-
Notifications
You must be signed in to change notification settings - Fork 24
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Bug]: Incompatible with the latest Automatic1111 updates #46
Comments
hey, so the send-to buttons were definitely broken but should be fixed as of 566a8a5 - that said, the base functionality (dream/img2img) is certainly working as expected in my main webUI with a billion extensions as well as a "clean" install... what are the console errors you're getting? |
The changes shown below seem to be causing issues for these colab instances: Fixed by adding:
|
Testing again with the following commits: All other extensions disabled: Restart webui, run test prompt in txt2img, click "Send to OpenOutpaint", getting the following in the console:
|
Same setup, reload webui, go to OpenOutpaint tab and attempt to dream with "test prompt" and same DeliberateV2 model loaded: 500 error on http://127.0.0.1:7860/sdapi/v1/txt2img
HAR link: https://drive.google.com/file/d/1p9l52kdWbXfOQfxgzJwg0j3d16MZ9yAN/view?usp=share_link |
Same process except I added content to the canvas with a resource stamp, response from Img2Img:
|
Just to be safe I also nuked the extension folder and did a fresh checkout, same issues present |
I also have a separate installation of Automatic1111 that is locked in at the pre-March 25'th commits where it works perfectly fine. That is where this installation came from, I duplicated that folder and then did a git pull to setup the second installation |
@kutyadog yeah, apologetically i have like almost no familiarity with colabs in general for the most part, but if a previous version of the extension works better for that there's actually a commandline flag specifically for those instances - just add @Zetaphor i can confirm that svelte throws up an error about v being null in web console but the extension still works as expected and i have no clue where to even start with minimized gradio lol; have you seen this issue? could you try going to http://127.0.0.1:7860/docs#/default/text2imgapi_sdapi_v1_txt2img_post, click
and execute it? it should return a base64 image |
Apologies for the delay @zero01101, looks like that's also failing on the server side: |
It also fails with the default parameters, looks like the API is busted for some reason? |
yeah, there's AUTOMATIC1111/stable-diffusion-webui#9319 open right now to deal with that, but openOutpaint definitely relies on the API functionality to be operational. tested send-to buttons on native linux as well which also work as expected on current commits for webUI, openOutpaint, and oo-extension so this is all seeming to point at colab vs webUI updates |
Strange, I'm running on Linux, not colab or docker. I guess I'll try reinstalling from the latest commit |
Hi bro, thanks a lot for the excellent work. I thought outpainting was cool but never had success using it. I am using automatic1111 colab version, I read the instructions carefully and know that the code --api is added to the web user.sh file, but every time after saving it, the --api disappears when I start it again, so it never works, how should I do it? |
Has there been any progress with this? I still can't run it in automatic, and the same issue seems to exist with vlad diffusion |
terribly sorry for the extended delay; this project isn't abandoned but i've had to step away from it due to Reasons for a while, and i admit additional responses may (aka will probably) be delayed but hopefully not nearly as long @253189147cc thanks for the kind words; i think that's a separate issue to the topic at hand, but i believe there's a different way of enabling API on colabs? looking at https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast_stable_diffusion_AUTOMATIC1111.ipynb it appears to be enabled by default in the bottom block? @Raivshard i can definitely at least confirm that base functionality should still work almost seamlessly in chrome/chromium-based browsers, and firefox just needs to get kicked in the refresh button on the openOutpaint tab to fire up on both windows and linux hosted instances, but is still kind of frustrating... i tried the extension in vlad's fork like RIGHT when the whole dustup was starting and it worked out of the box, but that was also about when i had to step back for a bit so things may have changed since then. are you getting any particular errors in f12 console logs or webUI console? does the txt2img API test mentioned above pass? |
If you had asked me this a couple of weeks ago, the answer would have been no, which is why I removed the extension in the first place, but it does seem to work n-- Nope. @zero01101 the API test is not useable on vlad's fork, as the link still returns {"detail":"Not Found"} edit: did a fresh install of vlad's and everything works fine now. Something stopped breaking something else, I guess. 😁 |
Is ths issue about the extension?
What happened?
Automatic1111 recently merged a large number of PR's, which included changes to the frontend, gradio version, and I believe even the API. Currently the "Send to OpenOutpaint" buttons are non-functional, and the Dream and Img2Img options are not functioning.
Steps to reproduce the problem
What should have happened?
The options should have produce output. Instead they are producing console errors
Commit where the problem happens
46534e8
What platforms do you use to access openOutpaint?
Windows, Linux
What browsers do you use to access the UI ?
Mozilla Firefox, Google Chrome
Browser Extensions/Addons
uBlock origin
AUTOMATIC1111 webUI Commandline Arguments
--api --cors-allow-origins=*
Additional information
My extensions also broke lol, the madlad merged like 40+ PR's in a single day and then they've been pushing bugfixes since
The text was updated successfully, but these errors were encountered: