多元素合成动态海报ComfyUI 工作流节点

简介

使用IPAdapter的focus layer mask功能，通过多个元素，合成图片。这种指定元素，指定画面位置，无痕合图的效果是非常不错的。大家可以自搭局部细化和高分修复，完成高清海报/写真。图片合成变的简单高效！同时使用SVD和Animatediff的动效做对比，并针对主体人物做了二次修复，效果丰富细腻。

工作流节点截图

附件文件一览

模型
Checkpoints
svd_image_decoder.safetensors👉前往下载

inswapper_128.onnx👉前往下载

LoRAs
more_details.safetensors👉前往下载

pytorch_lora_weights.safetensors👉前往下载

sd15_lora_beta.safetensors👉前往下载

v2_lora_ZoomIn.ckpt👉前往下载

内置节点
Display Int (rgthree)
List of strings [Crystools]
Reroute

自定义节点
AnimateDiff Evolved
– ADE_AnimateDiffLoaderWithContext
– ADE_AnimateDiffLoRALoader
– ADE_LoopedUniformContextOptions
ComfyUI
– LoadImageMask
– FreeU_V2
– VAEDecode
– SVD_img2vid_Conditioning
– CLIPSetLastLayer
– LoraLoader
– CLIPTextEncode
– KSampler
– EmptyLatentImage
– GrowMask
– CLIPVisionLoader
– LoadImage
– VideoLinearCFGGuidance
– ImageOnlyCheckpointLoader
– PreviewImage
– LatentUpscaleBy
– MaskToImage
– ImageToMask
ComfyUI Essentials
– MaskPreview+
– GetImageSize+
ComfyUI Frame Interpolation
– RIFE VFI
ComfyUI Impact Pack
– ImpactGaussianBlurMask
– MaskToSEGS_for_AnimateDiff
– SAMLoader
– ToBasicPipe
– DetailerForEachPipeForAnimateDiff
ComfyUI Layer Style
– LayerMask: MaskEdgeUltraDetail V2
– LayerColor: Levels
ComfyUI Nodes for Inference.Core
– FakeScribblePreprocessor
– AIO_Preprocessor
ComfyUI WD 1.4 Tagger
– WD14Tagger|pysssss
ComfyUI_IPAdapter_plus
– PrepImageForClipVision
– IPAdapterModelLoader
– IPAdapterTiled
– IPAdapterUnifiedLoader
– IPAdapterAdvanced
ComfyUI-Advanced-ControlNet
– ACN_AdvancedControlNetApply
– ACN_SparseCtrlLoaderAdvanced
– ControlNetLoaderAdvanced
ComfyUI-VideoHelperSuite
– VHS_VideoCombine
Efficiency Nodes for ComfyUI Version 2.0+
– HighRes-Fix Script
– LoRA Stacker
– Efficient Loader
– KSampler (Efficient)
Facerestore CF (Code Former)
– FaceRestoreModelLoader
– FaceRestoreCFWithModel
ReActor Node for ComfyUI
– ReActorFaceSwap
segment anything
– GroundingDinoModelLoader (segment anything)
– GroundingDinoSAMSegment (segment anything)
Use Everywhere (UE Nodes)
– Seed Everywhere