kohya-ss / kohya-ss/sd-scripts

Support REPA

Open
#1,694 2 comments 5 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
7.2k
Forks
1.2k
Avg merge
11m
Merged PRs (30d)
2

Description

https://github.com/sihyun-yu/REPA

REPresentation Alignment (REPA), a simple regularization technique built on recent diffusion transformer architectures.

reference code:
https://github.com/sihyun-yu/REPA/blob/0f6025751aae6746636113430077169abee26ceb/train.py#L265-L300
https://github.com/sihyun-yu/REPA/blob/0f6025751aae6746636113430077169abee26ceb/loss.py#L77-L88

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reading the referenced REPA implementation in train.py lines 265-300 and loss.py lines 77-88. Then inspect sd-scripts for the corresponding training and loss entry points, since the issue does not name target files or tests. Done means REPA support is integrated and its behavior is verified against the reference implementation.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.