gryf/coach
1
0
Fork 0
mirror of https://github.com/gryf/coach.git synced 2025-12-18 03:30:19 +01:00
Code Issues Projects Releases Wiki Activity

OPE: Weighted Importance Sampling (#299)

Browse Source
This commit is contained in:
Gal Leibovich
2019-05-02 19:25:42 +03:00
committed by GitHub
parent 74db141d5e
commit 582921ffe3
8 changed files with 222 additions and 51 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
rl_coach/agents/agent.py
View File

@@ -522,6 +522,7 @@ class Agent(AgentInterface):
self.agent_logger.create_signal_value('Inverse Propensity Score', np.nan, overwrite=False)
self.agent_logger.create_signal_value('Direct Method Reward', np.nan, overwrite=False)
self.agent_logger.create_signal_value('Doubly Robust', np.nan, overwrite=False)
self.agent_logger.create_signal_value('Weighted Importance Sampling', np.nan, overwrite=False)
self.agent_logger.create_signal_value('Sequential Doubly Robust', np.nan, overwrite=False)
for signal in self.episode_signals: