LTFloat_Step_02
A Coarser Everyday Step for Big-Knob Sweeps
- FLOAT
- STRING
LTFloat_Step_02 is the 0.2 member of the Latent Tools step-helper family - the "don't make me run fifty images" option. It's a calibrated constant: type a value, get the same number back as a FLOAT and a STRING, with the widget stepping by 0.2.
These nodes exist to sweep a parameter across a batch of images with the seed locked, and at 0.2 per run you get through a range fast. A cfg sweep from 5 to 11 is thirty runs; a denoise exploration is a handful. The catch is granularity: at 0.2 you'll see the overall shape of the response but you'll miss a sharp peak that sits between two points. Use it to establish the broad curve, then switch to the 0.1 or 0.05 sibling to map whatever looks interesting. That coarse-to-fine pattern is the whole game with this family.
How it works
The name implies it increments, so let's be precise: the node does not auto-increment. It takes the value you give it, rounds to five decimals, and returns it as a FLOAT and a STRING. The step is the widget's spinner increment and the label. The incrementing is yours - bump between runs, change nothing else.
The inputs and outputs
value- FLOAT, default 0, range ±1,000,000, step 0.2.FLOAT- the value.STRING- the value as text, for filename prefixes so each output records its parameter.
FLOAT into any float input (cfg, denoise, mean, std), STRING into a Save Image filename, seed fixed, compare.
Installing it
Part of xl0's Latent Tools pack. ComfyUI Manager → search Latent Tools → install → restart, or:
cd ComfyUI/custom_nodes
git clone https://github.com/xl0/latent-tools
Restart, and it's under LatentTools. The pack's only dependency is lovely-tensors (Manager installs it; manual cloners may need pip install lovely-tensors). No model downloads.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| value | FLOAT | 0.0-1000000–1000000 | — |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| FLOAT | FLOAT | — |
| STRING | STRING | — |