The VAE Decode node decode the latent images.

Introduction

This node serves as a pivotal component in the process of translating latent space images back into the realm of pixel space. Based on the provided VAE, it is capable of reconstructing images, making them ready for display or subsequent download.

Inputs

Name	Explanation
samples	The latent images to be decoded.
VAE	The VAE to use for decoding the latent images.

Outputs

Name	Explanation
image	The image reconstructed from the provided latent representation using the specified VAE model.

How to Use

The VAE Decode node is often integrated with other nodes to facilitate a seamless preview of the generated images. Here’s how they typically work in tandem:

Load Checkpoint: This node supplies the VAE model necessary for the VAE Decode node.

• KSampler: This node generates the samples that the VAE Decode node will decode.

• Preview Image: This node receives the decoded image from the VAE Decode node and enables the preview of images directly within the node graph.

In the typical workflow, the latent image processed by the KSampler node is decoded by the VAE Decode node, with the results then presented in the Preview Image node’s interface.