Different ways of writing Normal distribution in Edward


I want to know what’s the difference between these three Normal in Edward

from edward.models import Normal
import tensorflow as tf
D = 100
a = Normal(tf.zeros(D), tf.ones(D))
b = Normal(tf.zeros(D), tf.ones(1))
c = Normal(tf.zeros(D), 1.0)

Any difference among a,b and c?


Great question. No. Under the hood, we try to do the vectorization for you.