TenserTensor
f84e17ba0a
Implement VAE decoding nodes: - vae_decode_context.py: context-based VAE decoder - vae_decode_tiled.py: tiled VAE decoder for large images with VRAM optimization Add markdown documentation for VAE decoder nodes
1.6 KiB
1.6 KiB
TT_VaeDecodeTiled
This description was generated with AI assistance. If you spot any errors, please create an issue on GitHub.
Tiled VAE decoder for processing large images without running out of VRAM.
Inputs
| Parameter | Type | Required | Description |
|---|---|---|---|
vae |
VAE | Yes | VAE model for decoding |
latent |
LATENT | Yes | Latent image to decode |
tile_width |
INT | Yes | Width of each tile in pixels (320-4096, step: 64, default: 512) |
tile_height |
INT | Yes | Height of each tile in pixels (320-4096, step: 64, default: 512) |
overlap |
INT | Yes | Overlap between tiles to reduce seams (0-256, step: 8, default: 64) |
circular |
BOOLEAN | Yes | Enable circular padding for seamless tiling (default: False) |
Outputs
| Parameter | Type | Description |
|---|---|---|
IMAGE |
IMAGE | Decoded image tensor |
Usage
Decode large latent images by processing them in smaller tiles. Useful for generating high-resolution images that would otherwise exceed VRAM capacity. The overlap parameter helps reduce visible seams between tiles. Enable circular padding for seamless texture generation.