Sigmas CollatzIteration

This might be the most novelty node in the entire RES4LYF pack, and that's saying something given the company it keeps. Sigmas CollatzIteration runs your noise schedule through the Collatz conjecture - the famous unsolved "3n+1" problem from number theory.

What the Collatz conjecture is, quickly

Pick a positive integer. If it's even, halve it. If it's odd, triple it and add one. Repeat. The conjecture - still unproven, one of the most well-known open problems in mathematics - is that no matter what number you start with, you always eventually reach 1. It's a genuinely famous piece of math, which makes it a genuinely odd choice to bolt onto a diffusion sampler's noise schedule.

How it works

The node applies that even/odd iteration rule against your sigma values for iterations steps, with scaling_factor controlling how strongly each iteration moves the value, and normalize_output rescaling the (potentially chaotic, sequence-length-dependent) result back into a usable range.

The inputs and outputs that matter

• sigmas (SIGMAS, required) - the input schedule.
• iterations (default 3, range 1–20) - how many Collatz steps to apply.
• scaling_factor (default 0.1, range 0.0001–10) - how strongly each iteration affects the value.
• normalize_output (default true) - rescale into a usable range.

Output is a single SIGMAS list.

Should you use this?

I'll be straight with you: I don't have a recipe for this one, and neither does anyone else as far as I can find. There's no README coverage beyond the pack's general note that its sigma-manipulation category is "experimental and subject to further changes," and I checked whether anyone in the community has used or discussed RES4LYF's chaos/novelty-math sigma nodes at all - nothing came up, not for this one specifically and not for its siblings (Sigmas Attractor, Sigmas Chaos, Sigmas ConwaySequence). That's not a knock on the pack. RES4LYF ships a huge pile of sigma-math primitives because its actual premise is giving you raw material to build custom schedules, not because every single node in that pile is a recommended technique. This is a fun one to play with if that's your thing; it's not something you need to reach for to get a good image.

How to install it

• ComfyUI Manager - search "RES4LYF", install, restart.
• Manual - activate your venv, cd ComfyUI/custom_nodes && git clone https://github.com/ClownsharkBatwing/RES4LYF, cd RES4LYF, pip install -r requirements.txt (portable builds: use the embedded pip.exe). Restart.

Common issues & troubleshooting

Expect non-monotonic, possibly out-of-range output. The Collatz sequence for a given starting value can swing up before it swings down (that's the whole reason the conjecture is interesting), so there's no reason to expect the result respects "should decrease smoothly toward zero." Run it through Sigmas Abs and Sigmas Cleanup, and check step count and actual values with Sigmas Count and a preview, before it goes anywhere near a real render.

A low scaling_factor keeps the effect subtle; cranking it up amplifies exactly how erratic the underlying sequence is, for better or (more likely) worse.