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

### Firewatch Diffusion
This model is trained on in-game footages, and some game posters by OllyMoss. 
The model can make good looking nature sceneries similar to the game art style, with some objects.
It can not make portraits at all, since throughout the game you don't actually meet any characters.

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

Here are some samples.

**Sneneries**
![fwatch.png](https://s3.amazonaws.com/moonup/production/uploads/1670473030496-635eafb49f24f6db0a1eafd1.png)
![fwatch2.png](https://s3.amazonaws.com/moonup/production/uploads/1670473031879-635eafb49f24f6db0a1eafd1.png)

Prompt used:
Beautiful *mountain/lake/beach/etc* in fwatch style
Guidance: 7
Steps: 65 using DDIM

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

**Objects**
![fwatch3.png](https://s3.amazonaws.com/moonup/production/uploads/1670473031927-635eafb49f24f6db0a1eafd1.png)