Atari Games On Atari 2600 Assault

评估指标

Score

评测结果

各个模型在此基准测试上的表现结果

模型名称	Score	Paper Title	Repository
Reactor 500M	8323.3	The Reactor: A fast and sample-efficient Actor-Critic agent for Reinforcement Learning	-
Bootstrapped DQN	8047.1	Deep Exploration via Bootstrapped DQN	-
CGP	890.4	Evolving simple programs for playing Atari games	-
Ape-X	24559.4	Distributed Prioritized Experience Replay	-
A3C LSTM hs	14497.9	Asynchronous Methods for Deep Reinforcement Learning	-
DQN noop	4280.4	Deep Reinforcement Learning with Double Q-learning	-
A3C FF hs	5474.9	Asynchronous Methods for Deep Reinforcement Learning	-
Agent57	67212.67	Agent57: Outperforming the Atari Human Benchmark	-
Prior+Duel noop	11477.0	Dueling Network Architectures for Deep Reinforcement Learning	-
Prior+Duel hs	10950.6	Dueling Network Architectures for Deep Reinforcement Learning	-
Gorila	1195.8	Massively Parallel Methods for Deep Reinforcement Learning	-
SAC	350	Soft Actor-Critic for Discrete Action Settings	-
Advantage Learning	3661.51	Increasing the Action Gap: New Operators for Reinforcement Learning	-
DDQN (tuned) hs	6060.8	Deep Reinforcement Learning with Double Q-learning	-
Prior+Duel hs	10950.6	Deep Reinforcement Learning with Double Q-learning	-
Persistent AL	3304.33	Increasing the Action Gap: New Operators for Reinforcement Learning	-
C51 noop	7203.0	A Distributional Perspective on Reinforcement Learning	-
DDQN (tuned) noop	5393.2	Dueling Network Architectures for Deep Reinforcement Learning	-
Duel noop	4621.0	Dueling Network Architectures for Deep Reinforcement Learning	-
Duel hs	3994.8	Dueling Network Architectures for Deep Reinforcement Learning	-

0 of 45 row(s) selected.