ComfyUI Node
💪TB | Tiny Decode
Decode a latent representation back into an image using the provided VAE. This node optimizes memory usage by dividing the process into smaller tiles.
💪TB | Tiny Decode
- latent
- vae
- image
◄tile_size512px►
◄overlap100%►
Category💪TinyBreaker/latent
Inputs (4)
| Name | Type | Default | Description |
|---|---|---|---|
| latent | LATENT | The latent representation to decode back into image. | |
| vae | VAE | The VAE model used for decoding the latent representation. | |
| tile_size | COMBO | 512px | The size of the tiles used to divide the input latent into smaller regions for processing, expressed in pixels of the output image. A lower tile size reduces memory usage but may result in lower image quality. |
| overlap | COMBO | 100% | The percentage of overlap between adjacent tiles. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| image | IMAGE | The decoded image. |