ModelsOther

kohya_ss |【HOW TO】MAKE A LORA

kohya_ss |【HOW TO】MAKE A LORA

【HOW TO】MAKE A LORA
https://www.pixiv.net/artworks/107299229

You only need these 3 files , then you can make lora with kohya_ss_ui

  1. 01.install_sd_scripts_v5.bat

  2. 02.install_kohya_ss_ui_v4.bat

  3. 03.Xformers+troch.rar
    https://drive.google.com/file/d/1s18QvfK7ehJbTW7vWxDXThMi3KJM-G_T/view
    Password
    : 大叔愛蘿莉

I learn “how to make a lora” from this youtube video
https://youtu.be/089lLR46KCQ”

We edited his video and make a translation :
You need to choose [cc] Auto translate from chinese
https://youtu.be/AgajKybahHQ

Please modify it to your own path
At least 15GB must be reserved for the installation file

And This is the link for cutting those picture
https://www.birme.net/?target_width=512&target_height=512

The configuration file is here: [ .json ]
https://drive.google.com/drive/folders/119L4jM5BJ9_Z9a8jcUS79xAy5dHpeeCB?usp=share_link

I will still updated some imformation right here
https://www.facebook.com/media/set/?set=a.1648317618983237&type=3

But , In fact, my mother language is Chinese

中文的話我可以答你一些問題

Need Pictures?
15-20 pictures
better include, Full_body, Upper_body, Full_face, Right_face, Left_face, …
Outfit : shirt, skirt, …

In the “Stable_Diffusion” in “Training”
in “image preprocessing”
You can click : ” deepbooru”
for getting prompt.txt file

Output to named “/nameABC/100_EAV P” folder
(nameABC means as you want to make lora name)

then go to active “kohya_ss/gui.bat”

In the “Dreambooth”
Configuration file
to load “JLoraBasicSettings設定檔.json”

This is our Configuration file

and then
“Image folder” “/nameABC/”
“Output folder” “/nameABC/”
“Model output name” “nameABC_V1”

“Dreambooth Lora” in “Training parameters”
Make sure “LoRA type” “Kohya LoCon”

and you can change the “Network Rank (Dimension)”
to control the file size
20-128
lower then 20 is too low

Related Model

ControlNet Inpaint Example for ComfyUI
ControlNet Inpaint Example for ComfyUI
Negative Embedding for Realistic Vision v2.0
Negative Embedding for Realistic Vision v2.0
Tiefsee Image Viewer (Prompt Reader)
Tiefsee Image Viewer (Prompt Reader)
Lighting based picture control controlnet
Lighting based picture control controlnet
marduk191’s AnimagineXL 3 Workflow
marduk191's AnimagineXL 3 Workflow
ReVisionXL with Additive Prompting
ReVisionXL with Additive Prompting

No comments

No comments...