bytedance / bytedance/SuperEdit

Training Reproduce Suggestion

Open
#5 1 comment 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
165
Forks
10
PR merge metrics
No merged PRs in 30d

Description

Hi, thanks for your great work!

I've successfully reproduced the training based on sd1.5. However, I'm trying to migrate the SuperEdit into Sana (https://github.com/NVlabs/Sana) but the results look really bad (seems the edited image cannot preserve input image).

Wondering if you have investigated or studied on image editing based on Sana. Appreciate if there's any suggestion :)

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reading the SuperEdit training setup used for sd1.5 and the linked Sana repository, then compare their training and image-editing entry points. Reproduce the reported Sana migration and document the configuration changes needed for edited images to preserve the input image; the issue currently names no files, tests, or exact settings.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
computer-vision, machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.