Sign In

Google Emoji XL

10

59

7

4

Updated: Sep 4, 2025

styleemoji

Verified:

SafeTensor

Type

LoRA

Stats

39

7

14

Reviews

Published

Aug 21, 2025

Base Model

SDXL 1.0

Training

Steps: 2,498
Epochs: 10

Trigger Words

emoji

Training Images

Download

Hash

AutoV2
1177517EC0
default creator card background decoration
ayan4m1's Avatar

ayan4m1

This LoRA provides an "emoji" style.

It is trained on Noto Emoji Color, an open source font released by Google.

Square outputs are strongly recommended, anything from 768x768 up should give good results. I also recommend prompting for a background color that is not present on the edges of your subject (e.g. use a magenta background for a white object), so that background removal works properly.

Higher scale (0.8 - 1.4) works best for animals, plants, and people, while lower scale (0.6 - 1) works best for structures and clean lines.

NOTE: If you are using the CivitAI generation service, I recommend generating in batches of 4. One run should cost 27 Buzz, and if you provide feedback on each of the 4 results you will get 16 Buzz back leading to a net cost of 11, or 2.75 Buzz per image. You should get at least one satisfactory result in a batch of four.