Scheduler Selector
One Dropdown, Two Outputs, Zero Surprises
- scheduler
- scheduler_name
Scheduler Selector is the quiet half of the Rapid Scheduler Combo node - same idea, one dropdown instead of two. You pick a scheduler, and it hands you the same value twice: once as a typed scheduler output that wires into nodes which accept scheduler values, and once as a plain string so you can read it, log it, or feed a string input.
If that sounds redundant, it is, on purpose. Different nodes in the rapidfire ecosystem want the scheduler in different shapes, and this node is the adapter that lets you choose the scheduler in one place and push it everywhere without fighting input types.
How it works
Under the hood it's nearly nothing. The scheduler input is a dropdown over ComfyUI's KSampler.SCHEDULERS list - simple, sgm_uniform, karras, exponential, ddim_uniform, beta, normal, linear_quadratic, kl_optimal - and the node returns (scheduler, scheduler) as its two outputs. The first output is typed as a valid scheduler so it plugs straight into sampler nodes that take a real scheduler input; the second output, named scheduler_name, is a wildcard string you can wire anywhere or just eyeball.
What you actually set
One input, nine choices:
- scheduler - the noise schedule. Quick guide from the concepts doc: Karras is the classic default for SD 1.5 / SDXL DDPM models, but it's actively wrong on flow-matching models - if your checkpoint is one of those, simple or sgm_uniform is usually safer.
Outputs: scheduler (typed) and scheduler_name (string), both carrying the value you picked.
Install and when to use it
Ships with comfyui-rapidfire, no models or heavy deps involved - it's pure wiring. Get the pack through ComfyUI Manager (search "comfyui-rapidfire") or:
cd ComfyUI/custom_nodes
git clone https://github.com/Zeks/comfyui-rapidfire
then restart.
Honestly, if you're deciding between this and Rapid Scheduler Combo, the Combo is the one to reach for in a normal rapidfire graph because it also carries the sampler. This node earns its keep in the rarer setup where you only want to centralize the scheduler, or where you need the scheduler name as a plain string for logging and a separate node already owns the sampler. It's a tiny tool, but it's the right shape for that job.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| scheduler | COMBO | 9 options: simple, sgm_uniform, karras, exponential, ddim_uniform, beta, +3 |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| scheduler | simple,sgm_uniform,karras,exponential,ddim_uniform,beta,normal,linear_quadratic,kl_optimal | — |
| scheduler_name | * | — |