Use torch.svd_lowrank for large matrices in resize_lora.py#2240
Merged
kohya-ss merged 1 commit intokohya-ss:sd3from Mar 29, 2026
Merged
Use torch.svd_lowrank for large matrices in resize_lora.py#2240kohya-ss merged 1 commit intokohya-ss:sd3from
torch.svd_lowrank for large matrices in resize_lora.py#2240kohya-ss merged 1 commit intokohya-ss:sd3from
Conversation
Owner
|
Thank you, this is really nice! I'd like to merge this sooner. I'm a little busy so please wait a moment. |
Owner
|
Thank you for this! |
kohya-ss
added a commit
that referenced
this pull request
Mar 29, 2026
Allow users to control the number of iterations for torch.svd_lowrank on large matrices. Default is 2 (matching PR #2240 behavior). Set to 0 to disable svd_lowrank and use full SVD instead. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2 tasks
Owner
|
I've added an option |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This hugely reduces the time to prune a LoRA, without visible quality loss compared to the full SVD. The time is reduced from 3 hours (Wan) and 1 hour (Qwen-Image) to a few seconds on my laptop.
See https://docs.pytorch.org/docs/stable/generated/torch.svd_lowrank.html for details. Maybe we can increase
niterfor even better quality but I think the defaultniter=2already gives good enough quality. Kijai'sLoraExtractKJnode usesniter=7by default, see https://github.com/kijai/ComfyUI-KJNodes/blob/bb205d809b467307b8ec3bb1a22680a4873187f8/nodes/lora_nodes.py#L30 (but they don't yet do adaptive rank after svd_lowrank)