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

Reconsider the softmax distribution #24

Open
cheind opened this issue Dec 12, 2021 · 2 comments
Open

Reconsider the softmax distribution #24

cheind opened this issue Dec 12, 2021 · 2 comments

Comments

@cheind
Copy link
Owner

cheind commented Dec 12, 2021

PixelCNN++ argues that neighboring intensity usually correlate is not captured by the softmax distribution. Instead, they propose a mixture model consisting of the logitic distribution (like normal but with heavier tails).
https://arxiv.org/pdf/1701.05517.pdf
@cheind
Copy link
Owner Author

cheind commented Dec 12, 2021 •
edited
Loading

also add bits-per-dim metric that measures how many bits are required to encode pixel intensities.
see E.2 of https://arxiv.org/pdf/1705.07057.pdf
https://uvadlc-notebooks.readthedocs.io/en/latest/tutorial_notebooks/tutorial11/NF_image_modeling.html

@cheind cheind mentioned this issue Dec 12, 2021
Open
@cheind
Copy link
Owner Author

cheind commented Dec 13, 2021

See also for how to optimize based on cdf
https://github.com/Rayhane-mamah/Tacotron-2/issues/155

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@cheind