File size: 1,217 Bytes
4b0d9cd
f0d60be
 
 
4b0d9cd
f0d60be
 
 
 
cb5615e
 
4b0d9cd
f0d60be
 
 
37bf9e3
f0d60be
50c70b3
f0d60be
 
 
 
 
1ccf5b2
f0d60be
cb5615e
f0d60be
1ccf5b2
f0d60be
 
 
 
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
---
language:
- en
thumbnail: "https://huggingface.co/wavymulder/collage-diffusion/resolve/main/pics/thumb.jpg"
license: creativeml-openrail-m
tags:
- stable-diffusion
- text-to-image
- safetensors
- diffusers
inference: true
---



**Collage Diffusion**
![Header](https://huggingface.co/wavymulder/collage-diffusion/resolve/main/pics/page1.jpg)
[*CKPT DOWNLOAD LINK*](https://huggingface.co/wavymulder/collage-diffusion/resolve/main/collage-1.0.ckpt)  - - - [*SAFETENSORS DOWNLOAD LINK*](https://huggingface.co/wavymulder/collage-diffusion/resolve/main/collage-1.0.safetensors)

This is a dreambooth model trained on collage images.

Use the activation token **collage style** in your prompt (I recommend at the start)

I have the most fun with this model when I use simple prompts and let the model go crazy. If you want a model that strictly adheres to your prompt, this isn't that.

Trained from 1.5 with VAE. 

Please see [this document where I share the parameters (prompt, sampler, seed, etc.) used for all example images.](https://huggingface.co/wavymulder/collage-diffusion/resolve/main/parameters_for_samples.txt)



![Page2](https://huggingface.co/wavymulder/collage-diffusion/resolve/main/pics/page2.jpg)