stable-diffusion-webui
stable-diffusion-webui copied to clipboard
[Bug]: Random overbaking problem?
Is there an existing issue for this?
- [X] I have searched the existing issues and checked the recent builds/commits
What happened?
Every now and then I just randomly seem to get an extremely overbaked image. And it doesn't appear to be anything settings related (which I thought initially) because just now I got one overbaked image and one "correct" image in the exact same batch! I had batch size of 2, so both images being generated at the same time and one of them is overbaked and the other isn't.
What the heck can be going on here?
RTX 3090 ti if it matters.
Steps to reproduce the problem
Only seems to happen in img2img
What should have happened?
Both images should have been like the right image
Commit where the problem happens
0cc0ee1bcb4c24a8c9715f66cede06601bfc00c8
What platforms do you use to access the UI ?
Windows
What browsers do you use to access the UI ?
Mozilla Firefox, Google Chrome
Command Line Arguments
--api --xformers --opt-split-attention
List of extensions
StylePile, a1111-stable-diffusion-webui-vram-estimator, depthmap2mask, openOutpaint-webUI-extension, openpose-editor, sd-webui-ar, sd-webui-controlnet, sd-webui-depth-lib, sd-webui-llul, sd_grid_add_image_number, stable-diffusion-webui-artists-to-study, stable-diffusion-webui-composable-lora, stable-diffusion-webui-images-browser, stable-diffusion-webui-two-shot, unprompted
No extension features were used/enabled
Console logs
E:\stable-diffusion-webui>webui-user.bat
venv "E:\stable-diffusion-webui\venv\Scripts\Python.exe"
Python 3.10.5 (tags/v3.10.5:f377153, Jun 6 2022, 16:14:13) [MSC v.1929 64 bit (AMD64)]
Commit hash: 0cc0ee1bcb4c24a8c9715f66cede06601bfc00c8
Installing requirements for Web UI
Fetching updates for midas...
Checking out commit for midas with hash: b845b78...
Installing requirements for Unprompted - img2pez
Installing requirements for Unprompted - pix2pix_zero
Launching Web UI with arguments: --api --xformers --opt-split-attention
Loading Unprompted v7.6.0 by Therefore Games
(SETUP) Initializing Unprompted object...
(SETUP) Loading configuration files...
(SETUP) Debug mode is False
Loading weights [e04b020012] from E:\stable-diffusion-webui\models\Stable-diffusion\rpg_V4.safetensors
Creating model from config: E:\stable-diffusion-webui\configs\v1-inference.yaml
LatentDiffusion: Running in eps-prediction mode
DiffusionWrapper has 859.52 M params.
Loading VAE weights specified in settings: E:\stable-diffusion-webui\models\VAE\vae-ft-mse-840000-ema-pruned.safetensors
Applying xformers cross attention optimization.
Textual inversion embeddings loaded(17): Artgerm, ArtgermStyle-neg, ArtgermStyle, CharTurner, gta5-artwork, HyperStylizeV6, MakeItPopVA, micromini, star-tours, Style-Empire, Style-Glass, Style-Hamunaptra, Style-Moana-neg, Style-Moana, Style-Psycho-neg, Style-Psycho, Style-TronLegacy-8v-B
Model loaded in 4.8s (create model: 0.4s, apply weights to model: 1.4s, apply half(): 0.6s, load VAE: 0.3s, move model to device: 0.6s, load textual inversion embeddings: 1.4s).
[VRAMEstimator] Loaded benchmark data.
Running on local URL: http://127.0.0.1:7860
To create a public link, set `share=True` in `launch()`.
100%|██████████████████████████████████████████████████████████████████████████████████| 75/75 [00:38<00:00, 1.92it/s]
Total progress: 100%|██████████████████████████████████████████████████████████████████| 75/75 [00:42<00:00, 1.78it/s]
Additional information
No response
Question: Have you been using ControlNet, following the instructions to turn "Noise multiplier for img2img" to 0 or perhaps adjusted it too high after? It'll do VERY odd things to images.
This was not using ControlNet in the instance above, though I do use it often. What is this about noise multiplier? First I've heard of it.