ComfyUI Node
K2 Post Upscale
Scales the finished image to an exact final size. 'lanczos' is a deterministic high-quality resize. 'upscale_model' first runs the connected ComfyUI UPSCALE_MODEL (typically 2x or 4x) and then resizes to the exact requested factor, so a 4x model can serve a 2.5x target without a second node.
K2 Post Upscale
- image
- upscale_model
- image
◄scale2.00►
◄methodlanczos►
◄tile512►
CategoryFVM Tools/K2
Inputs (5)
| Name | Type | Default | Description |
|---|---|---|---|
| image | IMAGE | Image batch to enlarge. | |
| scale | FLOAT | 2.001–8 | Final width and height multiplier. |
| method | COMBO | lanczos | lanczos: deterministic high-quality resize. upscale_model: run the neural model first, then resize to the exact target. |
| upscale_modelopt | UPSCALE_MODEL | Required only for the 'upscale_model' method. | |
| tileopt | INT | 512128–2048 | Tile size for the neural upscaler (lower = less VRAM). |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| image | IMAGE | — |