Nodes/ComfyUI-FunPack/FunPack Hybrid Euler 2S Sampler
ComfyUI Node

FunPack Hybrid Euler 2S Sampler

Hybrid sampler: early Euler ancestral with order-2 denoised extrapolation for motion, late DPM-Solver++(2S) ODE for quality with progressive correction blending. Optional eta decay, anti-stiffness motion pulses, velocity bias, and reactive rescue. On rectified-flow models (CONST, e.g. LTXAV) it runs an RF-correct port of the same features (Heun corrector in place of 2S) instead of falling back to plain euler-ancestral.

By digital-garbage·Created about a year ago·Updated 24 days ago· 16
FunPack Hybrid Euler 2S Sampler
  • sigmas
  • sampler
  • sigmas
eta1.00
eta_final1.00
s_noise1.00
high_quality_pct0.35
correction_blend1.00
quality_sharpness0.00
motion_pulse_modeoff
motion_pulse_start_pct0.30
motion_pulse_count2
motion_pulse_spacing_pct0.22
motion_pulse_strength0.85
velocity_bias_modeoff
velocity_bias_strength0.00
velocity_bias_sourcemean
velocity_refinement_keydefault
rescue_modefalse
rescue_threshold0.15
rescue_strength0.20
normalize_strength0.00
normalize_start_sigma0.900
CategoryFunPack/Sampling

Inputs (21)

NameTypeDefaultDescription
etaFLOAT1.000–1Ancestral stochasticity at the start of sampling. Keep at 1.0 for classic ancestral behaviour.
eta_finalFLOAT1.000–1Eta value at the quality phase boundary. When below eta, ancestral noise decays linearly toward this value as sigma approaches the quality phase. Lower values give a cleaner hand-off into deterministic refinement. Set equal to eta to disable decay.
s_noiseFLOAT1.000–10Noise scale for ancestral noise injection.
high_quality_pctFLOAT0.350–1Fraction of late denoise steps that enter the quality phase. The first half of quality steps use single-eval Euler ODE; the second half use the full 2S correction.
correction_blendFLOAT1.000–1Blend between Euler ODE (0.0) and 2S correction (1.0) for the second half of quality-phase steps.
quality_sharpnessFLOAT0.000–1Restores fine detail lost to velocity-bias softening. Temporal-average unsharp on the x0 prediction during the quality phase only. 0 disables; 0.2-0.4 typical when velocity bias is on. Free (no extra model eval).
motion_pulse_modeCOMBOoffAdds early/mid anti-stiffness motion pulses. Off preserves legacy sampler behavior.
motion_pulse_start_pctFLOAT0.300–0.95Sampling progress where the first anti-stiffness pulse is applied.
motion_pulse_countINT21–6How many early/mid motion pulses to request before the late quality phase.
motion_pulse_spacing_pctFLOAT0.220.04–0.45Progress spacing between motion pulses.
motion_pulse_strengthFLOAT0.850–1How strongly motion pulses add monotonic noise kicks. Higher values push harder against stale image references.
velocity_bias_modeCOMBOoffExperimental: capture/apply averaged early model velocity around normalized sigma 0.9 and 0.8. Off preserves legacy sampler behavior.
velocity_bias_strengthFLOAT0.000–3Strength of the remembered velocity (action) injected at the structure sigma. 0 disables. ~0.15 = subtle spice; 0.3-1.0 = clear action crossover; 2-3 approaches full action replacement (capped so the current gen isn't wiped). Creative tool, not for consistency.
velocity_bias_sourceCOMBOmeanHow velocity bias / rescue pick a good direction. 'mean' = prompt-blind global average (legacy). 'nearest' = single best-matching prompt cluster — preserves one real good gen's detail instead of a washed-out average (less softening). Affects both apply and rescue.
velocity_refinement_keySTRINGdefaultMemory key used to capture/apply early velocity bias.
rescue_modeBOOLEANfalseReactive in-flight rescue, rating-gated. Steers each eligible step toward trajectories you rated good and away from ones you rated Awful (matched to the current prompt). Learns automatically from ratings while on — no separate capture step needed. A no-op until you've rated a few gens for this prompt/key (a positive rating builds the target, an Awful builds what to avoid).
rescue_thresholdFLOAT0.150–1Fires when the step has diverged from the good trajectory by more than this (1 - cosine) OR aligned with a bad trajectory by more than this (cosine). Lower = corrects more eagerly. 0.10-0.20 typical; raise toward 0.4+ for only severe cases.
rescue_strengthFLOAT0.200–0.5How hard to pull toward good / push away from bad when triggered (magnitude preserved, no energy injected). Keep moderate; 0.5 is a strong correction.
sigmasoptSIGMAS
normalize_strengthoptFLOAT0.000–1Video-only latent normalization (anti-overbake / oversaturation / colour drift) stacked on this RF loop. 0 = off. 0.5 = gentle. Audio is never touched (LTXAV/CONST path). ~zero overhead.
normalize_start_sigmaoptFLOAT0.9000–1Sigma at/below which latent normalization activates and anchors its reference. Only used when normalize_strength > 0.

Outputs (2)

NameTypeDescription
samplerSAMPLER
sigmasSIGMAS