hiro/ray
1
0
Fork 0
mirror of https://github.com/vale981/ray synced 2025-03-08 19:41:38 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions

[rllib] Entropy calculation for diag gaussian missing 0.5 term (#2968)

Browse source 
See: https://en.wikipedia.org/wiki/Multivariate_normal_distribution#Entropy
This commit is contained in:
Eric Liang 2018-09-29 22:57:47 -07:00 • committed by Richard Liaw
parent b3f0dcf20b
commit cb56f39070
1 changed files with 1 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
python/ray/rllib/models/action_dist.py
View file

@ -125,7 +125,7 @@ class DiagGaussian(ActionDistribution):
def entropy(self): def entropy(self):
return tf.reduce_sum( return tf.reduce_sum(
self.log_std + .5 * np.log(2.0 * np.pi * np.e), .5 * self.log_std + .5 * np.log(2.0 * np.pi * np.e),
reduction_indices=[1]) reduction_indices=[1])
def sample(self): def sample(self):