issues
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
Upgrade Unity environments to use gym API
#407
Closed
kengz
closed
5 years ago
kengz
commented
5 years ago
Unity gym envs
upgrade Unity environments to use gym API directly. remove old custom code
use
SLM-Env
as git submodule to distribute Unity binaries
add
UnityVecFrameStack
for unity gym env
TODO Unity benchmark next PR
Generalize PPO param inference
generalize PPO
minibatch_size
and
time_horizon
to safely split data for vec envs.
Unity gym envs
UnityVecFrameStackfor unity gym envGeneralize PPO param inference
minibatch_sizeandtime_horizonto safely split data for vec envs.