ComfyUI Node
HSWQ Ultimate SD Upscale
Upscales an image and runs image-to-image on tiles from the input image.
HSWQ Ultimate SD Upscale
- image
- model
- positive
- negative
- vae
- upscale_model
- IMAGE
◄upscale_byAuto►
◄target_height4320►
◄seed0►
◄steps20►
◄cfg8.00►
◄sampler_name▾►
◄scheduler▾►
◄denoise0.20►
◄mode_type▾►
◄tile_width512►
◄tile_height512►
◄mask_blur8►
◄tile_padding32►
◄seam_fix_mode▾►
◄seam_fix_denoise1.00►
◄seam_fix_width64►
◄seam_fix_mask_blur8►
◄seam_fix_padding16►
◄force_uniform_tilestrue►
◄tiled_decodefalse►
◄batch_size1►
◄tensor_boostfalse►
Categoryimage/upscaling
Inputs (28)
| Name | Type | Default | Description |
|---|---|---|---|
| image | IMAGE | The image to upscale. | |
| model | MODEL | The model to use for image-to-image. | |
| positive | CONDITIONING | The positive conditioning for each tile. | |
| negative | CONDITIONING | The negative conditioning for each tile. | |
| vae | VAE | The VAE model to use for tiles. | |
| upscale_by | COMBO | Auto | Choose 'Auto' to calculate the scale from target vertical pixels, or select a fixed magnification. |
| target_height | INT | 432064–8192 | Target output height in pixels. Used only when upscale_by is 'Auto'. |
| seed | INT | 00–18446744073709550000 | The seed to use for image-to-image. |
| steps | INT | 201–10000 | The number of steps to use for each tile. |
| cfg | FLOAT | 8.000–100 | The CFG scale to use for each tile. |
| sampler_name | COMBO | The sampler to use for each tile. | |
| scheduler | COMBO | The scheduler to use for each tile. | |
| denoise | FLOAT | 0.200–1 | The denoising strength to use for each tile. |
| upscale_model | UPSCALE_MODEL | The upscaler model for upscaling the image. | |
| mode_type | COMBO | The tiling order to use for the redraw step. | |
| tile_width | INT | 51264–8192 | The width of each tile. |
| tile_height | INT | 51264–8192 | The height of each tile. |
| mask_blur | INT | 80–64 | The blur radius for the mask. |
| tile_padding | INT | 320–8192 | The padding to apply between tiles. |
| seam_fix_mode | COMBO | The seam fix mode to use. | |
| seam_fix_denoise | FLOAT | 1.000–1 | The denoising strength to use for the seam fix. |
| seam_fix_width | INT | 640–8192 | The width of the bands used for the Band Pass seam fix mode. |
| seam_fix_mask_blur | INT | 80–64 | The blur radius for the seam fix mask. |
| seam_fix_padding | INT | 160–8192 | The padding to apply for the seam fix tiles. |
| force_uniform_tiles | BOOLEAN | true | Force all tiles to be the same as the set tile size, even when tiles could be smaller. This can help prevent the model from working with irregular tile sizes. |
| tiled_decode | BOOLEAN | false | Whether to use tiled decoding when decoding tiles. |
| batch_size | INT | 11–4096 | The number of tiles to process in a batch. Higher values can reduce processing time but use more VRAM. If you get tensor size mismatch with FP8/FP4 (quantized) models, set this to 1. |
| tensor_boost | BOOLEAN | false | Enable Blackwell Per-Weight CUDA Graph Tensor Boost during USDU tile upscaling. ON raises VRAM by several GB; RTX 5090 32GB+ recommended. Keep OFF for tiled upscale. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| IMAGE | IMAGE | The final upscaled image. |