[Bug]: vram not released after unable to create image due to lack of vram

[hungtooc]

Is there an existing issue for this?

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

What happened?

everything worked fine until i created an image with a high configuration that my machine couldn't handle, of course the image wouldn't be able to create but then the vram wasn't freed.

Steps to reproduce the problem

use --api and create txt2img via api
created an image with a high configuration that my machine couldn't handle, usually create very high resolution images

What should have happened?

image cannot be created, then release the vram

Version or Commit where the problem happens

version: v1.5.1  •  python: 3.11.3  •  torch: 2.1.0.dev20230727+cu121  •  xformers: N/A  •  gradio: 3.32.0  •  checkpoint: ed989d673d

What Python version are you running on ?

Python 3.11.x (above, no supported yet)

What platforms do you use to access the UI ?

Linux

What device are you running WebUI on?

Other GPUs

Cross attention optimization

None

What browsers do you use to access the UI ?

Google Chrome

Command Line Arguments

--api

List of extensions

Console logs

out of vram

Additional information

none

[catboxanon]

I think this was fixed on the dev branch with 0af4127 and ccb9233. When you can I would checkout the dev branch and see if that resolves it or not.

[hungtooc]

I think this was fixed on the dev branch with 0af4127 and ccb9233. When you can I would checkout the dev branch and see if that resolves it or not.

oh thanks, my mistake, I will checkout first

[catboxanon]

Just a heads up, I made two PRs to fix the (to my knowledge) remaining VRAM issues. Might want to wait until these are reviewed and merged too.
#12514
#12515