BASIC - All you need to know about Stable Diffusion (Mac & Windows)
ฝัง
- เผยแพร่เมื่อ 6 ก.ย. 2024
- 👨💻 This video covers the basics of Stable Diffusion, its unique interfaces like A1111, InvokeAI, and ComfyUI, to solutions for Mac compatibility issues. Plus, information on how to run it remotely using cloud-based platforms (such as Google Colab and RunPod) and cloud-based programs (such as RunDiffusion, DiffusionHub, and ThinkDiffusion).
____________________________________________________________________
🔗 Stay Connected:
[Discord] / discord
[Patreon] patreon.com/In...
[Instagram] / lacarnevali
[TikTok] www.tiktok.com...
____________________________________________________________________
🤙🏻 Learn More:
/ membership
/ lauracarnevali
____________________________________________________________________
📌 Links:
A1111: github.com/AUT...
InvokeAI: www.invoke.com
ComfyUI: github.com/com...
Colab: colab.research...
RunPod: www.runpod.io/...
PaperSpace: console.papers...
RunDiffusion: rundiffusion.com
DiffusionHub: diffusionhub.io
ThinkDiffusion: thinkdiffusion.us
____________________________________________________________________
🔥 Highlights:
00:17 What is Stable Diffusion in 30 secs
00:38 A1111
01:13 InvokeAI
01:27 ComfyUI
02:15 Stable Diffusion: PC requirements (Mac limitations)
03:18 Solution to Mac limitation: cloud-based platforms
03:45 Colab
04:56 RunPod
05:32 Cloud-based programs
05:55 RunDiffusion: pricing
06:34 DiffusionHub: pricing (Referral Code: LAURA20)
07:18 Thinkdiffusion: pricing
#AIart #StableDiffusion #MacUsers #DigitalArt #CreativeTech #AIonMac
I am obsessed with your accent. It's the cutest accent I've ever heard
Try italian. It´s cuter.
I've been using Stable Diffusion locally for a while. I'm actually building a better computer right now specifically for AI. I'll use it for large language models and Stable Diffusion.
Too expensive unfortunately
My favorite image generation TH-camr is back. Great. Here is my wishlist for the next videos: InstantID & FaceID, consistency of characters, implementation of specific objects into the image, training a style with kohya, another video on what's behind the numbers in koyha. Thanks. Not a demand... Maybe an inspiration. You are great
Laura, you are an expert. Would you not argue that with your expertise, i would make sense to switch to ComfyUI?
I personally find it hard to switch from A1111 to Comfy, feels like starting from the beginning. But I am not as good as you. What is your opinion, if someone wants to be really good, should one change to Comfy?
Thanks so much for all the valuable information always!
Laura, great video as always. However, I noticed that you didn't mention options for Linux users. I personally only use Linux, but my AMD 5xx graphic card is not supported for IA acceleration. Regardless, Linux is the best operating system for developing or running AI applications. These codes usually run faster on Linux than on Windows, especially on computers with AMD graphics cards, although Nvidia is by far the best choice. As for Mac, even the most powerful ones are outperformed by a Linux or Windows computer with an Nvidia 4070 graphic card. The only advantage of recent Mac-ARM computers is their shared RAM between CPU, GPU, and ANE. This means you could buy a 128GB RAM Apple notebook to run large AI models or Stable Diffusion when requires a significant amount of RAM, but at a very high cost for a normal user, although it would be a better solution than an Nvidia GPU for enterprise.
True... on my Mac studio m1 max I can render huge pictures. But it takes forever.
I Just have to sort of disagree on one tiny bit of your video 😁
I run ComfyUI locally and it runs just fine* on 4GB VRAM. I'm using Linux (alternative to Windows/Mac)...so maybe that has something to do with it? But yeah ComfyUI is for geeks like me and so is Linux 😁
Great video, keep them coming!
*sometimes I have to scale things down and then upscale later.
💜💜💜
Hello, what is forge ui ???
How to run it through colab ???
Thank you for your time.
ForgeUI is just another interface, which should perform better than A1111 and ComfyUI. The interface is basically the same as A1111.
Yes, you can run it on Colab (change the runtime to GPU) and run this code:
!pip install pygit2==1.12.2
%cd /content
!git clone github.com/lllyasviel/stable-diffusion-webui-forge
%cd /content/stable-diffusion-webui-forge
!python launch.py --share
kudos: github.com/lllyasviel/stable-diffusion-webui-forge/discussions/189#discussioncomment-8886546
@@LaCarnevali Thank you for your time
@@LaCarnevali hello again, I really appreciate your response, but I'm about to tell you that it does'nt work, i don't know if the problem is the "free" Google colab notebook or what.
If you can share a Google colab notebook with us it will be great act from you.
Thank you again, I wish the best for you.
@@youcefamarache4801 does it initialise? It was working for me
@@LaCarnevali NO, AT ALL
Grazie per i tuoi video sempre chiarissimi!
👋