AUTOMATIC1111 / stable-diffusion-webui

Stable Diffusion web UI
GNU Affero General Public License v3.0
140.66k stars 26.61k forks source link

[Bug]: Why does the quality on pics look bad? #15986

Open blehh666 opened 3 months ago

blehh666 commented 3 months ago

Checklist

What happened?

I freshly installed everything and my pics still look bad. I'm not sure what to do. image (10)

Steps to reproduce the problem

I'm not sure why it's happening.

What should have happened?

It should make a good looking and clean picture and I have no clue as to why it doesn't.

What browsers do you use to access the UI ?

Google Chrome

Sysinfo

sysinfo-2024-06-10-07-08.json

Console logs

venv "C:\Users\sydne\Desktop\bewbies\stable-diffusion-webui\venv\Scripts\Python.exe"
Python 3.10.11 (tags/v3.10.11:7d4cc5a, Apr  5 2023, 00:38:17) [MSC v.1929 64 bit (AMD64)]
Version: v1.9.4
Commit hash: feee37d75f1b168768014e4634dcb156ee649c05
Installing sd-webui-controlnet requirement: changing opencv-python version from 4.10.0.82 to 4.8.0
Launching Web UI with arguments:
no module 'xformers'. Processing without...
no module 'xformers'. Processing without...
No module 'xformers'. Proceeding without it.
Tag Autocomplete: Could not locate model-keyword extension, Lora trigger word completion will be limited to those added through the extra networks menu.
[-] ADetailer initialized. version: 24.5.1, num models: 10
ControlNet preprocessor location: C:\Users\sydne\Desktop\bewbies\stable-diffusion-webui\extensions\sd-webui-controlnet\annotator\downloads
2024-06-10 02:48:54,074 - ControlNet - INFO - ControlNet v1.1.449
Loading weights [67ab2fd8ec] from C:\Users\sydne\Desktop\bewbies\stable-diffusion-webui\models\Stable-diffusion\ponyDiffusionV6XL_v6StartWithThisOne.safetensors
2024-06-10 02:48:54,587 - ControlNet - INFO - ControlNet UI callback registered.
Running on local URL:  http://127.0.0.1:7860

To create a public link, set `share=True` in `launch()`.
Creating model from config: C:\Users\sydne\Desktop\bewbies\stable-diffusion-webui\repositories\generative-models\configs\inference\sd_xl_base.yaml
C:\Users\sydne\Desktop\bewbies\stable-diffusion-webui\venv\lib\site-packages\huggingface_hub\file_download.py:1132: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
  warnings.warn(
IIB Database file has been successfully backed up to the backup folder.
Startup time: 16.7s (prepare environment: 5.8s, import torch: 4.4s, import gradio: 1.0s, setup paths: 1.3s, initialize shared: 0.3s, other imports: 0.8s, load scripts: 1.9s, create ui: 0.7s, gradio launch: 0.4s).
Loading VAE weights specified in settings: C:\Users\sydne\Desktop\bewbies\stable-diffusion-webui\models\VAE\sdxl_vae.safetensors
Applying attention optimization: Doggettx... done.

Additional information

My GPU driver is up to date.

kalle07 commented 3 months ago

what you expect .. its the RAW first pony model ;) its same for me ... download newer once

with embeding may you tune a bit https://civitai.com/models/332646?modelVersionId=509253

AlUlkesh commented 3 months ago

I agree. If you want anime, use an embedding or a specialized pony merge.

Also, for Euler samplers more steps is usually better. And you might also get better results if you use the recommended resolutions for SDXL.

For example, these are with 50 steps at 768x1344: image

However I am not convinced the models know what good or bad hands/anatomy are. If I leave those terms out, I get this: image

viking1304 commented 3 months ago

That is not a bug.

For a usage tips, better go to Discord and ask for help there, where someone with more experience can point you in the right direction. Ask in #general-with-images or #prompting-help, since this is not a techical issue and it is not for #tech-support