Martin Rizzo
1cda0f59d7
This commit makes minor updates to the docsify documentation, including adjusting formatting, adding links to the repository on the homepage, and updating some outdated node images.
1.3 KiB
⚡| VAE Encode (for Soft Inpainting)
This node is designed to encode images into a latent representation suitable for inpainting tasks. It includes additional preprocessing steps to adjust the input image and mask, ensuring optimal encoding by the VAE model.
Inputs
pixels
The input image to be encoded.
vae
The VAE model used to encode the image into its latent representation.
mask
A mask indicating the regions of the input image that will be inpainted.
output_size
Determines the size of the final image on which the inpainting process will be applied.
Available options:
safe_as_input: Uses the same dimensions as the supplied image.large: Scales up to an equivalent area of 1600x1600 pixels.medium: (recommended): Scales up to an equivalent area of 1300x1300 pixels.small: Resizes to fit within an area equivalent to 1024x1024 pixels.
mask_blur_pixels
The number of pixels used for blurring the edges of the mask, creating a softer transition between masked and unmasked areas. Larger values result in smoother edges with increased masked regions.
Outputs
latent
The latent representation of the input image, with the mask embedded to indicate the inpainting regions. This output is essential for subsequent inpainting tasks.