example_Deep Gaussian Process #2227
Answered
by
gpleiss
findoctorlin
asked this question in
Q&A
-
|
Hey guys! I am running the code for Deep Gaussian Process. Can someone tell me why the predictive mean here is 10x3320 instead of 1x3320? Since the output dimension is 1. |
Beta Was this translation helpful? Give feedback.
Answered by
gpleiss
Jan 4, 2023
Replies: 1 comment
-
|
The deep GP model outputs 10 samples of the (approximate) predictive posterior distribution. (We draw samples from the hidden layer, and pass them into the final layer, which outputs a MVN. The output is |
Beta Was this translation helpful? Give feedback.
0 replies
Answer selected by
findoctorlin
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The deep GP model outputs 10 samples of the (approximate) predictive posterior distribution. (We draw samples from the hidden layer, and pass them into the final layer, which outputs a MVN. The output is
num_samplesMVN distributions.)