ComfyUI Node
ImageMask_Align
Preserve 1:1 original pixels and pad outward to the nearest multiple. Model alignment reference: • 8: SD 1.5, SD 2.1, SDXL (8x VAE) • 16: FLUX.1-Kontext, SD3 / SD3.5 (Patch 2x2 + 8x VAE) • 32: Qwen-Image 2.1, Qwen2-VL, HunyuanVideo • 64: CogVideoX, Wan2.1 (3D VAE / Video DiT)
ImageMask_Align
- image
- mask
- image
- mask
- align_data
◄round_to_multiple32►
CategoryComfyPanel/Process & Restore
Inputs (3)
| Name | Type | Default | Description |
|---|---|---|---|
| image | IMAGE | — | |
| round_to_multiple | COMBO | 32 | 4 options: 8, 16, 32, 64 |
| maskopt | MASK | — |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| image | IMAGE | — |
| mask | MASK | — |
| align_data | ALIGN_DATA | — |