Atari Games On Atari 2600 Kangaroo

평가 지표

Score

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름	Score	Paper Title	Repository
NoisyNet-Dueling	15227	Noisy Networks for Exploration
QR-DQN-1	15356	Distributional Reinforcement Learning with Quantile Regression
GDI-I3	14500	GDI: Rethinking What Makes Reinforcement Learning Different From Supervised Learning	-
Ape-X	1416	Distributed Prioritized Experience Replay
MuZero	16763.60	Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model
DQN hs	4496.0	Deep Reinforcement Learning with Double Q-learning
Nature DQN	6740.0	Human level control through deep reinforcement learning
MuZero (Res2 Adam)	13838	Online and Offline Reinforcement Learning by Planning with a Learned Model
Advantage Learning	10809.16	Increasing the Action Gap: New Operators for Reinforcement Learning
IQN	15487	Implicit Quantile Networks for Distributional Reinforcement Learning
Prior noop	16200.0	Prioritized Experience Replay
Prior+Duel noop	1792.0	Dueling Network Architectures for Deep Reinforcement Learning
Agent57	24034.16	Agent57: Outperforming the Atari Human Benchmark
ASL DDQN	13027	Train a Real-world Local Path Planner in One Hour via Partially Decoupled Reinforcement Learning and Vectorized Diversity	-
SARSA	8.8	-	-
A3C FF hs	94.0	Asynchronous Methods for Deep Reinforcement Learning
ES FF (1 hour) noop	11200.0	Evolution Strategies as a Scalable Alternative to Reinforcement Learning
DDQN (tuned) hs	11204.0	Deep Reinforcement Learning with Double Q-learning
DDQN (tuned) noop	12992.0	Dueling Network Architectures for Deep Reinforcement Learning
A3C FF (1 day) hs	106.0	Asynchronous Methods for Deep Reinforcement Learning

0 of 47 row(s) selected.