Open
Conversation
- Introduced a new script to investigate the impact of classifier-free guidance (CFG) on image collapse during extended EqM sampling. - Implemented baseline sampling with CFG and a switch experiment to analyze the effects of transitioning to null class labels. - Captured intermediate latents and saved generated images at specified steps. - Added argument parsing for various experiment configurations including model selection, image size, and sampling parameters.
…tput folder handling - Updated IntermediateImageSaver to accept a folder pattern for dynamic folder naming. - Modified GradientNormTracker's finalize method to directly accept parameters for sampling statistics. - Adjusted calls to these classes across multiple scripts to ensure consistent usage of the new parameters.
…rameter for consistency across scripts
…ross multiple scripts for improved clarity and default behavior
- Moved IntermediateImageSaver, GradientNormTracker, and other sampling hooks from utils.sampling_utils to utils.sampling_hooks for better organization. - Updated import statements across multiple scripts to reflect the new module structure. - Improved documentation to clarify the usage of the new hook system.
- Updated the `forward_with_cfg` method in `models.py` to include a new argument `return_components`, allowing the return of individual conditional and unconditional outputs. - Modified `GradientNormTracker` in `sampling_hooks.py` to track and compute statistics for conditional and unconditional gradient norms when CFG is enabled. - Adjusted sampling functions in `sampling_utils.py` to pass conditional and unconditional outputs to hooks, improving monitoring capabilities during sampling. - Enhanced documentation to clarify the new functionality and its implications for gradient norm tracking.
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
No description provided.