To string (FLOAT)

The settings you change most - CFG, denoise, LoRA strength - are all floats, and the floats you can't see in the final image are exactly the ones you'll wish were recorded. To string (FLOAT) (RF_FloatToString) converts a float into text, which means you can bake cfg_7.5_denoise_0.6 straight into a filename or a graph note. It's the float-flavored member of the RF Nodes ToString family, and its whole purpose is making settings visible outside the node editor.

How it works

Plain str() conversion. FLOAT 7.5 in, "7.5" out. No rounding, no decimal padding - what you see is Python's default representation of the float. That's the one gotcha to know about: a float that's been through arithmetic can print as 0.6000000000000001 because that's genuinely what it is. If you want 0.6 in the filename, round the value before converting, or use a formatting node.

The inputs that matter

• value - a FLOAT, forceInput, so it must be wired from a node output.

Output: STRING - the float as text.

Where it earns its keep

Comparison and iteration. When you're sweeping CFG values across a batch, concatenating the float into the Save Image filename gives you a folder full of cfg_7.5.png, cfg_8.0.png, cfg_8.5.png - the whole experiment readable from the file listing, no metadata spelunking. Same trick works for denoise schedules and strength ramps. It's the same "self-documenting filename" pattern as the pack's int-to-string node, just for the settings that aren't integers.

Installing RF Nodes

No dependencies, no models. Via ComfyUI Manager search "RF Nodes", or:

cd ComfyUI/custom_nodes
git clone https://github.com/foxtrot-roger/comfyui-rf-nodes

Restart ComfyUI; it's under RF in the menu.

Common issues

• "Why does it print 0.6000000000000001?" - Float arithmetic, honestly reported. Round upstream if the label needs to be clean.
• "There are three float-ish ToString nodes." - RF_FloatToString, RF_NumberToString, and the ANY RF_ToString. For a guaranteed FLOAT wire, use this one; for a generic NUMBER, use the Number variant.