SYSTEM NOTICE

Auto translation by AI. Be sure, accuracy, nuances and authorial intent may not be fully reflected.
見出し画像

[For AI Beginners] Just unzip the ComfyUI Portable version on your PC! A manga-style guide to installing the Windows version. What is ComfyUI? Includes NSFW workflows too.

手久野 愛のAIちゃんネル
This article covers "1-1. Running in a local environment" and "2-1. Running in a local environment"

For AI Beginners | What is ComfyUI?

ComfyUI is a tool for AI image generation that allows you to control the process by connecting blocks called "nodes" .

Simply put,

It is "software that lets you visually build and customize the flow of image generation"
*Currently, it is becoming one of the mainstream environments for AI image and video generation.

.

For example,

  • loading a model

  • writing a prompt

  • generating an image

  • upscaling

  • fixing faces

You can perform these processes (called nodes) by connecting them with lines.

For AI Beginners | What is ComfyUI?

Features of ComfyUI

◎ High performance yet free

Because it runs on your own PC, you can use it without monthly fees or usage limits.

◎ Fully customizable

By simply adding nodes,

  • video generation

  • Face swap

  • ControlNet

  • LoRA

  • Lip sync

  • Animation generation

  • NSFW generation

and more, you can expand various functions.

◎ Visual and easy to understand

Since you can see the "processing flow" on the screen,

it is also easy to understand

"what is happening and where".

Click here for PCs without a GPU

Differences from Stable Diffusion WebUI

Stable Diffusion WebUI is "button-operation centered," while ComfyUI is a "freely customizable type."

It might look a bit difficult at first, but once you get used to it,

"your own custom AI image generation environment"

you will be able to build.

Recommended for people like this

  • Those who want to do AI image generation seriously

  • Those who want to use it for free and without limits

  • Want to try video generation and the latest AI features

  • Want to customize it to your liking

  • Want to create high-quality images

The portable version is recommended first

For beginners, we recommend starting with

"ComfyUI Portable version"

as it is easy to install.

Just unzip and launch, so you can use it without having to build a Python environment yourself.

ComfyUI Portable version (Windows) easy installation!

"I want to start using image generation AI, but the installation seems difficult..."
For beginners like you, this time I will explain how to introduce the ComfyUI Portable version (Windows) as clearly as possible.

The biggest feature of the Portable version is that it does not require difficult Python environment setup, and you can use it just by downloading and unzipping it.

Furthermore, since it runs on your own PC, it has the advantages of:

  • no monthly fees

  • no usage limits

  • can be used offline

.

What is the ComfyUI Portable version?

The ComfyUI Portable version is a package for Windows that includes the ComfyUI core and the necessary Python environment from the start.

In a normal AI environment setup,

  • installing Python

  • CUDA settings

  • Library installation

and so on are required, but the Portable version automates almost all of that for you.

It is easy for beginners to set up which is its appeal.
*Manual installation can lead to version hell...

1. Download the ComfyUI Portable version

First, download the ComfyUI Portable version.

Choose the one that matches your PC environment.

Main types

Basically, if you are using a GeForce RTX series card, you can just choose the
Nvidia GPU version.

2. Unzip the downloaded file

Extract the downloaded 7z file using
unzipping software like 7-ZIP.

Once unzipped, the folder will contain the following files.

  • ComfyUI -> Main program

  • python_embeded -> Python environment

  • update -> Update files

  • run_nvidia_gpu.bat -> For launching with Nvidia GPU

  • run_cpu.bat -> For launching with CPU

The portable version contains almost everything you need inside.

3. Launching ComfyUI

Next, launch ComfyUI.

Double-click one of the following files according to your PC environment.

For those using an Nvidia GPU

run_nvidia_gpu.bat

For those using only a CPU

run_cpu.bat

Most gaming PCs and PCs equipped with an RTX card
will use run_nvidia_gpu.bat.

4. A black screen appears

When you launch it, a black screen (command prompt) will open.
*The first launch may take several tens of minutes due to settings and additional model downloads. Let's wait patiently while doing some exercise with a stepper.

If the following appears there, it is a success.

To see the GUI go to: http://127.0.0.1:8188

This means,

"ComfyUI has started successfully"

.

You might be surprised by the black screen at first, but please rest assured that it is not an error.

5. Open ComfyUI in your browser

Normally, the browser will open automatically and the ComfyUI screen will be displayed.

If it does not open automatically, please enter the following address into your browser's URL bar.

http://127.0.0.1:8188

Then, the ComfyUI screen will be displayed.

Advantages of the Portable version

The Portable version is highly recommended for beginners.

Advantages

  • Easy to install

  • No Python configuration required

  • Ready to use just by unzipping

  • Easy to delete if issues occur

  • Easy to separate multiple environments

It is especially ideal for those who just want to try out AI.

Summary

The ComfyUI Portable version is a convenient environment for beginners that can be used just by:

  • Downloading

  • Unzipping

  • Launching the bat file

It might feel difficult at first, but it is surprisingly easy once you actually try it.

It might feel difficult at first, but it is surprisingly easy once you actually try it.

If you want to freely enjoy AI image generation in a local environment, please give it a try.

How to install ComfyUI Manager (for beginners)

To use ComfyUI conveniently, the first thing you should install is
ComfyUI Manager. (It is essential.)

Once you install ComfyUI Manager,

  • Adding custom nodes

  • Updating

  • Automatic installation of missing nodes

and more can be done easily from the screen.

Installation Method

1. Open the ComfyUI folder

First, open the folder where you unzipped ComfyUI.

Inside it, please open the

ComfyUI/custom_nodes

folder.

2. Right-click in the custom_nodes folder

Right-click inside the custom_nodes folder and select

  • "Open in Terminal"

  • "Open PowerShell"

  • "Open in Terminal"

or similar options.

3. Paste the following command

git clone https://github.com/ltdrdata/ComfyUI-Manager.git

Once pasted, press the Enter key.

4. Restart ComfyUI

When the installation is finished, close ComfyUI and restart it.

Then, on the screen

  • Manager

  • Custom Nodes Manager

and other menus will be added.

The installation is now complete.

If Git is not installed

If 

git は認識されていません

is displayed, you need to install Git first.

In that case, install Git for Windows and then run the command again.

Benefits of installing ComfyUI Manager

Once you install Manager,

  • add nodes with one click

  • automatically install missing nodes

  • manage updates

and more, so it is very convenient for beginners.

If you are using ComfyUI, it is recommended to install it first.
*It is at an essential level.

When you launch ComfyUI, frames are displayed from the start

When you launch ComfyUI for the first time, it will be displayed with several frames (nodes) already placed.

This is a basic sample for image generation, and includes

  • model loading

  • prompt input

  • Image generation

  • Save

and other processes are pre-configured from the start.

However, since there are no AI models included at first, the "Load Checkpoint" section may appear in red.

This is not a malfunction; it just means "no models have been added yet," so please don't worry.

Once you add the models, it will work normally.

What is a workflow?

A workflow is

  • which model to use,

  • what order to process in,

  • how to generate the image,

and other such blueprints for the process.

In ComfyUI, you build processes by connecting nodes with lines.

What is a model?

A model is the actual AI engine that generates images and videos.

For example,

  • anime-style images,

  • photorealistic images,

  • video generation,

  • character generation

and so on, the model you use changes depending on your purpose.

ComfyUI is like an "AI production studio."

Think of it like this:

  • ComfyUI → Workbench

  • Workflow → Blueprint

  • Model → AI Engine

That is the relationship.

In other words, only after adding workflows and models to ComfyUI can you start generating AI images or AI videos.

First, let's try adding one workflow

For beginners, don't overthink it from the start;

  • Image generation workflow

  • Video generation workflow

I recommend adding one and giving it a try.

Learning by actually using it makes the ComfyUI mechanism much easier to understand.

◎ Recommended image generation workflows

◎ Recommended video generation workflows

いいなと思ったら応援しよう！