Async's MIX XL

Async 的 MIX XL 合并检查点模型。

该模型专注于生成动漫图像。

sdxl_vae 已集成。

故障排除

如果你使用 WebUI 且图像为黑色或有异常色调。

在设置中，关闭 StableDiffusion 部分的“启用强调”选项。在这种情况下，结果将与上述示例图像不同。

图像结果与上述示例图像不同。

我创建示例图像时使用的 WebUI 版本为 1.7.0，且未添加任何扩展。我使用了如 an14 或 an12 等负向嵌入来生成图像，因此请下载并将其放置在指定的嵌入文件夹中，以重新生成图像。

an14、an12 负向嵌入

Async 的负向 - v14 | Stable Diffusion 嵌入 | Civitai

设置

clip skip：2

ENSD：0

质量提示

(Depth of field hdr 8k 4k wallpaper cinematic angle, cinematic lighting,:1.5) (masterpiece, best quality:2.0), (Depth of field hdr 8k 4k wallpaper cinematic angle, cinematic lighting,:1.5) (masterpiece, best quality:2.0)

负向提示

n12_light (an12:1.2) (kneeling  lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name,:1.5),, an12_light an12 ( lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name,:1.5)

示例

anime,1girl,(Depth of field hdr 8k 4k wallpaper cinematic angle, cinematic lighting,:1.5) (masterpiece, best quality:2.0), (Depth of field hdr 8k 4k wallpaper cinematic angle, cinematic lighting,:1.5) (masterpiece, best quality:2.0)
Negative prompt: an12_light (an12:1.2) (kneeling  lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name,:1.5),, an12_light an12 ( lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name,:1.5)
Steps: 45, RNG: CPU, Size: 896x1152, Seed: 2079619126, Model: Async's MIX XL, Version: v1.7.0, Sampler: DPM++ 2M Karras, CFG scale: 7.5, Clip skip: 2, Model hash: 9abe366cab, Hires steps: 45, Hires upscale: 1.25, Hires upscaler: Latent (nearest-exact), Denoising strength: 0.6