[SD 1.5][Checkpoint]viewer-mix_v1.3 viewer-mix_v1.3
<h1>v1.3</h1><p>改动：</p><p>1、 将vae文件烘焙进去了。</p><p>2、调整了原料的比例重新进行混合，优化了色调和氛围，使其看起来整体光影效果更具戏剧性。</p><p>3、训练了一个名为<a target="_blank" rel="ugc" href="https://civitai.com/models/11772/verybadimagenegative">veryBadImageNegative</a>的负面嵌入，所有的例子均使用了这个负面嵌入。</p><p></p><p>Changes:</p><p>1. Baked the vae file.</p><p>2. After adjusting the weight, re-mixing is performed to optimize the tone and atmosphere, making the overall light and shadow effect more dramatic.</p><p>3. A negative embedding called <a rel="ugc" href="https://civitai.com/models/11772/verybadimagenegative">veryBadImageNegative </a>is trained, and all examples use this negative embedding.</p><p></p><h1>v1.0</h1><p></p><p>该模型可以生成高质量的动漫风格壁纸、动漫风格角色、自然风景插图。</p><p>推荐使用pastel-waifu-diffusion.vae.pt，所有例子都是使用这个vae生成的。</p><p></p><p>this model can generate high-quality anime style wallpaper, anime style characters, and natural scenery illustrations.</p><p>pastel-waifu-diffusion.vae.pt is recommended, all examples are generated using this vae.</p>
页:
[1]