search

kengz / SLM-Lab

Modular Deep Reinforcement Learning framework in PyTorch. Companion library of the book "Foundations of Deep Reinforcement Learning".
https://slm-lab.gitbook.io/slm-lab/
MIT License
1.25k stars 264 forks source link

Full Roboschool benchmark #405

Closed kengz closed 5 years ago

kengz commented 5 years ago

Feature / Fix

Env. \ Alg. A2C (GAE) A2C (n-step) PPO SAC
RoboschoolAnt
graph
 1029.51
graph
 1148.76
graph
 1931.35
graph
 2903.32
graph
RoboschoolAtlasForwardWalk
graph
 68.15
graph
 73.46
graph
 148.81
graph
 942.39
graph
RoboschoolHalfCheetah
graph
 895.24
graph
 409.59
graph
 1838.69
graph
 2350.94
graph
RoboschoolHopper
graph
 286.67
graph
 -187.91
graph
 2079.22
graph
 1826.50
graph
RoboschoolInvertedDoublePendulum
graph
 1769.74
graph
 486.76
graph
 7967.03
graph
 6664.53
graph
RoboschoolInvertedPendulum
graph
 1000.0
graph
 997.54
graph
 930.29
graph
 955.16
graph
RoboschoolInvertedPendulumSwingup
graph
 1000.0
graph
 997.54
graph
 930.29
graph
 958.29
graph
RoboschoolReacher
graph
 14.57
graph
 -6.18
graph
 19.18
graph
 19.73
graph
RoboschoolWalker2d
graph
 413.26
graph
 141.83
graph
 1368.25
graph
 1982.05
graph

NOTE There's a bug in RoboschoolInvertedPendulumSwingup that causes it to use the basic Pendulum. To be filed. See source code here
sac ant sac halfcheetah sac hopper sac humanoid
Ant HalfCheetah Hopper Humanoid
sac doublependulum sac pendulum sac reacher sac walker
Inv.DoublePendulum InvertedPendulum Reacher Walker