elumixor · May 24, 2020 22:45
diff --git a/TRPO_critic_update.py b/TRPO_critic_update.py
 from torch.optim import Adam


 critic_optimizer = Adam(critic.parameters(), lr=0.005)


 def update_critic(advantages):
    loss = .5 * (advantages ** 2).mean()
    critic_optimizer.zero_grad()
    loss.backward()
    critic_optimizer.step()
	from torch.optim import Adam


	critic_optimizer = Adam(critic.parameters(), lr=0.005)


	def update_critic(advantages):
	loss = .5 * (advantages ** 2).mean()
	critic_optimizer.zero_grad()
	loss.backward()
	critic_optimizer.step()