Nodes/FUSE Face Enhancer/FUSE KSampler (Video)
ComfyUI Node

FUSE KSampler (Video)

A ComfyUI node in Face Enhancement/Video with 23 inputs and 2 outputs.

By WASasquatch·Created about a year ago·Updated 9 months ago· 13
FUSE KSampler (Video)
  • model
  • vae
  • images
  • positive
  • negative
  • mask_optionals
  • yolo_optionals
  • images
  • masks
use_cachetrue
seed0
steps20
cfg3.50
sampler_name
scheduler
denoise0.50
yolo_detector
sam_segmenter
sam_model_typevit_b
face_id0
face_orderlinear
face_size512
face_padding20
force_squaretrue
temporal_trackingtrue
CategoryFace Enhancement/Video

Inputs (23)

NameTypeDefaultDescription
modelMODELBase model for sampling. Should support video if input is 5D.
vaeVAEVAE model for the sampler. Should support video if input is 5D.
imagesIMAGEInput image batch (4D) or video (5D) for face detection and upscaling.
positiveCONDITIONINGPositive conditioning for the sampler.
negativeCONDITIONINGNegative conditioning for the sampler.
use_cacheBOOLEANtrueUse internal caching to speed up workflow iteration.
seedINT00–18446744073709550000Seed for the sampler.
stepsINT20Number of steps for the sampler.
cfgFLOAT3.50Classifier-Free Guidance scale for the sampler.
sampler_nameCOMBOChoose the sampler to use for sampling.
schedulerCOMBOChoose the scheduler to use for sampling.
denoiseFLOAT0.500–1Denoising strength for the sampler.
yolo_detectorCOMBOChoose the YOLO detector to use for face detection. Prefix 'face/' for face-specific models, 'yolo/' for general YOLO models.
sam_segmenterCOMBOChoose the SAM segmentation model to use for face segmentation.
sam_model_typeCOMBOvit_bSAM model type
face_idINT0Index of the face to process (-1 for all faces).
face_orderCOMBOlinearOrder to process detected faces.
face_sizeCOMBO512The resolution to sample the face crop at.
face_paddingINT200–16384Padding in pixels to pad the face crop with.
force_squareBOOLEANtrueForce 1:1 square face crops
temporal_trackingBOOLEANtrueEnable temporal face tracking for video inputs.
mask_optionalsoptDICTOptional masking and blending settings from FUSESamplerMaskOptions node.
yolo_optionalsoptDICTOptional YOLO detection settings from FUSEYOLOSettings node.

Outputs (2)

NameTypeDescription
imagesIMAGE
masksMASK