README.md

---
language:
- en
license: creativeml-openrail-m
tags:
- Marvel
- stable-diffusion
- text-to-image
- diffusers
inference: true
---

### Marvel What_If Diffusion
This model is trained on images from the animated Marvel Disney+ show What If.
Which includes characters, background, and some objects.

To reference the art style, use the token: whatif style

Here are some samples.

**Portraits**
![watif4.png](https://s3.amazonaws.com/moonup/production/uploads/1670730628768-635eafb49f24f6db0a1eafd1.png)
![watif3.png](https://s3.amazonaws.com/moonup/production/uploads/1670730629466-635eafb49f24f6db0a1eafd1.png)
![watif5.png](https://s3.amazonaws.com/moonup/production/uploads/1670730628689-635eafb49f24f6db0a1eafd1.png)

Prompt used:
Portrait of *name* in whatif style
Guidance: 7
Steps:  50 using DDIM

I'm not a prompt wizard so you can definitely get better results with some tuning.

**Landscapes**
![watif.png](https://s3.amazonaws.com/moonup/production/uploads/1670730629152-635eafb49f24f6db0a1eafd1.png)

**Objects**
![watif2.png](https://s3.amazonaws.com/moonup/production/uploads/1670730629101-635eafb49f24f6db0a1eafd1.png)