Nodes/WanImageToVideoTiledVAE for ComfyUI/Wan22ImageToVideoLatent (Tiled VAE encode)
ComfyUI Node

Wan22ImageToVideoLatent (Tiled VAE encode)

A ComfyUI node in conditioning/inpaint with 10 inputs and 1 output.

By stduhpf·Created about a year ago·Updated 10 months ago· 16
Wan22ImageToVideoLatent (Tiled VAE encode)
  • vae
  • start_image
  • LATENT
width1280
height704
length49
batch_size1
tile_size512
overlap64
temporal_size64
temporal_overlap8
Categoryconditioning/inpaint

Inputs (10)

NameTypeDefaultDescription
vaeVAE
widthINT128032–16384
heightINT70432–16384
lengthINT491–16384
batch_sizeINT11–4096
tile_sizeINT512128–4096
overlapINT640–4096
temporal_sizeINT648–4096Amount of frames to encode at a time.
temporal_overlapINT84–4096Amount of frames to overlap.
start_imageoptIMAGE

Outputs (1)

NameTypeDescription
LATENTLATENT