×
SeaArt AI Enterprise Version
article cover

Using LoRas to add changes to images

avatar
T
avatar_frame
TCO
Updated on Dec 5, 2024
36
#Articles
#Generation guide
#Feature

When generating an image with SeaArt, there are many options for adjusting the image output including adding one or multiple LoRas. LoRas provide the AI additional image and text information, and your image outputs can change drastically depending on the LoRas you use.

For this guide, here is the example LoRa:  https://www.seaart.ai/models/detail/ct6cg85e878c73caqln0

This LoRa was trained on ***paint face paint as well as rattlesnakes and adds distinct alterations to the face of generations. This applies to images uploaded for image to image, as well as text to image generations.

____


This is an example of using the LoRa with image to image, to change the look of an image. The setting and train maintain their general look, while the character has a drastic transformation to match the style preset of the LoRa. There are some slight changes such as the front of the train, these changes are based on the LoRa strength.

___


For this progression of Taz in a tornado, each step has more strength of the LoRa added, to the point where full 2 strength is nearly unrecognizable to the original source image. This comes down to personal taste and the direction you wish to take the image.

___

From this example from my workflow, you can see how I use three separate LoRas at once, each with different strength settings, this allows greater control and ability to shift the image in any direction you see fit. Pressing the + icon next to LoRa in the LoRa box in Image generation will take you to the Select Model (LoRa) page, where you can search styles and experiment. You can also train your own LoRa models with SeaArt and use those as well with ones already uploaded.

LoRas are not required to make an image with SeaArt, they are an optional ability to introduce specific changes in style to the images you are making. Additionally, you can emphasize the same changes in your text prompting. There are many use cases for LoRas, from complimentary aesthetics, to juxtaposition and clashing by prompting the opposite style of your LoRas. It is up to you the creator to decide which types of LoRas will be chosen to fit the image you are making.

Different models (Flux, SDXL) have different LoRas that work with them, so you can select different models and search the LoRas that are compatible.

LoRas can add realism, a certain medium or style, different VFX, it takes patience and experimentation to find the right blends that work for you. Have fun creating and finding out for yourself.


596
36
0 comment
36
596
0
0/1000
Related Guides
cover
COMPREHENSIVE GUIDE DIRECTORY
avatar
S
avatar_frame
SeaArt Official
15627
5999
cover
Illustrious LoRA Training Advanced Guide
avatar
椎
avatar_frame
椎名
67
1071
cover
【基礎】illustrious プロンプト解説
avatar
椎
avatar_frame
椎名
25
916
cover
LTX2.3 PROMPTING GUIDE
avatar
W
avatar_frame
wiz.X
30
211
cover
Train Your Own LoRA in 3 Simple Steps
avatar
A
avatar_frame
Ai-GMQ
21
622
cover
【初心者向け】生成で使えるプロンプト一覧
avatar
椎
avatar_frame
椎名
18
538
cover
Illustrious XL Character Generation Guide - Prompt Techniques Without LoRA
avatar
S
avatar_frame
SeaArt Official
172
559
logo
English
Application
Create Image AI Characters Swift AI Model Training Canvas AI Apps Workflow
About
Studio Rankings AI Chat AI Blog AI News
Help
Guides Customer Service
Get App
icon
Download on the
APP Store
icon
GET IT ON
Google Play
Follow Us
iconiconiconiconiconiconiconicon
© 2026 SeaArt, Inc.
Copyright Policy
Terms
Privacy 特定商取引法 資金決済法に基づく表示
More
SUNRISEAI PTE. LTD.