（二）Stable Diffussion 图生图模块-换头发颜色

局部重绘

第一篇中已经完成了相关环境部署，现在开始体验SDwebUI的img2img功能

原图人物模型

使用相同的参数得到相同的图片

(masterpiece:1.0), (best quality:1.0), (ultra highres:1.0) ,(8k resolution:1.0),(realistic:1.0),(ultra detailed1:0), (sharp focus1:0), (RAW photo:1.0),full body,simple background,beautifull girl,solo focus,tall,skirt,high heels,sea side,sky,tree
Negative prompt: (easynegative:1.2), (worst quality: 1.2), (low quality: 1.2),nsfw,by  , by  , by , by < bad_prompt_version2:0.8>
Steps: 40, Sampler: DPM++ SDE Karras, CFG scale: 7, Seed: 1845414120, Face restoration: CodeFormer, Size: 512x786, Model hash: 7234b76e42, Model: chilloutmixni, Denoising strength: 0.45, Hires upscale: 1.4, Hires steps: 11, Hires upscaler: Latent (bicubic antialiased)

通过配置生成图片后, 点击send to img2img或者send to inpaint

图生图 （img2img）

1. 可以在上一步的txt2img生成后发送的img2img，好处是生成图片所需要的提示词模型等等参数会直接传过去
   
2. 也可以手动切换到img2img, 上传图片，然后配置和原图一样的参数
   

参数设置

在上一步，不管是本地还是重新生成，我们都配置了参数，如果是其他图片（没有图片参数），那么最起码设置一下图片大小保持和上传的图片一直

局部重绘（Inpaint）

颜色用：red,pink (银色，粉红色)

• blue hair, red hair 多个描述词组合
• (blue hair:1.1), (red hair 1.3) 多个描述词加权重
• [blue:red:0.16] hair 步数权重
• [blue|red] hair 轮流交替

blue hair, red hair

Negative prompt: (easynegative:1.2), (worst quality: 1.2), (low quality: 1.2),nsfw,by  , by  , by , by < bad_prompt_version2:0.8>

Steps: 20, Sampler: DPM++ SDE Karras, 
CFG scale: 7, 
Seed: 1845414120, 
Size: 512x786, Model hash: 7234b76e42, 
Model: chilloutmixni, 
Denoising strength: 0.65, 
Mask blur: 4, 
ControlNet Enabled: True, 
ControlNet Preprocessor: canny, 
ControlNet Model: control_sd15_canny [fef5e48e], 
ControlNet Weight: 1, ControlNet Starting Step: 0, ControlNet Ending Step: 1, 
ControlNet Resize Mode: Crop and Resize, 
ControlNet Pixel Perfect: True, 
ControlNet Control Mode: My prompt is more important, 
ControlNet Preprocessor Parameters: "(512, 100, 200)"

上述参数无法体现的设置，很重要
Masked content: Original
Inpaint area: Only masked