santa hat
deerdeer nosedeer glow
Sign In

尘歌盆 Potted World

80
414
4
Type
LoRA
Stats
414
Reviews
Published
Jun 9, 2023
Base Model
SD 1.5
Trigger Words
potted
Hash
AutoV2
3BBDD5D29F

简介 Introduction

此模型擅长画花盆上的中式古建筑,如果你喜欢,请不要吝啬你的赞和收藏!如果你要支持我,可以通过微信(Anson_NgAu)联系我,让我有电费可以继续跑图训练。

This LoRA excels in depicting traditional Chinese architecture on potted plants. If you enjoy it, please don't hesitate to give it your likes and save it! If you would like to support me, you can contact me via WeChat (Anson_NgAu) to contribute towards my electricity bill for further training.

使用指引 Guidelines

  • 由于我使用了additional network来加载lora,因此你看到的演示图提示词并不包含lora名,在你使用时,记得要在提示词里调用lora,权重0.6到0.75之间为宜。你也可以通过更高的权重来画一些大模型中包含的著名地标性建筑,如埃菲尔铁塔。You can incorporate the following prompts: potted, East Asian architecture, water, bird, garden, cloud, blue sky, sun (to highlight), and other related elements. Additionally, if you prefer to draw modern or Western architecture, include "East Asian architecture" along with negative prompts and use keywords like church, castle, palace, etc.

  • 由于我使用了additional network来加载lora,因此你看到的演示图提示词并不包含lora名,在你使用时,记得要在提示词里调用lora,权重0.6到0.75之间为宜。你也可以通过更高的权重来画一些大模型中包含的著名地标性建筑,如埃菲尔铁塔。Please note that I have utilized an additional network to load Lora. Therefore, the demonstration prompts you see may not include Lora names. When you use it, remember to invoke Lora in the prompts, with a weight between 0.6 and 0.75 being appropriate. You can also use higher weights to draw famous landmarks included in larger models, such as the Eiffel Tower.

  • 此lora与以下大模型配合较好:chilloutmix(采样方法 - DPM++ 2M SDE Karras)、revAnimated(采样方法 - DPM++ 2M SDE Karras)、以及AnythingV3或V5(VAE - animevae.pt)。This Lora model works well with the following large models: chilloutmix (sampling method - DPM++ 2M SDE Karras), revAnimated (sampling method - DPM++ 2M SDE Karras), and AnythingV3 or V5 (VAE - animevae.pt).

  • 欢迎评论你使用该模型生成的作品,如果你发现配合哪些其他大模型有更好的效果,请一定不要吝啬,尽管分享!!!I welcome your comments on the artwork generated using this model. If you discover better results when combined with other large models, please do not hesitate to share your findings!