[Bug]: Resizing the image in img2img

[alexbofa]

Is there an existing issue for this?

• I have searched the existing issues and checked the recent builds/commits

What happened?

Resizing the canvas does not work as it should.
https://github.com/AUTOMATIC1111/stable-diffusion-webui/assets/58225118/c2cdbd67-f158-402c-8fc3-a953286e97ac

Steps to reproduce the problem

1. Drop image in im2img→tab img2img
2. Press resize canvas

What should have happened?

How it worked before the bugs:
https://youtu.be/ubvFXuPJk5k?t=218

Commit where the problem happens

2b3fc24

What platforms do you use to access the UI ?

Windows

What browsers do you use to access the UI ?

Google Chrome

Command Line Arguments

set COMMANDLINE_ARGS=--opt-sdp-no-mem-attention --autolaunch

List of extensions

Console logs

From https://github.com/AUTOMATIC1111/stable-diffusion-webui
 * branch              dev        -> FETCH_HEAD
error: Your local changes to the following files would be overwritten by merge:
        style.css
Please commit your changes or stash them before you merge.
Aborting
Updating 2b3fc246..1ceb82bc
venv "venv\Scripts\Python.exe"
fatal: No names found, cannot describe anything.
Python 3.10.9 (tags/v3.10.9:1dd9be6, Dec  6 2022, 20:01:21) [MSC v.1934 64 bit (AMD64)]
Version: <none>
Commit hash: 2b3fc246b050f4649a643cfcaf16df60a9ac615a
Installing requirements
Launching Web UI with arguments: --opt-sdp-no-mem-attention --autolaunch
No module 'xformers'. Proceeding without it.
Loading weights [9aba26abdf] from D:\Portable-Stable-Diffusion-dev\models\Stable-diffusion\MAIN--\Deliberate_v2.safetensors
Creating model from config: D:\Portable-Stable-Diffusion-dev\configs\v1-inference.yaml
LatentDiffusion: Running in eps-prediction mode
fatal: No names found, cannot describe anything.
DiffusionWrapper has 859.52 M params.
Applying scaled dot product cross attention optimization (without memory efficient attention).
Running on local URL:  http://127.0.0.1:7860

To create a public link, set `share=True` in `launch()`.
Startup time: 10.6s (import torch: 2.0s, import gradio: 1.7s, import ldm: 0.5s, other imports: 1.1s, load scripts: 1.7s, create ui: 0.9s, gradio launch: 2.5s).
Textual inversion embeddings loaded(0):
Model loaded in 4.3s (load weights from disk: 0.4s, create model: 0.9s, apply weights to model: 0.6s, apply half(): 0.6s, move model to device: 0.6s, load textual inversion embeddings: 1.2s).
100%|██████████████████████████████████████████████████████████████████████████████████| 30/30 [00:17<00:00,  1.72it/s]
Total progress: 100%|██████████████████████████████████████████████████████████████████| 30/30 [00:14<00:00,  2.06it/s]
Total progress: 100%|██████████████████████████████████████████████████████████████████| 30/30 [00:14<00:00,  2.13it/s]

Additional information

It worked for the last time 2-3 months ago before the version of Gradio 3.23.1

[animeking527]

Having the same behavior. Makes resizing, finding areas and relighting techniques impossible.

[inverherive]

Can confirm the issue here too. Alternate resize fail and bounce back to previous canvas. Also dropping an image onto the canvas fails if another one is already present.

[andrewssdd]

The issue is in gradio itself. It has been there for quite some time.

Issue:
gradio-app/gradio#3810

If you really need this function, the best bet is to revert to an older commit that uses gradio 3.16.2

[Beelkic]

I have the same problem

[catboxanon]

Since Gradio has not repaired this functionality in months it's just been removed from the web UI for the time being. #12792

Closing.

[catboxanon]

Re-opening so it's easier to keep track of when this is actually fixed. It should be after Gradio 4 support is merged in. #14184