ComfyUI Node
Flow Matching Progressive Upscaler
A ComfyUI node in latent/upscaling with 28 inputs and 5 outputs.
Flow Matching Progressive Upscaler
- model
- positive
- negative
- latent
- latent
- next_seed
- model
- positive
- negative
◄seed0►
◄steps_per_stage16►
◄cfg4.5►
◄sampler_name▾►
◄scheduler▾►
◄total_scale4.00►
◄stages2►
◄renoise_start0.35►
◄renoise_end0.15►
◄skip_blend_start0.80►
◄skip_blend_end0.05►
◄upscale_methodbicubic►
◄noise_schedule_override►
◄skip_schedule_override►
◄denoise1.00►
◄enable_dilated_samplingenable►
◄dilated_downscale2.00►
◄dilated_blend0.25►
◄dilated_min_steps1►
◄dilated_seed_modederive►
◄dilated_denoise0.50►
◄cleanup_stagedisable►
◄cleanup_noise0.00►
◄cleanup_denoise0.40►
Categorylatent/upscaling
Inputs (28)
| Name | Type | Default | Description |
|---|---|---|---|
| model | MODEL | Flow-matching diffusion model to drive refinement. | |
| positive | CONDITIONING | Positive conditioning for CFG. | |
| negative | CONDITIONING | Negative conditioning for CFG. | |
| latent | LATENT | Low resolution latent to progressively upscale. | |
| seed | INT | 00–18446744073709550000 | Base seed controlling the re-noising at each stage. |
| steps_per_stage | INT | 161–256 | Denoising steps executed at every progressive stage. |
| cfg | FLOAT | 4.50–20 | Classifier Free Guidance strength per stage. |
| sampler_name | COMBO | Sampler backend leveraged during refinement. | |
| scheduler | COMBO | Noise schedule applied during denoising. | |
| total_scale | FLOAT | 4.001–16 | Overall scale factor from the input latent to the final output. |
| stages | INT | 21–5 | Number of progressive stages to reach the total scale. |
| renoise_start | FLOAT | 0.350–1 | Noise ratio applied at the first stage. |
| renoise_end | FLOAT | 0.150–1 | Noise ratio applied at the last stage (interpolated in-between). |
| skip_blend_start | FLOAT | 0.800–1 | Blend weight for the upsampled latent at the first stage. |
| skip_blend_end | FLOAT | 0.050–1 | Blend weight for the upsampled latent at the last stage. |
| upscale_method | COMBO | bicubic | Algorithm used when resizing the latents between stages. |
| noise_schedule_overrideopt | STRING | Comma-separated override for per-stage noise ratios. Empty uses interpolated start/end. | |
| skip_schedule_overrideopt | STRING | Comma-separated override for per-stage skip blend weights. | |
| denoiseopt | FLOAT | 1.000–1 | Denoising strength supplied to the sampler per stage. |
| enable_dilated_samplingopt | COMBO | enable | Optionally run a dilated refinement pass for global coherence. |
| dilated_downscaleopt | FLOAT | 2.001–4 | Factor used when downscaling for the dilated pass (>=1.0). |
| dilated_blendopt | FLOAT | 0.250–1 | Blend weight of the dilated refinement result. |
| dilated_min_stepsopt | INT | 11–256 | Minimum sampling steps for dilated refinement. Actual steps = max(min_steps, steps // 2). Increase for lightning models to ensure adequate sampling (e.g., 4). |
| dilated_seed_modeopt | COMBO | derive | Seed mode for dilated sampling. 'derive' adds 10,000 to base seed, 'same' uses identical seed. |
| dilated_denoiseopt | FLOAT | 0.500–1 | Denoising strength for dilated sampling. Lower values preserve more spatial structure from the original. |
| cleanup_stageopt | COMBO | disable | Run an extra non-scaling clean-up denoise pass at the end. |
| cleanup_noiseopt | FLOAT | 0.000–1 | Noise ratio for the optional clean-up stage (set to 0 to disable re-noising). |
| cleanup_denoiseopt | FLOAT | 0.400–1 | Denoising strength used during the clean-up stage. |
Outputs (5)
| Name | Type | Description |
|---|---|---|
| latent | LATENT | — |
| next_seed | INT | — |
| model | MODEL | — |
| positive | CONDITIONING | — |
| negative | CONDITIONING | — |