ComfyUI Node
Eric Qwen-Edit Spectrum Accelerator
Spectrum Accelerator — training-free ~3–5× diffusion sampling speedup. Predicts transformer outputs on skipped steps using Chebyshev polynomial regression instead of running all 60 transformer blocks. Wire between Loader and any execution node.
Eric Qwen-Edit Spectrum Accelerator
- pipeline
- pipeline
◄enabletrue►
◄warmup_steps3►
◄window_size2►
◄flex_window0.75►
◄w0.50►
◄lam0.10►
◄M4►
◄min_steps15►
CategoryEric/QwenEdit
Inputs (9)
| Name | Type | Default | Description |
|---|---|---|---|
| pipeline | QWEN_EDIT_PIPELINE | — | |
| enable | BOOLEAN | true | Enable/disable Spectrum acceleration. |
| warmup_stepsopt | INT | 31–20 | Number of initial denoising steps that always run the full transformer. More warmup = better forecaster initialization but less speedup. 2–4 recommended. |
| window_sizeopt | INT | 22–10 | Base period between actual transformer evaluations (in steps). window_size=2 means every other step is cached after warmup. Higher = more aggressive caching. |
| flex_windowopt | FLOAT | 0.750–2 | Window growth rate — the window_size increases by this amount after each actual eval. Later diffusion steps change features less, so larger windows are safe. 0 = fixed window. |
| wopt | FLOAT | 0.500–1 | Blend weight between Chebyshev predictor (w=1) and Newton forward-difference predictor (w=0). 0.5 = equal blend (recommended). |
| lamopt | FLOAT | 0.100.001–10 | Ridge regularization for Chebyshev regression. Higher = smoother predictions, lower = more responsive to recent data. |
| Mopt | INT | 41–8 | Chebyshev polynomial degree. Higher M captures more complex feature trajectories but risks overfitting with few data points. |
| min_stepsopt | INT | 155–100 | Spectrum auto-disables when num_inference_steps < min_steps. Low step counts (e.g. 8-step lightning LoRA) don't benefit. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| pipeline | QWEN_EDIT_PIPELINE | — |