Eden_IntToFloat
The type bridge that keeps your math from rounding behind your back
- FLOAT
Eden_IntToFloat takes an integer and returns it as a float. That's the entire node - float(a) - and like its sibling Eden_FloatToInt, it exists because ComfyUI's typing is strict in the places that matter. When a node expects a FLOAT input and you're feeding it an integer-typed value, this is the two-wire fix.
Why you'll actually need it
The case that bites people: you compute something with a math node and it comes out INT, but the consumer - a denoise slider, a strength input, a conditioning weight - only accepts FLOAT. You could retype the widget, but if the value is being computed or driven from a seed, there's no widget to edit; you need a typed conversion. This node is that conversion, and it's lossless. 7 becomes 7.0, no rounding, no surprise precision loss - ints up to the 64-bit max convert cleanly into double-precision floats.
Inputs and outputs
- a (
INT, default 0) - the value to convert. - Output FLOAT - the same value, now floating-point.
Nothing optional, nothing hidden. The schema is one required input and one output.
Where it fits in a workflow
Pair it with Eden_Math or Eden_Int as the source of an integer, and the outputs of that node go into any float-socket. It's also the natural bookend to Eden_FloatToInt - if you're building a round-trip (int β float for math β int for a consumer), these two keep the types honest at each boundary. In the Eden pack's own Logic lineup, the constants come in both flavors for exactly this reason: Eden_Int and Eden_Float next to each other, so you can pick the type the graph demands rather than fight it.
Installing it
Ships in the Eden.art nodesuite (edenartlab/eden_comfy_pipelines). Install via ComfyUI Manager (search "Eden.art nodesuite") or manually:
cd ComfyUI/custom_nodes
git clone https://github.com/edenartlab/eden_comfy_pipelines
cd eden_comfy_pipelines && pip install -r requirements.txt
Restart ComfyUI. Same note as every node in this pack: requirements.txt pulls a heavy stack (clip-interrogator==0.6.0, open_clip_torch, transformers, scikit-image, opencv-python), so a three-line converter comes with a full suite install.
Common issues
This one genuinely can't fail in interesting ways - no range violations to hit, no rounding to argue about. The only "issue" is the inverse one: if you're seeing a float where an int is required downstream, you've reached for the wrong direction, and Eden_FloatToInt (or rounding inside a math expression) is the actual fix. Know which type each socket wants and this node is boring in exactly the right way.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| a | INT | 0 | β |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| FLOAT | FLOAT | β |