diff --git a/rllib/agents/ppo/appo.py b/rllib/agents/ppo/appo.py index f963cc7eb..61510eb04 100644 --- a/rllib/agents/ppo/appo.py +++ b/rllib/agents/ppo/appo.py @@ -31,7 +31,7 @@ DEFAULT_CONFIG = impala.ImpalaTrainer.merge_trainer_configs( { # Whether to use V-trace weighted advantages. If false, PPO GAE # advantages will be used instead. - "vtrace": False, + "vtrace": True, # == These two options only apply if vtrace: False == # Should use a critic as a baseline (otherwise don't use value