AUTOMATIC1111
[Bug]: Extra networks - Clicking "replace preview" shows all cards instead of those filtered by search keywords.
Is there an existing issue for this?
- [X] I have searched the existing issues and checked the recent builds/commits
What happened?
In the feature of "extra networks", there's a button to "replace preview" image of an extra network card with the current image shown in the gallery.
When I type some keyword in the "Search..." textbox, only the extra network cards with the keyword I typed will be shown. But at this point I click on "replace preview", all the cards will be shown regardless of the keyword typed in "Search..." textbox.
Steps to reproduce the problem
- Go to txt2img, type some prompts.
- Press "Generate" button to generate a new image.
- Click on "Show extra networks".
- Type some keyword to filter some extra networks.
- Click on "replace preview" of an extra network card.
What should have happened?
I expected that only the cards with the keywords typed in "Search..." box will be shown.
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 ?
Microsoft Edge
Command Line Arguments
webui-user.bat
List of extensions
No
Console logs
> webui-user.bat
venv "...\stable-diffusion-webui\venv\Scripts\Python.exe"
Python 3.10.6 | packaged by conda-forge | (main, Oct 24 2022, 16:02:16) [MSC v.1916 64 bit (AMD64)]
Commit hash: 0cc0ee1bcb4c24a8c9715f66cede06601bfc00c8
Installing requirements for Web UI
Launching Web UI with arguments:
No module 'xformers'. Proceeding without it.
Loading weights [bb6e6362d8] from ...\stable-diffusion-webui\models\Stable-diffusion\....safetensors
Creating model from config: ...\stable-diffusion-webui\configs\v1-inference.yaml
LatentDiffusion: Running in eps-prediction mode
DiffusionWrapper has 859.52 M params.
Applying cross attention optimization (Doggettx).
Textual inversion embeddings loaded(27): ...
Model loaded in 5.1s (create model: 0.6s, apply weights to model: 0.6s, apply half(): 0.8s, move model to device: 1.8s, load textual inversion embeddings: 1.3s).
Running on local URL: http://127.0.0.1:7860
To create a public link, set `share=True` in `launch()`..
Additional information
No response
Can confirm this is an "relatively" old issue. What happens is when you click replace preview, on Embeddings/Lora/Hypernetworks tab. That tab will refresh and not carry over the filter. SO you have to to delete 1 letter or add 1 new one every time you replace/add a preview.
As a sidenote, could we it filter in alphabetical order as well?
Updated UI looks to have resolved this issue. If the problem persists as of the latest commit on the dev branch, open a new issue.
