Randomized-Ensembled-Double-Q-learning-REDQ- icon indicating copy to clipboard operation
Randomized-Ensembled-Double-Q-learning-REDQ- copied to clipboard

Published 20 hours ago • verified publisher icon BY571
→

Metadata

Pytorch implementation of Randomized Ensembled Double Q-learning (REDQ)

Results 2 Randomized-Ensembled-Double-Q-learning-REDQ- issues
Sort by recently updated
recently updated
newest added

Actor Update Bug

Should the actor update not utilise idx[0] and idx[1] for Q1 and Q2? currently it just gets the same value of Q from the same critic # ---------------------------- update actor...

beardyFace avatar beardyFace

Application to discrete action space

Hi, I was wondering that have you applied this idea to environments with discrete action space? or have any idea how it would perform? Thanks

merv22 avatar merv22

← Metadata

20
Stars
0
Forks
Watchers

Owner

verified publisher icon BY571

Metadata

Pytorch implementation of Randomized Ensembled Double Q-learning (REDQ)

Back