The Easiest Stable Diffusion With Quality. Fooocus Tutorial.
ฝัง
- เผยแพร่เมื่อ 14 ส.ค. 2023
- github.com/lllyasviel/Fooocus
This is the recommended Fooocus fork now: github.com/runew0lf/RuinedFoo...
FREE Prompt styles for Stable diffusion a1111 & Vlad/SD.Next: / sebs-hilis-79649068
FREE ComfyUI workflow for 1.5 models: / comfyui-1-5-86145057
FREE ComfyUI Workflow for SDXL: / comfyui-workflow-86104919
Support me on Patreon to get access to unique perks! / sebastiankamph
Chat with me in our community discord: / discord
My Weekly AI Art Challenges • Let's AI Paint - Weekl...
My Stable diffusion workflow to Perfect Images • Revealing my Workflow ...
ControlNet tutorial and install guide • NEW ControlNet for Sta...
Famous Scenes Remade by ControlNet AI • Famous Scenes Remade b...
Early access to videos as a Patreon supporter www.patreon.com/sebastiankamph
is it possible to upload an image of myself and create new images with my face? i tried uploading an image in the input tab but it changes my face
I love the fact that you put the dadjokes mid through the video, because you KNOW that's the reason we are here for!
Got to help me with the algorithm! But you're already doing that, aren't you? 🌟😊
@@sebastiankamph Man, i shared lots of yours tutorials, the problem is no one I know has interest in or is pro-AI, so i just talk about AI in facebook groups that already know you better than I do. Not my fault, sorry.
@@IlRincreTeam Haha, that's fantastic! Keep up with that great work 😊 I love it when people share my content
Haha The Dad jokes are a bonus, I love Sebastian's content to. But I've told the 'Camera has Photos, Socks have Five Toes' joke to many people to great success. So yeah...
My ex-wife the Digibetic was asking me when I would install comfyUI on her laptop, I procrastinated that not because I hate her or because her laptop can't handle it BUT simply because she is a digibetic, for every thing that doesn't work she comes to me to solve it and imagine how many things we need to look up and search and do not want to be disturbed with nuiscance calls when we are enjoying ourselves. So yeah Ima install Fooocus on her laptop, it will save me a lot of frustration, Ima keep ComfyUI for myself because of more freedom and infinite variations upon variations
Thanks for the tutorial videos man, it has been really easy to catch up on AI stuff with your help.
Glad to hear it!
This is my favorite Stable Diffusion channel I've found... I'm a beginner with SD and once i get up to speed, I'll try to contribute my recommendations on using it with a "potato computer". I'm running it on an NVidia GTX 1060 3gb so will do my best to see if I can max it out. Love your videos!
Hey Pete, glad to have you board and thanks for all the engagement and comments! 🌟😊
Wow! Awesome. I was hoping this had both in & outpainting but maybe that will come pretty soon?
You should make a video on how to create new styles for the "styles" tab, since it´s quite easy to do and very useful.
Thank you. Very helpful.
You're welcome! Thank you for the never-ending support 😁
Thanks for the informative walkthrough of using Stable Fusion with Focus! It seems like a powerful tool. Can you share some tips or techniques for optimizing the settings in Stable Fusion to achieve even more impressive results with image generation?
I appreciate the fact that you included the dad jokes in the middle of the video because you understand that's exactly what we're here for!
Awesome tool, thanks
Installed easily and looks great. What it needs is the ability to combine styles, and to create custom resolutions/aspect ratios. So many styles but it seems only one at a time. Might also be more usable in one or more drop-downs. An excellent start!
You can definitely select multiple styles [all if you want]. From my understanding styles are really just hidden text prompts... so using multiple at a time is probably just a lot of trial and error.
Another great find master Kampf; this is great for noobs, but I prefer the power and flex of A1111 or Comfy... just the same, for beginners who want midjourney-like point and shoot 1 shots from their prompts? This is fantastic. Also the lineart styles (actually, several of the styles) are really great and accomplish stuff that's more or less a PITA if working from scratch in A1111. Many thanks!
comfy is a mess of error messages for me and A1111 takes 3 minutes to load models on 3060 12gb rn so I have to use this gen to even use sdxl 😅
Hey Sebastian,
Please provide another AI Tools tutorials (Open Source) like Voice/Video AI Tools.
Thanks for your best channel 🥰
Every time I try to run it, I get a message that just says "Error". My system meets all the requirements and I've followed all of the steps. I even tried putting the image generation on the lowest quality settings, but It just does not work.
I Have the same problem, could you find a way to fix it?
@@aleks378 i found out that you need an nvidia gpu for it to work
@@penguin395yeah i tried it again and it worked for me when started to generate it in online browser, but thank you for the quick ansewer
@@penguin395 that's odd, I have nvidia GTX 1050 gpu and still get the "error" message
@@ryu10000 Maybe you need a GPU thats less than 8 years old.
Fooocus is runelady approved, therefore i approve (you need to keep the good lady happy!! :D cheers for the video seb!
Always! And thank you for always being there. You guys are the real mvp 😘🌟
Will give it a try. Will SD1.5 models also work?
Thanks for the tutorial, can you tell me how to get the dark mode?
Very nice!! But how do you access the WebUI remotely? I install it on my server at home but want to access it from work.
I have installed but it's ridiculously slow. I have 3060 12gb. Do you know how to fix it?
Does this do image to image? I've been wanted a simple option to change the style of my renders
Does it only support SDXL? It seems pretty simple! Might give it a try, thanks :)
Someone actually said it. "Stable Diffusion is much better than Midjourney!" I concur lol.
I love Midjourney t hough, it does some amazing stuff. You should do a workflow that trains images from midjourney that you like to a style to use on SD. :)
you can't make nsfw stuff with midjourney, kek
YOu can do it accidently lol@@funfunnier3037
lol noob
Hey Sebastian, i did all the steps here but when i go to use it, it gives me a connection error for some reason.
Someone know a away to share the same models folder between this, a1111 and comfyui? like a1111 and comfy already did?
Does anyone know why input image doesn't do anything for me, and how can i fix that? Did anyone have a similar problem?
Hey Sebastian, what GPU are you using? Or maybe you can tell me how much VRAM you have? I'm thinking about buying a new one, and I noticed that yours is impressively fast
Hey hey, I use the RTX 3080 10gb
Good video
Very nice
subbed! ty sir!
Welcome!
With so many UI's pushing SDXL is a 6gb GPU still viable? I've been sitting pretty comfortable with A1111 and 1.5 since SDLX doesnt have all the features yet, but I'm unsure how long my home set up is going to last at this rate.
You can run SDXL on Compfy reasonably well even with 4G of VRAM. If you want quicker results colab or similar may be more cost efficient than upgrading your hardware
can you try a custom checkpoint? im trying to use one but is doest start the work. thanks
Hello. Thank you for this video. I am using Windows 10, AMD R9 M380 and Fooocus keep asking for ( NVidia). So is there is a solution for this situation? Best regards.
How do you install more models ? id like more realism in mine but not sure how to install realism and i cant find anything on youtube
can you share the file of results that come out from different steps in sampling method
your teaching is so good that I almost forgive your dad jokes
Does it have limitations, like for me it will not create whole body image or a tattoo with expitives?
Does it provide img2img, inpaint, negative prompts etc?
I'm not sure what the point of generating on your own computer is, if there's no image-to-image or controlnet.
How do we change the output directory of the images and where does this app look for LORA files?
Find out the location for the lora files?
For some reason, I which selections I’ve made they don’t show up like they do on your video. On your video they show up as blue checks. Mine doesn’t show up at all. Can you help?
Stable Diffusion Rocks
I can’t get it to work. When launched and trying to generate anything the message connection errored out shows up. I closed this and tried launching it through webui but then 2 modules are missing from the code: “gradio :1” and “comfy” so I am just stuck in the code page without the software working… anyone with the same error or somewhat similar that knows how to fix this?
o yea gonna generate all kinds of ai art now
Installing looks so easy now. I remember tearing my hair out for days trying to get the original SD installed and working when it was first launched. My PC struggles with A1111 to run SDXL, locks up and freezes a lot so I might give fooocus a try.
All these easy tools are going to put us tutorial TH-camrs out of work! 😂
@@sebastiankamphdon’t let AI replace you bro!
do you have a recommendation for AI 3 or 2d character designs to animate, or even better able to animate? Either route to interact with live action character within photogrammetry “real photography” spaces using compositing.. …
Any part of this scenario done in AI, even fake live action or fake photogrammetry, im curious about. Ultimately want to experiment with AI see what it can/ cant do aside from repurposing still photos and paintings it learned from online.
It is still the same issue for me on Linux at least with Fooocus(needs the conda junk). The reason is quite simple. AI started in the lab and lab people are no actual programmers. So they use fake languages like Python (or Lua for games) that got the worst dependency management system of all languages on Earth. They are fake languages that do not compile before they are shipped but after. So essentially you are the fool who has to do it on your own computer after shipping so when you downloaded it. That is the reason why you can see the source code in their files.
That is why other applications are usually little to no hassle at all with getting them to run. They are made with real languages like C#, C, C++ or the best of them all called Rust.
Question: So do I have to have GB extension to run this? Such as a NVIDIA RTX 3070? Or does my standard Windows laptop do the trick? Because once I download Fooocus, I cant seem to access it. An answer from anyone would be greatly appreciated.
You need a dedicated gpu, yes. If you don't have that, you can use cloud solutions like ThinkDiffusion
Im having 4GB graphic memory with 3050 AMD card but still it is not working in my laptop as it say’s not enough memory.
is there a similar thing for MAC (intel) ?
Everytime again when I go to the Previously used fooocus url. It doesn't work and says reload.
So I have to go to the file manager and again install the same file to run it.
This happens everytime.
Any solution?
how to put models inside checkpoints?
it does NOT run automatically after install; do you have to install it again and again each time you want to access the app??? Why does no one tell how to run this after the initial install??
I installed it, the speed is faster than A1111 but slower than comfyUI. Also you cannot choose how many steps you need for generation. Only 30steps or 60steps. I think it still needs a lot of improvements. The one who created this software is the one who made controlnet so it still have potential. 6/10
liked and subscribed 😲
Happy to hear it, welcome aboard!
cool
Hello mate. Can you please describe how can we change model in foooocus ? Suppose i like some model from tensor art and want similar there ( without faceswap ) how can i do it ?
Download and put in models checkpoints folder
Are there any Mac installation instructions?
can you do the tutorial for mac book ?
Young John Malkovich
I downloaded and extracted but the UI didn't open up when it was done with the command prompt
Did i hear you right when you said at least 4gbs of vram is recommended? I have 6 (laptop gpu) and struggle to run SDXL at all
I only have 4 GB on my laptop, add this in your run file: --normalvram
@@robertcurrie1026 how??
Can we uae another models
Someone please help me im getting error that says
RuntimeError: Failed to import transformers.models.clip because of the following error (look up to see its traceback):
DLL load failed while importing _sentencepiece: The specified module could not be found.
After a1111 this seems like generative-images-for-dummies
Agreed - I'm used to A1111 and this seems overly simplefied. I can see how I might suggest it to someone who just wants to try a few prompts to start out with, but it's not about to get me to switch over.
It's a huge difference for sure. But there still are A LOT of users that have never tested generative AI. I love all the options we're getting! 😊
doesnt work for amd gpu users :(
mine was running perfecly and then all the sudden it stop and its asking for GPU nvidia, I have already updated the drivers and still the same helppp.
Oh! oh! I think I know the developer. He's a very handsome gentleman!
Thanks for sharing! I see it downloads the "0.9vae" models rather than the ones without "0.9vae", I know it doesn't mean "beta" or SDXL 0.9, and the release date is newer, but, does anyone know the differences?
People have found that the sdxl 1.0 model works pretty well with the 0.9 vae. So it's all in order.
The 1.0 VAE had a problem, it caused something like a watermark on the images, a thin green border around the images, and the 0.9 VAE did not have that, so the most used is that now.
Thanks for your reply! @@sebastiankamph
Thanks for your reply! @@ArtificialSinger
When i try to generate 2 person in same picture for example, mom carries her baby on hip. It creates a mutated photo, is there any solution for that?
Do we need to install Python somehow beforehand? I am trying to press run but I am getting an error FileNotFound related to Outputs and config file.
Same!
🔥🔥🔥
after submitting the prompt it gives connection time out error ,plss help
Hello! Could you help me please? I have an issue with running Fooocus. Its says RuntimeError: Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU and installed a driver. I have amd gpu.
What is the difference with iLLyasviel version of Fooocus ?
Noob question. How can I set the interface to dark mode?
How do you get it in dark mode?
Just use a browser in darkmode
it just keep giving me error.
RuntimeError: Could not allocate tensor with 52428800 bytes. There is not enough GPU video memory available!
GPU AMD RX580 8GB
RAM 24GB
I am totally unable to install this. I downloaded Fooocus. I unpacked it. I double-clicked on the run.bat file, and eventually it said the system path wasn't found - hit any key to continue. It closed the DOS window, and I'm stuck here. No models downloaded during any of this process. I initially unpacked it to my Programs folder. I then tried installing it as-is in the downloads folder. Neither way worked. I also tried unpacking/running the python_embedded file.
Further, I went to the Fooocus Github page and registered an account to get help, but the owner doesn't allow new users to post. So how am I supposed to get help installing it?
I'm at a loss.
Get RuinedFooocus instead. Runewolf is author on github
I'm sorry to hear about your photography career 😥
What about image to image
fooocus 2.0 is out ;) possible to cumulate styles ... yep
so its only running on nvidia cards?
I tried fooocus, it's incredibly slow compared to my SD 1.5. Plus the SD ecosystem, I have to give up.
thanks for the video, wont run for me as it seems to not work well with amd
help it says nvidia drivers not found
I got error when unzip files. to many phyton error I got. dunno why
how do you change Fooocus at to dark mode? thanks
Based on your browser settings
My pc got stuck after launching
I use rtx 2050 4Gb
foocus generate ou have nsfw blocked?
I'm testing Fooocus, but it's melting my computer (32GB RAM, Ryzen 9 5900, Radeon 6800XT). Temperatures are off the charts, I can boil water on my GPU. There were a few times when the computer just shut down and I wondered if it would restart.
Any ideas on how I can slow down Fooocus to render a little longer but not burn my computer to the ground? :)
I'm not sure if you can change that in Fooocus. However you could undervolt your gpu through afterburner.
@@sebastiankamph Thank you for responce :). Yeah, I thought so, that I would have to go 'outside' of Fooocus to create some thresholds for it.
Btw: love you channel :)
If your joke wasn't about focus that would have been a missed opportunity lol xD
As long as face repair is turned on, an error will be reported:“TypeError: 'NoneType' object is not subscriptable”
Ele consome toda a memória RAM do meu PC! Inutilizável! Dizem que é mais leve que os outros, mas não tem como usar! Parece que não é verdade! A coisa certa é consumir a minha GPU que é muito boa e não toda a minha memória RAM. Só de abrir o link já usa 16gb de RAM a ponto de congelar todo o pc. Eu não sei se é um bug ou se isso é normal, mas é assustador, eu uso SDXL e Automatic 1111 sem problemas, eu realmente queria testar isso.
i cant match the requirements, any other way to do it
thinkdiffusion.com/
Is it available for Android
Would this use less vram than comfyui
as far as I can tell it is a bit slower on a low vram set up such as mine (4GB vRAM), its a bit slow but the results are imprerssive
so it doesnt work it with amd?
Did you install Nvidia Game Ready Driver or Studio Driver?
I think I'm currently on the studio driver. But game driver has worked for me as well.
@@sebastiankamph Thank you for answer my question. I have switched from "Game Ready Driver" to "Studio Driver." I wish you all the best, good health, and the creation of new video clips.
I apologize if I said something wrong because my English is quite poor, Chat GPT Translate
So I am currently running Stable.D through foocus this same way, do you know if it is possible to run extensions within the application? I'm not seeing settings for this like how the standard Stable.D interface has. Example video showing someone installing extensions and setting them up in the standard Stable.D set up: th-cam.com/video/CuJV8hU6_O4/w-d-xo.html
i have not nvidia, how can i use it?
Check out online providers like ThinkDiffusion
no controlnet?
No Controlnet for Fooocus no. At least not at this time. Use another UI for that.
It says I have low vram and doesn't load
Then you need to upgrade your gpu