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

		Paper Title	Repository
MuZero	143972.03	Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model
R2D2	108197.0	Recurrent Experience Replay in Distributed Reinforcement Learning	-
GDI-H3	97155	Generalized Data Distribution Iteration	-
Agent57	67212.67	Agent57: Outperforming the Atari Human Benchmark
GDI-I3	63876	Generalized Data Distribution Iteration	-
MuZero (Res2 Adam)	33292.22	Online and Offline Reinforcement Learning by Planning with a Learned Model
IQN	29091	Implicit Quantile Networks for Distributional Reinforcement Learning
Ape-X	24559.4	Distributed Prioritized Experience Replay
DreamerV2	23625	Mastering Atari with Discrete World Models
QR-DQN-1	22012	Distributional Reinforcement Learning with Quantile Regression
IMPALA (deep)	19148.47	IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures
DNA	16293	DNA: Proximal Policy Optimization with a Dual Network Architecture
A3C LSTM hs	14497.9	Asynchronous Methods for Deep Reinforcement Learning
ASL DDQN	14372.8	Train a Real-world Local Path Planner in One Hour via Partially Decoupled Reinforcement Learning and Vectorized Diversity
Prior+Duel noop	11477.0	Dueling Network Architectures for Deep Reinforcement Learning
NoisyNet-Dueling	11231	Noisy Networks for Exploration
Prior+Duel hs	10950.6	Dueling Network Architectures for Deep Reinforcement Learning
Prior+Duel hs	10950.6	Deep Reinforcement Learning with Double Q-learning
DDQN+Pop-Art noop	9011.6	Learning values across many orders of magnitude	-
Reactor 500M	8323.3	The Reactor: A fast and sample-efficient Actor-Critic agent for Reinforcement Learning	-

0 of 45 row(s) selected.