Visit ComfyUI Online for ready-to-use ComfyUI environment
Encode images into latent representations using a Variational Autoencoder for AI art applications with storage options.
The BLVAEEncode node, also known as "VAEEncode (Bundle Latent)," is designed to encode images into latent representations using a Variational Autoencoder (VAE). This process is essential for various AI art applications, as it allows for the efficient compression and manipulation of image data in a lower-dimensional latent space. The node supports both standard and tiled encoding, making it versatile for different image sizes and resolutions. By encoding images into latent space, you can perform advanced operations such as image generation, transformation, and inpainting with greater ease and efficiency. The node also offers options to store or load latent representations, providing flexibility in managing your workflow.
This parameter specifies the Variational Autoencoder (VAE) model to be used for encoding the image. The VAE is responsible for converting the image into its latent representation, which is a compressed version of the image that retains essential features.
This boolean parameter determines whether the image should be encoded in a tiled manner. Tiling can be useful for handling large images by breaking them into smaller, more manageable pieces. The default value is False
.
This integer parameter sets the size of the tiles when the tiled
option is enabled. It defines the dimensions of each tile in pixels. The default value is 512, with a minimum of 320 and a maximum of 4096, adjustable in steps of 64. Tiling can help in processing high-resolution images more efficiently.
This boolean parameter indicates whether the latent representation should be stored or loaded. When set to True
, the latent representation is stored for future use. The default value is True
.
This boolean parameter specifies whether the latent representation should be removed after it is loaded. This can help in managing memory usage by clearing latent data that is no longer needed. The default value is True
.
This boolean parameter determines whether the latent representation should be deleted as part of the workflow. This can be useful for cleaning up and ensuring that no unnecessary data is retained. The default value is False
.
This optional parameter allows you to provide the image that you want to encode. The image is converted into its latent representation by the VAE.
This hidden parameter is used to store additional PNG information that might be required during the encoding process.
This hidden parameter is used to assign a unique identifier to the encoding process, which can be useful for tracking and managing different encoding tasks.
The output parameter latent
represents the encoded latent space representation of the input image. This latent representation is a compressed version of the image that retains its essential features, making it suitable for various AI art applications such as image generation, transformation, and inpainting. The latent output is crucial for performing advanced operations in a more efficient and manageable manner.
tiled
option and adjust the tile_size
parameter according to your needs.store_or_load_latent
parameter to save the latent representation for future use, which can save time in workflows that require repeated encoding.remove_latent_on_load
to True
to clear latent data that is no longer needed.delete_workflow_latent
parameter to clean up latent representations as part of your workflow, ensuring no unnecessary data is retained.vae
parameter is required but not provided.vae
parameter.tile_size
parameter is set to a value outside the allowed range.tile_size
parameter to be within the range of 320 to 4096, in steps of 64.image
parameter is required for encoding but not provided.image
parameter to proceed with the encoding process.store_or_load_latent
option if not needed.© Copyright 2024 RunComfy. All Rights Reserved.