3D_everything
5.0
0 reviews
Description
基于SV3D的一套优化流程
将图片高清放大至1152*1152
增加了脸部优化功能，如需要跑其他物品可以关掉或者无视
增加了seg遮罩功能，能让主体更好的独立出
An optimization process based on SV3D includes the following features:
- Upscaling images to 1152*1152 resolution for enhanced clarity.
- Added facial optimization capabilities, which can be turned off or ignored if optimizing other items.
- Introduced segmentation (seg) masking functionality, allowing the subject to stand out more prominently.
Discussion
(No comments yet)
Loading...
Reviews
No reviews yet
Versions (1)
- latest (5 months ago)
Node Details
Primitive Nodes (42)
Anything Everywhere (2)
Anything Everywhere3 (1)
GetNode (23)
IPAdapterApply (3)
InsightFaceLoader (1)
Reroute (1)
SetNode (11)
Custom Nodes (73)
- ADE_AnimateDiffUniformContextOptions (1)
- ADE_AnimateDiffLoaderWithContext (1)
ComfyUI
- VideoTriangleCFGGuidance (1)
- KSampler (3)
- VAELoader (1)
- CLIPTextEncode (2)
- VAEDecode (3)
- PreviewImage (8)
- ControlNetLoader (5)
- VAEEncode (2)
- LoraLoaderModelOnly (1)
- ImageBlend (2)
- CheckpointLoaderSimple (2)
- SV3D_Conditioning (1)
- CLIPVisionLoader (1)
- UpscaleModelLoader (1)
- LoadImage (1)
- InvertMask (1)
- MaskToImage (1)
- RIFE VFI (1)
- FaceDetailer (1)
- UltralyticsDetectorProvider (1)
- SAMLoader (1)
- ImpactImageBatchToImageList (1)
- ImageListToImageBatch (1)
- AIO_Preprocessor (2)
- ImageGenResolutionFromImage (1)
- PixelPerfectResolution (1)
- HintImageEnchance (1)
- IPAdapterModelLoader (2)
- ACN_AdvancedControlNetApply (5)
- VHS_VideoCombine (7)
- ColorMatch (3)
- ImageGridComposite2x2 (1)
- SDXLPromptStyler (1)
- GroundingDinoModelLoader (segment anything) (1)
- SAMModelLoader (segment anything) (1)
- GroundingDinoSAMSegment (segment anything) (1)
- GetImageSize (1)
- UltimateSDUpscale (1)
Model Details
Checkpoints (2)
A_2D\ghostmix_v20Bakedvae.safetensors
Z other\sv3d_u.safetensors
LoRAs (1)
Y 功能\ip-adapter-faceid-plus_sd15_lora.safetensors