lllyasviel / lllyasviel/ControlNet
Fixed prompt leads to poor performance？
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 34.1k
- Forks
- 3k
- PR merge metrics
- No merged PRs in 30d
Description
I tried using ControlNet to improve the quality of FFHQ images. When the model started training, the sampling results were normal, but when it reached the later stage of training, strange output effects appeared. Please help me solve this problem. Thanks you very mush!!
Fixed prompt: photo of portrait, HDR, UHD, 8K, highly detailed eyes, high detailed skin, high detailed mouth, high detailed nose
Image Input: FFHQ, 512x512
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
No source file, test, training configuration, or reproducible command is provided. Start by reproducing the FFHQ 512x512 ControlNet training with the fixed prompt and compare early and late sampling outputs. Done would require isolating the cause of the degradation and documenting a verified correction.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 18/100