Add TemporalScoreRescaling node #10351
Merged
Conversation
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
adlerfaulkner
pushed a commit
to LucaLabsInc/ComfyUI
that referenced
this pull request
Oct 16, 2025
* Add TemporalScoreRescaling node * Mention image generation in tsr_k's tooltip
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
Resolve #10214.
TSR’s mechanism can be interpreted as a form of non-uniform Epsilon Scaling (for diffusion), so this PR places it in
nodes_eps.py. The default parameters are taken from their t2i demo notebook.Since TSR is an SNR-dependent (schedule-dependent) scaling method, it can produce different effect strengths under the same parameter settings with different schedulers.
According to its formulation, TSR gradually converges to the value of k over time.
Therefore, the same scale used in Epsilon Scaling can be used as its constant version for comparison.
V-Prediction ztsnr
sampler: er_sde, schedule: simple, cfg: 2.5
sampler: er_sde, schedule: beta, cfg: 2.5
SD 3.5 Medium
sampler: euler, schedule: simple, cfg: 5.0
Using
sigma = 3.0should have similar effect shown in their SD3 examples.