From 5b1bb15b406111be831839caf9abe0a48b8d01b1 Mon Sep 17 00:00:00 2001
From: Ervin T <ervin@unity3d.com>
Date: Tue, 17 Nov 2020 16:01:44 -0800
Subject: [PATCH] [bug-fix] Separate critic only for PPO (#4661)

---
 ml-agents/mlagents/trainers/ppo/trainer.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/ml-agents/mlagents/trainers/ppo/trainer.py b/ml-agents/mlagents/trainers/ppo/trainer.py
index 401dc4b1fb..3e2b913207 100644
--- a/ml-agents/mlagents/trainers/ppo/trainer.py
+++ b/ml-agents/mlagents/trainers/ppo/trainer.py
@@ -254,7 +254,7 @@ def create_torch_policy(
             behavior_spec,
             self.trainer_settings,
             condition_sigma_on_obs=False,  # Faster training for PPO
-            separate_critic=behavior_spec.action_spec.is_continuous(),
+            separate_critic=True,  # Match network architecture with TF
         )
         return policy