From 9c5719fc42549f583c0a99ef0e3e88ee0406f670 Mon Sep 17 00:00:00 2001 From: =?utf8?q?J=C3=A9r=C3=B4me=20Benoit?= Date: Mon, 22 Dec 2025 20:02:33 +0100 Subject: [PATCH] fix(ReforceXY): reset last reward shaping on neutral self loop MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit When the environment is reset, ensure that the last reward shaping value Signed-off-by: Jérôme Benoit --- ReforceXY/user_data/freqaimodels/ReforceXY.py | 1 + 1 file changed, 1 insertion(+) diff --git a/ReforceXY/user_data/freqaimodels/ReforceXY.py b/ReforceXY/user_data/freqaimodels/ReforceXY.py index dbde6c6..db83ef1 100644 --- a/ReforceXY/user_data/freqaimodels/ReforceXY.py +++ b/ReforceXY/user_data/freqaimodels/ReforceXY.py @@ -2351,6 +2351,7 @@ class MyRLEnv(Base5ActionRLEnv): self._last_next_potential = float(self._last_potential) self._last_entry_additive = 0.0 self._last_exit_additive = 0.0 + self._last_reward_shaping = 0.0 return base_reward def _set_observation_space(self) -> None: -- 2.53.0