EmptySD3LatentImage - ComfyUI Built-in Node Documentation - ComfyUI

EmptySD3LatentImage creates a blank (all-zero) latent image in the layout that Stable Diffusion 3 models expect. Because the latent is empty, it is normally used as the starting point that a generation workflow fills in with an image. The width and height you choose determine the size of the final image.

Inputs

Parameter	Description	Data Type	Required	Range
`width`	The width of the latent image in pixels (default: 1024). Values are stepped in increments of 16.	INT	Yes	16 to MAX_RESOLUTION (step: 16)
`height`	The height of the latent image in pixels (default: 1024). Values are stepped in increments of 16.	INT	Yes	16 to MAX_RESOLUTION (step: 16)
`batch_size`	The number of latent images to generate in the batch (default: 1).	INT	Yes	1 to 4096

Outputs

Output Name	Description	Data Type
`LATENT`	A latent tensor containing blank (all-zero) samples in the SD3-compatible format. The tensor has 16 channels, is scaled down by a factor of 8 relative to `width` and `height`, and carries a spatial downscale ratio of 8.	LATENT

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): 694ede56f43e3f3889b4d23e636fa6b33b490bcbd214584557f0dc883fa0a32d