Output min/max sigmas
The exact noise range your model actually lives in
- model
- Sigmas_max
- Sigmas_min
Two numbers, straight from the checkpoint
Every diffusion model has a fixed range of noise levels it was trained with - a minimum sigma and a maximum sigma that bound its entire denoising schedule. ComfyUI reads those from the checkpoint's model sampling object, but it doesn't exactly hand them to you in the UI. Output min/max sigmas from sigmas_tools_and_the_golden_scheduler is the node that surfaces them: give it a model, and it returns two FLOAT values, Sigmas_max and Sigmas_min, in that order.
That's the entire node. One input (model), two float outputs. No knobs, no toggles, no schedule generation.
Why you'd want to know
Two genuinely useful reasons. First, the manual scheduler in this same pack lets you write formulas using sigmax and sigmin, and it reads those from the model internally - this node gives you the same values to see, so you can sanity-check a formula before you run it, or just know what range your model works in. Second, it makes cross-model comparison concrete: flow-matching models and SD-class models live in quite different sigma worlds, and a pair of numbers makes that obvious fast. Wire both outputs into text/display nodes and you've got a readout sitting in your workflow.
It's also a nice reality check on the pack's own schedulers. The Aligned Scheduler and the Golden Scheduler both derive their curves from the model's real min/max rather than published defaults - the author makes a point of that in the README. This node is the way to see exactly what those bounds are.
The honest take
It's the quietest node in the pack - pure introspection, nothing creative about it - and its value is almost entirely diagnostic. The outputs are plain floats, so if you find yourself needing a model's sigma range inside a formula or an automation, this is the missing piece. But it's also one of those nodes you'll install, use once to satisfy curiosity about your checkpoints, and mostly leave alone after. That's fine. It does one thing correctly, and there's a place for that.
Install & gotchas
ComfyUI Manager → search sigmas_tools_and_the_golden_scheduler, or:
cd ComfyUI/custom_nodes
git clone https://github.com/Extraltodeus/sigmas_tools_and_the_golden_scheduler
Restart, no model downloads. Note this node isn't even mentioned in the pack README - it's only discoverable in the source - which tells you it's the least-advertised member of the family. Pack-level gotcha still applies: requirements.txt lists only asteval while the code imports matplotlib and scipy, so pip install matplotlib scipy asteval if a minimal install errors on import.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| model | MODEL | — |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| Sigmas_max | FLOAT | — |
| Sigmas_min | FLOAT | — |