Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Generated samples are very homogenous for conditional sampler #9

Open
pat-alt opened this issue May 2, 2023 · 0 comments
Open

Generated samples are very homogenous for conditional sampler #9

pat-alt opened this issue May 2, 2023 · 0 comments
Labels
help wanted Extra attention is needed question Further information is requested

Comments

@pat-alt
Copy link
Member

pat-alt commented May 2, 2023 •
edited
Loading

I've noted that generated samples for MNIST are quite homogenous and originally thought this was an implementation issue:

image

Having looked at this a little more and playing around with the way losses are computed (on an instance basis vs. aggregated across batch) I'm no longer sure. I also noted that the unconditional samples in Tom are also homogenous:

image
@pat-alt pat-alt added bug Something isn't working help wanted Extra attention is needed question Further information is requested and removed bug Something isn't working labels May 2, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
help wanted Extra attention is needed question Further information is requested
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@pat-alt