Sign In


Checkpoint Trained
Sep 6, 2023
Base Model
SD 1.5

AWPainting 1.4

This iteration focuses on enhancing scene diversity and visual impact, with improvements in picture details and color saturation compared to version 1.3. The variety of character faces has also increased. Using keywords such as "Top view" and "perspective composition" for controlling the camera can achieve better comprehension and visual effects compared to version 1.3.

Recommended Resolutions: 512x768, 576x864, 640x960

For low-resolution generation, you can use the "hire" HD repair feature to increase the resolution, with the recommended algorithm being ESRGAN_4x.

AWPainting 1.4

本次迭代主要在于提升场景多样性以及画面张力,画面细节以及色彩饱和度相较1.3均有一定提升,人物面部种类更加多样化。使用类似Top view, perspective composition的词汇控制镜头可以获得相较于1.3有更好的理解力和画面效果。

推荐分辨率: 512*768, 576*864, 640*960



AWPainting 1.3 What's New

在1.2的基础上优化了角色面部的多样性,支持男、女及不同肤色,角色动作的类型以及协调性、优化了光照和整体的着色效果,对自然语言响应度良好(可使用GPT写自然语言的Prompts使用)。 翻译成英文

AWPainting 1.3 has optimized the diversity of character faces based on version 1.2, supporting male, female, and different skin colors. It has improved the types of character actions and their coordination, optimized lighting and overall shading effects, and responds well to natural language (can use GPT to write natural language prompts for use).

AWPainting 1.2 What's New


Building on 1.1, we offer more refined visuals, a more lifelike representation and gentler lighting. Also, enhanced response to Prompts

AWPainting 1.1 What's New


搭配新出的LoRA DynamicKungfu可以很好的复现之前我一些作品的风格。

The light effect of the picture and the diversity of character movements have been improved, the facial features of the characters are more distinct and recognizable, and the picture accuracy has been improved to a certain extent. With the new LoRA DynamicKungfu, it can reproduce the style of some of my previous works very well.


AWPainting 1.0


比较有特色的是对niji风的复现,搭配LoRA nijiexpress以及不同LoRA的复用有很好的相性,直出图就会有很高的质量,如例图所示均为t2i直出+高清修复(无controlnet)。

推荐搭配LoRA: Nijiexpress V1 V2,Xiaorenshu小人书,墨心moxin,沁彩

This model has a wide range of applicability, illustrations, two-dimensional and general 2D can be well drawn

What is more distinctive is the reappearance of the niji style. With LoRA nijiexpress and the multiplexing of different LoRAs, there is a good compatibility, and the direct image output will have high quality. As shown in the example images, they are all t2i direct output + HD Fix (no controlnet). Recommended collocation with LoRA: Nijiexpress V1 V2, Xiaorenshu villain book, Moxin moxin, Qincai Fisheyes, masterpiece, top quality, best quality, official art, beautiful and aesthetic, animation

推荐Recommended Prompts:Fisheyes, masterpiece, top quality, best quality, official art, beautiful and aesthetic, animation

Neg_Prompts: EasyNegativeV2, ng_deepnegative_v1_75t, (low_quality:1.4), (worst_quality:1.4), (badhandv4:1.1),collage, artist_name, signature, artist_logo, watermark

免责声明 Disclaimer


"You are responsible for maintaining and independently assessing the text content generated by this model service. You must use it in compliance with applicable laws and regulations and the terms of this service. Any intellectual property issues arising from content generated by this service are your responsibility. I am not responsible for any losses incurred as a result.

Unless otherwise specified by laws and regulations, you may not integrate this model service into any third-party software or service without my written consent. You are also not allowed to disclose the association between the generated content and this service to any third party, directly or indirectly, in any manner, without my consent. If you violate this agreement, I have the right to seek compensation for any losses incurred."