Type: huggingface.image_to_image.VAEDecode

Namespace: huggingface.image_to_image

Description

Decodes latents into an image using a VAE. tensor (TorchTensor) -> image

Properties

Property Type Description Default
model hf.vae The VAE model to use. {'type': 'hf.vae', 'repo_id': '', 'path': None, 'variant': None, 'allow_patterns': None, 'ignore_patterns': None}
latents torch_tensor Latent tensor to decode. {'type': 'torch_tensor', 'value': None, 'dtype': '<i8', 'shape': [1]}
scale_factor float Scaling factor used for encoding (inverse is applied before decode) 0.18215

Outputs

Output Type Description
output image  

Metadata

Browse other nodes in the huggingface.image_to_image namespace.