Тёмный

1 Your Ultimate ComfyUI Toolkit: A Comprehensive Guide to Stable Diffusion for Absolute Beginners 

Arch_AI_3D
Подписаться 4 тыс.
Просмотров 1,9 тыс.
50% 1

Playlist: • ComfyUI Ultimate Guide...
Patreon: / archai3d
Join me for the most in-depth beginner's guide to ComfyUI and Stable Diffusion on RU-vid! This 2-hour video is designed to take you from zero to hero, covering every aspect of the software you need to know. We'll explore:
************
00:00:00 How to install ComfyUI
00:01:09 Install ComfyUI Manager extension
00:02:25 download checkpoint
00:03:34 Civitai.com find the latest checkpoint and update
00:04:40 Check the recommended setting for each checkpoint
00:05:21 Move to the models folder
00:05:33 Add extra folder path for loading models inside ComfyUI, use models from Automatic1111 or others
00:07:02 my extra_model_paths.yaml after configuring
00:08:08 First Run ComfyUI
00:09:25 Save image node setting
00:10:08 active preview
00:10:32 Check Vram usage for preview
00:12:17 Underclock GPU and keep it cool
00:13:07 How does Stable Diffusion work
00:17:10 Inside of stable diffusion
00:17:52 convolution 3x3
00:19:23 Unet diagram
00:20:05 inside Unet
00:26:04 in background of ComfyUI
00:32:24 Difference between samplers
00:38:40 image to image workflow
00:40:00 denoise effect
00:44:16 Prompt in image-to-image
00:50:49 What is the steps in Ksampler
00:53:47 What is CFG
00:55:57 VAE Encode and Decode
00:57:54 ComfyUI nodes setting
00:59:58 How to copy nodes
01:00:33 Convert to input
01:00:57 Primitive node and data format, Int, float, string, combo, boolean
01:06:35 Load workflow from saved PNG files
01:07:05 How to move a selection of nodes
01:07:51 Grid size and snap
01:08:18 UI setting
01:08:28 default setting for double-click
01:09:03 Save and load Workflow in json format
01:09:19 UI setting
01:09:56 Change line drawing type in ComfyUI
01:10:35 clipspace
01:11:09 Mask Editor
01:12:17 Check CMD window to see the rendering time
01:13:57 What is batch size
01:17:03 What is batch count
01:18:11 Queue setting
01:21:24 Find Converge sampler, a trick to test and compare strings inside ComfyUI
01:27:34 Mute nodes
01:29:53 bypass nodes
01:31:05 reroute node
01:31:25 Adjust prompt attention
01:32:50 Negative prompt compare
01:34:08 Embedding, how to use
01:35:39 What is LORA
01:36:39 How to find the best LORA models
01:37:48 difference in LORA
01:38:15 Load LORA in ComfyUI
01:39:44 LORA trigger prompt
01:40:35 LORA for contrast
01:42:00 LORA for line art drawing
01:45:02 What is HyperSD
01:48:31 How to install missing nodes in ComfyUI
01:49:21 1 STEP, Generate an image on the fly
01:50:49 HyperSD, test CFG, and negative prompt
01:52:21 HyperSD implementation as LORA
01:54:46 1 Step, LORA
01:57:07 HyperSD Best implementation
01:57:55 Real-time painting by using HyperSD
***********************************
The fundamentals of Stable Diffusion: What it is, how it works, and why it's revolutionizing AI art.
Getting started with ComfyUI: Installation, setup, and navigating the interface.
Essential settings: Understanding the key parameters and how to adjust them for different results.
Deep dive into parameters: A closer look at advanced settings, explained in simple terms. This includes an exploration of ComfyUI's node-based architecture and how it interacts with Stable Diffusion.
Understanding the core concepts: We'll break down samplers, schedulers, steps, denoising, and other essential terms in a way that's easy to understand.
Tips and tricks: My personal insights and best practices for getting the most out of ComfyUI.
This is the only beginner's guide to ComfyUI you'll ever need. Whether you're an aspiring artist, a curious tech enthusiast, or just someone who wants to have fun with AI, this video is your ultimate starting point.
--------------------------------
URL
about Stable diffusion: jalammar.github.io/illustrate...
poloclub.github.io/diffusion-...
Git: git-scm.com/downloads
7Zip: www.7-zip.org/download.html
ComfyUI: github.com/comfyanonymous/Com...
ComfyUI Manager: github.com/ltdrdata/ComfyUI-M...
Civitai.com
HyperSD: huggingface.co/ByteDance/Hype...

Опубликовано:

 

27 июн 2024

Поделиться:

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист
Посмотреть позже
Комментарии : 16   
@captainpike3490
@captainpike3490 Месяц назад
Thank you for the great tutorial. You have excellent teaching skills presented in a calm way. I have learned a lot of tips that no one has shown before. Most of the ComfyUI tutors race to show us how complicated a workflow can be. When I watch them, I always have the same feeling that I should stay with Fooocus because ComfyUI is too complicated. You showed me that things can work in a simple way. I will definitely come back and watch some parts again. You might consider creating some short videos from this tutorial. Thank you again.
@ArchAi3D
@ArchAi3D Месяц назад
thank you
@user-ef4df8xp8p
@user-ef4df8xp8p Месяц назад
Thank you.....Good work...This is a very detailed video. Very well done...
@ArchAi3D
@ArchAi3D Месяц назад
Welcome
@seifalislam3645
@seifalislam3645 День назад
Amazing guide, how can I access the 3D interactive map in 20:17?
@cemilhaci2
@cemilhaci2 Месяц назад
Perfect
@ArchAi3D
@ArchAi3D Месяц назад
Thank you
@dadekennedy9712
@dadekennedy9712 Месяц назад
This is a recommend video to watch.
@ArchAi3D
@ArchAi3D Месяц назад
Thanks 🙏
@notionto4638
@notionto4638 Месяц назад
Thanks could you share how can access the 3d environment about explaining stable diffusion like in the video?
@ArchAi3D
@ArchAi3D Месяц назад
also I am interested :)
@HarborSchool-ey6gt
@HarborSchool-ey6gt 22 дня назад
@@ArchAi3D Is it not available to access to the 3D space like you show in 17:11?
@ArchAi3D
@ArchAi3D 22 дня назад
@@HarborSchool-ey6gt I made it in 3DS Max for this video, if you need the 3D model send me email: amir84ferdos@gmail.com
@ArchAi3D
@ArchAi3D 22 дня назад
I made it in 3DS Max for this video, if you need the 3D model send me email: amir84ferdos@gmail.com
@mehradbayat9665
@mehradbayat9665 27 дней назад
How can I access the resource you are showing at 18:00?
@ArchAi3D
@ArchAi3D 27 дней назад
Please check this: setosa.io/ev/image-kernels/#:~:text=Below%2C%20for%20each%203x3%20block%20of%20pixels%20in%20the%20image%20on%20the%20left%2C%20we%20multiply%20each%20pixel%20by%20the%20corresponding%20entry%20of%20the%20kernel%20and%20then%20take%20the%20sum.%20That%20sum%20becomes%20a%20new%20pixel%20in%20the%20image%20on%20the%20right.%20Hover%20over%20a%20pixel%20on%20either%20image%20to%20see%20how%20its%20value%20is%20computed.
Далее
Спасибо Анджилишка, попил😂
00:19
Monster shops mini mart 👻🛒 #shorts
00:49
Просмотров 11 млн
Dissecting SD3
20:25
Просмотров 13 тыс.
The U-Net (actually) explained in 10 minutes
10:31
Просмотров 83 тыс.
Coding Stable Diffusion from scratch in PyTorch
5:03:32