Hugging Face's logo

nerijs
/
pixel-art-xl

Text-to-Image
Diffusers
stable-diffusion
lora
Model card Files Community
20
pixel-art-xl
File size: 1,009 Bytes 
630650d
 
8460168
 
 
 
 
 
 
 
133cbbf
630650d
bc0dce7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
 
---
license: creativeml-openrail-m
tags:
  - text-to-image
  - stable-diffusion
  - lora
  - diffusers
base_model: stabilityai/stable-diffusion-xl-base-1.0
instance_prompt: pixel art
widget:
  - text: pixel art, mafalda, simple, flat colors
---
# Pixel Art XL
## Consider supporting further research on [Patreon](https://www.patreon.com/user?u=29466374) or [Twitter](https://twitter.com/nerijs)

![F1hS8XHXwAQrMEW.jpeg](https://cdn-uploads.huggingface.co/production/uploads/6303f37c3926de1f7ec42d3e/SSOQ9lfB1PVhXVWJiL7Mx.jpeg)
![F1hS489X0AE-PK5.jpeg](https://cdn-uploads.huggingface.co/production/uploads/6303f37c3926de1f7ec42d3e/tY19J3xWDlSY2hhTTHySc.jpeg)


Downscale 8 times to get pixel perfect images (use Nearest Neighbors)
Use a fixed VAE to avoid artifacts (0.9 or fp16 fix)

### Tips:
Don't use refiner

Works great with only 1 text encoder

No style prompt required

No trigger keyword require

Works great with isometric and non-isometric

Works with 0.9 and 1.0

#### Changelog
v1: Initial release