Implementation of Score Distillation via Inversion #508
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
thuliu-yt16
reviewed
Nov 10, 2024
thuliu-yt16
reviewed
Nov 10, 2024
docs/Pumpkin_head_zombie_skinny_highly_detailed_photorealistic.gif
Outdated
Show resolved
Hide resolved
|
Appreciate your contribution, nice work! Looks good to me except for some redundant files. |
|
Good catch, and thank you for reviewing it so fast. Removed the extra files |
thuliu-yt16
reviewed
Nov 11, 2024
There was a problem hiding this comment.
I think we have two better ways here to avoid changing the default behavior.
- add an option defaulted as False to specify ambient_only in test time
- just keep the original behavior if it is not necessary to keep ambient_only during testing
There was a problem hiding this comment.
Good point. I returned it to the default behavior and went with option 1 by adding a new option in the config of the material
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Here is implementation of our paper on analyzing Score Distillation by viewing it as re-parametrization of DDIM, what suggests that the noise sample should be carefully chosen in SDS. In the PR we implement an improved method called SDI where the noise term is regressed via inverting DDIM.
The paper is accepted to NeurIPS 2024.
ArXiv: https://arxiv.org/abs/2405.15891
Webpage: https://lukoianov.com/sdi
The pull request contains implementation of SDI as a module in threestudio and an additional notebook with 2D optimization using different strategies of noise term sampling.