How to Install AUTOMATIC1111 + SDXL1.0 - Easy and Fast!
Вставка
- Опубліковано 5 лип 2024
- With the launch of SDXL1.0, there's never been a better time to test out Stable Diffusion. Check out this Beginner friendly install guide for Automatic 1111, SD's most popular webUI.
In this video I'll walk you through installing A1111 using git, installing python, show you all the best startup commands and run through all of A1111's settings to get you started.
NOTE: The newest release of A1111 (1.6) has native SDXL Refiner support, so you no longer need to install the extension, you can just skip that part!
🔗 Links: 🔗
--------------------
Python 3.10.6: www.python.org/downloads/rele...
Git: git-scm.com/download/win
A1111: github.com/AUTOMATIC1111/stab...
A1111 Commit History: github.com/AUTOMATIC1111/stab...
SDXL 1.0 Model: huggingface.co/stabilityai/st...
SDXL Refiner: huggingface.co/stabilityai/st...
SDXL Width & Heights: / sdxl_resolution_cheat_...
📖Chapters:📖
--------------
00:00 START
00:20 DOWNLOADING PYTHON & GIT
01:22 DOWNLOADING A1111
02:21 UPDATING AND REVERTING WITH GIT
03:10 DOWNLOADING SDXL
04:12 LAUNCH SETTINGS
05:35 REFINER EXTENSION
06:14 A1111 SETTINGS - PROMPT BOXES
07:10 SAMPLING METHODS
07:37 SAMPLING STEPS
07:58 WIDTH AND HEIGHT SETTINGS
08:30 BATCH COUNT AND BATCH SIZE
08:49 CGF SCALE
09:05 SEED
09:15 SDXL REFINER EXTENSION
09:41 GENERATING AN IMAGE
10:02 THANKS & OUTRO
---------------- - Наука та технологія
Was this VAE problem your describing at 4:40 addressed in CIVITAI's Jul 28, "SD XL 1.0 VAE fix" ? Would post the link to it but I don't want this comment to get autoflagged by YT
Good Question! I think the issue with the original 1.0 VAE had more to do with color desaturation I, than black outputs, and the fix that was released improves that. I had been using --no-half-vae for 1.5, as default so I left it in the tutorial because I find it useful in general. I might take a deeper dive into this subject at some stage 👍
Things i learned from your video: you can put CMD in the explorer, and git pull. Thanks.
After trying to install StableDiffusion with automatic1111 from different videos, I have to say, this is the most complete one. THANK YOU
Great tutorial!
I've had A1111 for a good year now, over time it was cluttered with so much junk I was unable to update it with "git pull" as well as many other things I wanted to add and remove. Happy to report the new install is great and running much faster for me, thanks so much! for those who are willing to learn... ANYTHING IS POSSIBLE!
Nice tutorial buddy straight and simple👍
nice job! you got a new subscriber. 🙂
Thanks! Much appreciated 😀
THnaks for the Tutorial, very clear and easy. looking foward for more videos from you, maybe explainning how to use the tool more deeply, like imgtoimg tab and so on. thanks man!!
Thanks for the feedback! I was thinking my next video would be on img2img, so watch this space 😀
if you guys are getting an error when running the webui-user batch file saying "runtime error: could not install torch", to fix this look up how to install pytorch manually and follow the video tutorials. Also deleting the venv file in the a1111/stable diffusion file seemed to help too
Bro, Much Obliged for this video
Thanks for watching!
I am new and your very helpful tutorial worked! I can create cool images. In this tutorial, you briefing mention Civitai - What's the best video to watch to learn how to use what is there?
Following your tutorial, all I had to do was dl the SDXL files, as I already had everything else. But when web ui is open, and I select the SDXL checkpoint, the processing loads and when it's finished, it switches to SD 1.5. So I can't use SDXL. Any idea why? 🙏
What about the CUDA drivers? Does the installer install those on its own or do I have to download them separately?
can you start what is an SDXL and how is that different from all the other models?
If SDXL is the checkpoint then how do I use other checkpoints that require SDXL? such as Pony Diffusion
ive been running v1.5, so can I just drop SDXL base and refiner into the program files and use sdxl models
Hi, i followed the tutorial step by step and everything seems to look the same, but in the last step when i launch the webui-user batch file, it loads perfectly but when it finishes nothing opens up. What should i do? or where shoul i try to open it up? or should it pop up automatically?
Thank you very much!
I thought this was going to be the tutorial that finally works for me... but when I go to open the bat to finally run it, I get some code in the command prompt with other stuff that says "fatal: detected dubious ownership in repository at". omg... I'm not a coder... I hit a big wall after so much time dedicated to getting this to work. Where could I go for help on that?
Everything was fine, 10 seconds after launch I got the following error: : Error connection errored out. And nothing works
Thanks for the fresh install help. Not sure why but I don't see the "refine faces" option. Was there an update that eliminated that?
Hey there! On newer versions you have to enable it in settings. Click the Settings Tab > Look for Face Restoration and you can toggle it on there. If you like toggling it on and off, you can add into your quick settings: Go to the settings tab > user interface > Quick settings list > then add face restoration. Hope that helps!
Thank you! @@InciteAI
I've followed the exact steps, prompt, settings etc but just getting "OutOfMemoryError: CUDA out of memory" when hitting the Generate button. I have two NVIDIA GeForce GTX 1080 GPUs (8 GB each) but couldn't figure out how to use both if it's possible. Do you have any workarounds?
To my knowledge there is no support for using dual GPUs for a single render for any SD WebUIs currently available, but I know it's a feature request for both a1111 and InvokeAI. At the moment you're limited to 8GB VRAM then, but you should be able to get away with it. You can check out my video on speeding up SDXL, but the main thing is to make sure as much of your VRAM is available as possible before starting a generation, I'll often get that error if I forget to close Photoshop etc. You can also try --lowvram in your launch options instead of --medvram.
bro specification required to download sdxl ??
When i try to open de webui user i says that has an error, unrecognized arguments, what should i do?
Very good video and got to the point much much faster than other videos. I got A1111, and the SDXL 1.0 installed (and moved this and refiner to the 'Models' folder).
The images i'm seeing are good. However, there is still a lot of issues with the eyes and fingers of the outputs. Like some of them are rather deformed (even after i included this in the Negative prompt box). Am i the only one?
Hi try adding this to the prompt, detailed eyes, detailed face, If you add (detailed eyes), to the prompt it makes the image stronger in that instruction to the A.I. Fingers & hands are not perfect however there are tutorials on youtube on how to fix this. The learning curve is very steep with A1111, time & patience is the key.
Thank you for the excellent explanation. The installation went well. I followed all the steps. The whole process took some time, but in the end, I managed to open SDXL in the web browser. However, when I tried to install the refiner and generate an image, I received an error message: 'SSLError: not enough data: cadata does not contain a certificate (_ssl.c:3989)'. If anyone knows anything about this, please help. Thank you in advance.
ddi you figure it out?
@@TheKillingPerfection In essence, when the installation is initiated, the program will start downloading the files necessary for generating images. I encountered a problem, likely due to a slower internet connection. During the installation, the download of certain files was interrupted, causing errors upon startup. I identified the missing files and independently downloaded them from websites such as Hugging Face and similar sources. Subsequently, I organized these files into their respective folders, such as checkpoints, clip_vision, etc. It can be a bit tedious to figure out what exactly is missing, but in the end, I successfully launched SDXL. Everything is now functioning properly.
I get an error when i try to simply load the SDXL model, i select it and it says error
There could be a number of reasons for that. If you keep an eye on the command prompt window that opens when you launch a1111, it should give you a more detailed error that should give you more of an idea of what's going on.
Hi, thanks for the instruction. However Im presuming this is NOT going to be a worthwhile experience with my 4 year old GPU an AMD RX 5700 XT ? Is there anywhere I can find the minimum PC requirements?
My PC is a Ryzen 1700 at 3.8mhz 16GB ram, with a nvidia 2070 8GB vram , My channel has my work on it... Judge me!
I have rtx 3070 GPU and trying to create a page for colouring book but it's taking 30 minutes to create one image. Can can I make it fast?
Please help me.
Thanks
ı have 360 and same problem
@@claylgruber7994 I got the solution. I'll explain you shortly.
doesn't seem to work with AMD card?
I'm stuck at this error when trying to run the batch file saying that Python is not recognized as an internal or external command, operable program or batch file
Anyway to fix this?
First thing comes to mind is to check if you have multiple versions of python installed, or was "add to path" checked when installing python?
Best vide thus far... I keep getting this error... when clicking webui-user batch file. I have python loaded. I downloaded the version but why I inster "py" into the cmd prompt. I am said that I have my anaconda 3.11 version running. Any help with this?
"exit code: 9009
stderr:
'python' is not recognized as an internal or external command,
operable program or batch file."
Hmm , not sure, I don't use anaconda so I'm not that familiar with it, but it sounds like it's not finding your python file. Found this on reddit so I take no credit, or responsibility if it works or doesn't xD:
Open Webui-user.bat in notepad
You'll see Python=
Find your python.exe install directory and add it after the equals
Example
Python=C:/python310/python.exe
Some people's python installs to program files, if that's the case, uninstall it and install it to a folder that has no spaces. The installer won't work if any file paths have spaces. Example that wouldn't install "C:/program files/python"
@@InciteAI I was able to get it to work through Anaconda's CMD prompt. It has been working great. Thanks for the video.
I was wondering is there a reason why mine is all white? I like how yours looks better. Is there a "dark" setting as I can't find one?
Also, any specific videos of yours you would want me to share when I start doing videos sharing my experience trying this out on FB?
I keep getting the error
exit code: 1
stderr:
D:\python.exe: No module named pip
Launch unsuccessful. Exiting.
Press any key to continue . . .
I have followed the instructions to the letter I am not entirely sure what this error means
Hi! PIP is a package manager for Python and should be included when you install Python 3.10.6. If you installed python just like in the video (Ticking the box for PATH), you can try opening a command box: Windoes menu > run > type in 'cmd'. In the command box type in 'python -m ensurepip' (withouth the quotation marks). Let me know if that works!
If you have a graphics card with 8 gigs or less , type dash dash ??? (what??) vram? I can't hear what you said there. Plus your text was so small I could not read it. Please help
You should probably go full screen, or take a screenshot
Hmmm. Tried your generation settings and it took me 10mins. Why :(
so no need to load lora ?
Nope, the Lora is optional! I actually haven't been using it much at all!
yeeeeh mine does not work at all
There no more refiner on the huggingface site. Why is that?
Check video description it's a different link now
I got an error and don't understand
Hey there, what was the error that you got and at what stage? Maybe there's an easy fix?
@@InciteAI I try to send a link but message doesn't come to you
You have to be online to use this.... it's kind of pointless. I thought this was so you can use it without Internet connection. It does clearly say ( web ui )
Following your tutorial, all I had to do was dl the SDXL files, as I already had everything else. But when web ui is open, and I select the SDXL checkpoint, the processing loads and when it's finished, it switches to SD 1.5. So I can't use SDXL. Any idea why? 🙏