Install ComfyUI and run your 1st generation - Stable Diffusion for Professional Creatives
Ever wondered how to use Stable Diffusion in your traditional workflows in professional environments as a creative?
Stable Diffusion for Professional Creatives is a series of videos about how we can apply generative AI in production environments, with real life examples and guides.
If you like my contributions, you can buy me a coffee here: ko-fi.com/risunobushi
In this introductory video, we'll learn how to download and run Stable Diffusion on our hardware, and we will run our first text2image and image2image generations.
Useful links from the video, by category.
Python:
- Download Python 3.11.5: www.python.org/downloads/rele...
- How to install Python on Windows: • How to Install Python ...
- How to install Python on macOS: • How to Install Python ...
- How to install Python on Linux (Ubuntu): • How to install Python ...
ComfyUI:
- Download ComfyUI: github.com/comfyanonymous/Com...
- Download ComfyUI Manager: github.com/ltdrdata/ComfyUI-M...
- Civitai: civitai.com/
- Download epicRealism: civitai.com/models/25694?mode...
- Download VAE: huggingface.co/stabilityai/sd...
Other hardware / OS configurations:
- Install comfyUI on macOS: • Install ComfyUI on Mac...
- Install comfyUI on windows with an AMD card: • AMD GPU + Windows + Co...
Good all-around models (SD 1.5):
- Realistic Vision: civitai.com/models/4201/reali...
- Dreamshaper: civitai.com/models/4384/dream...
Good all-around models (SDXL):
- Juggernaut XL: civitai.com/models/133005/jug...
- Dreamshaper XL: civitai.com/models/112902/dre...
- Animagine XL: civitai.com/models/260267/ani...
Timestamps:
00:00 - Intro
00:42 - What is Stable Diffusion
03:19 - What's needed
05:18 - Download and install Python
05:57 - Download and install comfyUI
08:02 - Download and install comfyUI Manager
08:47 - Download a model
10:16 - Download a VAE
10:55 - Launching comfyUI
11:44 - Setting up your first text to image generation
21:51 - Setting up your first image to image generation
25:54 - Outro
#stablediffusion #stablediffusiontutorial #ai #generativeai #generativeart #comfyui #comfyuitutorial #risunobushi_ai #risunobushi #andreabaioni
Пікірлер: 40
I feel like you are a golden source in the area. I have used focus and automatic 1111 but I wanna start using comfy uı and start learning with you. 🎉
@risunobushi_ai
12 сағат бұрын
thank you for the super kind words!
Great tutorial for those of us who are starting out so far, thank you for making this type of content, I hope you continue creating these explanatory videos for us beginners. Thanks!!!!!!
Thanks! The best presentation of the material from all that I have seen. Without further ado. Everything is specifically step-by-step. A really professional approach.
@risunobushi_ai
2 ай бұрын
Thank you for the kind words!
Great! Finally someone who can explain this clearly. Thank you!!!
Excellent video and very much looking forward to the whole series!! Thanks!!
Hi! I'm really happy too see someone who talks through the idea and then starts from scratch. And it only helps that you're using it for professional purpose where control is important. You have my subscription :) Keep up the awesome work!
@risunobushi_ai
Ай бұрын
Thank you for the kind words!
Спасибо за видео, начал понимать как работать с ComfyUI
Great material, perfect pace, everything clearly explained.Thank You!!!
@risunobushi_ai
28 күн бұрын
Thank you for the kind words!
Clear explanation & Nice job
This is the channel I needed, Thank you, finally vae and latent image makes lil bit of sense to me
@risunobushi_ai
2 ай бұрын
Happy to have been helpful! That's precisely what I'm trying to do with this channel.
i'll not be able to download Comfyui it is 1.3 GB, my daily data is 1. GB lol. Anyway video was great i was searching for a long time and I found it with a full explanation. u learn my subs
Thanks for this.!!
Perfect, thank you. Just tried today and the whole initial node set up is loaded when I first started ComfyUI. I ¨m finally able to generate my images so I'm very thankful for this.
@risunobushi_ai
2 ай бұрын
Happy to have been helpful! ComfyUI is very intimidating at first, but I think that if it's explained properly it can be a great tool
Thanks a lot. Installation made on Mac and results are outstanding
@risunobushi_ai
2 ай бұрын
I’m glad you made it through the installation at least! Are you using the same exact models as I am? Edit: I saw your edit, great that it’s working perfectly!
@regiszapp8758
2 ай бұрын
@@risunobushi_ai yep all good, I will now need to install few more LDM things I believe ...Lora etc...
"How many specifications are required for running Smooth ComfyUI with VRAM?"
@risunobushi_ai
16 күн бұрын
I've seen some people get away with 4gb VRAM for simple workflows, but I wouldn't recommend it. You can use it in low VRAM mode to squeeze some more out of it.
Hey brother, in my case the preview image node is not working, it does not show any output, although when I use save image node the images are saved in the output folder of comfyUI. Can you please help me with this? Edit: Preview method in manager is set to "none" (just like yours in the video)
@risunobushi_ai
27 күн бұрын
That's weird, I've never seen that happen before. Preview method shouldn't have any impact on this, as it's just how the preview image is rendered (none is a faster one that bypasses some bells and whistles the other types have, but it's still a preview type). What OS are you on? Have you got the (iirc) impact pack installed? if so, could you test the preview bridge node and check if you have the same issue with that type of preview node as well?
@pranavahuja1796
27 күн бұрын
@@risunobushi_ai Yes I have impact pack installed and same issue is faced in preview bridge node as well. I am using windows 11, 16GB ram, gtx 1650, 8gb vram
@pranavahuja1796
27 күн бұрын
I am trying figure out whether it can add some value to my workflow, I will then deploy it to some cloud service, if you think my pc specs are not good enough can u please suggest me some cloud service
@risunobushi_ai
27 күн бұрын
no, I was just trying to figure out if you had some non standard hardware / OS config. I'm sorry I can't be of much help, I can't replicate the issue and honestly even by googling the symptoms I can't find a single similar issue :/
is there any other process
@risunobushi_ai
2 ай бұрын
You can run it on some cloud services, like RunDiffusion, OpenArt, and more, but each one has a different pricing / credits structure
I used same prompt and same setting then why my model face is giving me ugly , i have wrote also in negative prompt " ugly face " then face is making more bad
@risunobushi_ai
2 ай бұрын
Are you using the same models as I’m using? Another user had the same issue and they weren’t using the same models, that solved it.
@mohammadbishalahmed8784
2 ай бұрын
@@risunobushi_ai yes i used same model from your link civitai
@risunobushi_ai
2 ай бұрын
Did you build the workflow from scratch or did you import my workflow from the Json file? (I’m asking so I can help you troubleshoot the issue)
@mohammadbishalahmed8784
Ай бұрын
@@risunobushi_ai i am going to mail you
... can all of this be done on a Mac -_- ? .. or just PC
@risunobushi_ai
2 ай бұрын
You can install ComfyUI (or most other UIs) on macOS, or you could run workflows on the cloud (there’s a ton of services for that). On a Mac it’s just a bit more time consuming when setting everything up, some features may not be available (anything related to CUDA for example, for which you need a Nvidia GPU), and generally way slower.
@adrienbiosseduplan9666
2 ай бұрын
@@risunobushi_ai as I thought. Thanks for this though
@risunobushi_ai
2 ай бұрын
@@adrienbiosseduplan9666 If I were to give my honest advice to macOS users, it'd be to test out locally where possible, and if they feel like generative AI is something that adds enough value to justify a dedicated workstation clients-wise, then go for it only once it's been established that it makes sense to do so. Unfortunately CUDA has no competition in the space yet, with no end to its reign in sight, so both macOS and AMD users are kinda in a weird spot right now.