快速生成精美AI图像ComfyUI 工作流节点

简介

这个工作流的诞生是为了需要批量处理最少 8 个到最多 20 个图像，尺寸从 256×256 到最大 512×512 像素，但创建必须要快，所以我选择了 SD Hyper，出于图形原因修改为Turbo。从图片中可以看出，如果我们将草稿图与成品图进行比较，细节还是相当不错的。

工作流节点截图

附件文件一览

模型
dreamshaperXL_v21TurboDPMSDE.safetensors👉前往下载

内置节点
DF_Integer
Display Int (rgthree)
Fast Bypasser (rgthree)
GetNode
JWFloat
JWInteger
KSampler Config (rgthree)
Label (rgthree)
Note
Power Lora Loader (rgthree)
PrimitiveNode
Reroute
SetNode
Switch image [Crystools]

自定义节点
Allor Plugin
– ImageNoiseGaussian
– ImageEffectsAdjustment
Comfyroll Studio
– CR Text Concatenate
– CR Simple Image Compare
ComfyUI
– ImageUpscaleWithModel
– UpscaleModelLoader
– CheckpointLoaderSimple
– VAEDecodeTiled
– CLIPTextEncode
– EmptyLatentImage
– KSampler
– PreviewImage
– VAEEncodeTiled
– ImageSharpen
– VAEDecode
– ImageBlur
– LoadImage
– EmptyImage
ComfyUI Essentials
– SimpleMath+
– ImageResize+
– GetImageSize+
ComfyUI Impact Pack
– UltralyticsDetectorProvider
– FaceDetailer
– ImpactMakeImageBatch
comfyui-art-venture
– ImageScaleDownBy
Hakkun-ComfyUI-nodes
– Any Converter
Image chooser
– Preview Chooser
KJNodes for ComfyUI
– ColorMatch
mape’s helpers
– mape Variable
Mikey Nodes
– FaceFixerOpenCV
ntdviet/comfyui-ext
– gcLatentTunnel
pythongosssss/ComfyUI-Custom-Scripts
– ShowText|pysssss
– MathExpression|pysssss
SDXL Prompt Styler
– SDXLPromptStyler
SDXL Prompt Styler (customized version by wolfden)
– SDXLPromptStylerbyLighting
Various ComfyUI Nodes by Type
– JWImageResizeByLongerSide
VLM_nodes
– SimpleText
WAS Node Suite
– Text Multiline