diff --git a/sac.py b/sac.py
index aef23f0..1aa523f 100644
--- a/sac.py
+++ b/sac.py
@@ -73,9 +73,6 @@ class SAC(object):
         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)
 
         qf1_pi, qf2_pi = self.critic(state_batch, pi)
@@ -87,9 +84,6 @@ class SAC(object):
         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()