---
license: creativeml-openrail-m
language:
- en
- ja
tags:
- Stable-Diffusion
- lora
---
# 【LoRA】witchpot-citynight-sd-1-5
LoRA for 2D game city silhouette night stage    
[witchpot-citynight-sd-1-5](https://huggingface.co/Witchpot/CitySilhouette_Night/resolve/main/witchpot-citynight-sd-1-5.safetensors)    
All training data is generated by Midjourney    

## Trigger
- citynight

## Sample Prompts
- citynight, jump game level design, house and buildings
- Negative prompt: character, enemy, sun, multiple moon

## Sample Images
![sample1](https://huggingface.co/Witchpot/CitySilhouette_Night/resolve/main/76b6b17c-f428-4b60-bdd1-b2bddb14c69f.png)
![sample2](https://huggingface.co/Witchpot/CitySilhouette_Night/resolve/main/7a39bd4e-83a1-4072-879f-5a1fdee1f918.png)    
Left one is depth to image input depth 3D model data on stable-diffusion-for-unity    
![sample3](https://huggingface.co/Witchpot/CitySilhouette_Night/resolve/main/CityNight.png)

## Model Description
- Model type: [LoRA]
- Base Model: Model trained with runwayml/stable-diffusion-v1-5/v1-5-pruned.ckpt (https://huggingface.co/runwayml/stable-diffusion-v1-5/blob/main/v1-5-pruned.ckpt)

## Recommendations
This LoRA model has been trained to generate game stages made of silhouette city at night, based on specific patterns.    
You can find evening version from [this](https://huggingface.co/Witchpot/CitySilhouette_Evening)    
By combining it with Depth2Image, you can create consistent game stages.     
This LoRA is supposed to use with [stable-diffusion-for-unity](https://docs.witchpot.com/)

## Information
- https://twitter.com/Witchpot_