<h2>这是试验品，并不稳定</h2><h2 id="this-is-an-experimental-sample-and-is-not-stable">This is an experimental sample and is not stable</h2>所有示例图均未使用插件None of the example images used plugins.<s>瞎几把融，发出来主要是把C站当网盘使</s><h3 id="negative-prompt">negative prompt</h3><pre><code>worst quality, low quality, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, jpeg artifacts, signature, watermark, username, blurry, artist name</code></pre>第一次尝试融合模型，是比较简陋的试验验型，本意是为了解决3D、2.5D模型使用动漫LORA容易出现崩坏、画风剧变等问题<s>wai佬实在太强了</s>This was my first attempt at fusion modeling, and it was a rather rudimentary experimental model. The original intention was to solve the problems that often occurred when using anime Lora models in 3D and 2.5D, such as the breakdown of the animation and drastic changes in art style欢迎反馈Feedback is welcome.

X1MIX

好吧。。。我所预想中的那个V2比较失败，只能把那些设想放到下一个轮测试中了。。。相较V1，这个版本调整了对色彩和光影的控制，有大幅改善Okay... my envisioned V2 was a bit of a failure, so I'll have to put those ideas into the next round of testing...Compared to V1, this version has significantly improved the control over color and lightingDPMPP_2M+Karras25-40step

x1mix_v20O.safetensors

x1mix_v20.safetensors

v2.0O

v2.0O DMD2

v2.0 DMD2

v1.0

CFG:1.5-5.5STEP:8+大于1024*1536的图特别是横图需要抽卡，但也能出1920*1080、2048*1024之类的，小抽几张就行Images larger than 1024*1536, especially horizontal images, may require multiple attempts, but you can still get images like 1920*1080 or 2048*1024. Just try a few times

Files with SHA256: dc4777990eeb112b31138d4dfe5318a132aa74dad46dc112ee344fd6bbe285ff