.5(Q1(st,at) - r(st,at) - γ(𝔼st+1~p[V(st+1)]))^2]
         qf2_loss = F.mse_loss(qf2, next_q_value)  # JQ = 𝔼(st,at)~D[0.5(Q1(st,at) - r(st,at) - γ(𝔼st+1~p[V(st+1)]))^2]
+        qf_loss = qf1_loss + qf2_loss
+
+        self.critic_optim.zero_grad()
+        qf_loss.backward()
+        self.critic_optim.step()
+
+        for c_param in self.critic.parameters():
+            c_param.requires_grad = False
 
         pi, log_pi, _ = self.policy.sample(state_batch)
 
@@ -75,18 +83,13 @@ class SAC(object):
 
         policy_loss = ((self.alpha * log_pi) - min_qf_pi).mean() # Jπ = 𝔼st∼D,εt∼N[α * logπ(f(εt;st)|st) − Q(st,f(εt;st))]
 
-        self.critic_optim.zero_grad()
-        qf1_loss.backward()
-        self.critic_optim.step()
-
-        self.critic_optim.zero_grad()
-        qf2_loss.backward()
-        self.critic_optim.step()
-
         self.policy_optim.zero_grad()
         policy_loss.backward()
         self.policy_optim.step()
 
+        for c_param in self.critic.parameters():
+            c_param.requires_grad = True
+
         if self.automatic_entropy_tuning:
             alpha_loss = -(self.log_alpha * (log_pi + self.target_entropy).detach()).mean()