XY Plot Pixaroma
Compare settings side by side in one labeled grid - no extra wiring.
What it does
Drop this at the end of your workflow and wire your final image into it, like a Preview node. Pick what changes ACROSS (X = columns) and DOWN (Y = rows), press Run once, and every combination fills a labeled grid right here on the node.
How to use
- Wire your workflow's final image into the
imageinput. - In the X card, pick a setting to vary across the columns. Do the same in the Y card for the rows (you can use just one axis if you like).
- Enter the values you want to try in the value box.
- Press Run once. The grid builds as each run finishes.
The value box adapts to what you pick
- The checklist shows ALL your installed options (every lora, sampler, checkpoint, etc.), not just the one in the node - that is how you compare against ones you have not loaded yet. The
now:line under the picker shows what the node is set to right now.
- Number
- A
Range(Start / End / Steps) or aListof values. - Dropdown
- A checklist of VALUES to try, not a list of things to load at once: tick the samplers / models / loras you want to compare and each tick becomes one square.
- LoRA
- A checklist of lora files, or a number when you pick a weight. Loras get a section of their own further down, because they work one level differently from every other setting.
- Prompt text
Full list(one full prompt per line) orFind & replace(swap a word for each value). Long prompts wrap onto several rows, and the numbers down the left count your VALUES, so you can always see where the next one starts. The numbers are just a guide - you cannot type in them.
Examples: what to pick for what you want
Two things to hold on to and the rest follows. The picker chooses WHICH setting changes. The value box under it chooses WHAT that setting becomes, and you get one square per value, not all of them at once.
So X and Y are how many settings can change at once, not how many things you can compare. One axis with 5 ticks gives you 5 squares, and you can go up to 100 on an axis. Use the second axis when you want a SECOND setting changing as well.
Two habits worth having. Watch the counter above the buttons, because it tells you how many images you just asked for and 4 across by 3 down is 12 runs. And leave Lock seed on unless you are plotting the seed itself, or two things change at once and you cannot tell which one did it.
- Which sampler suits this image?
- X: your KSampler's
sampler_name. Tick the 4 or 5 you actually use. One row of squares, one sampler each. - How many steps do I really need?
- X:
steps,Range, Start 10, End 40, Steps 4. That gives 10, 20, 30, 40, and you can see where it stops improving. - Steps against cfg, all at once
- X:
steps(try 15, 25, 35). Y:cfg(try 3, 5, 7). Nine squares showing every pairing, which is far quicker than nine separate runs. - Same prompt, one word different
- Pick your prompt text and choose
Find & replace. Put the word to swap in Find, and your alternatives in the values. Everything else in the prompt stays put. - Which checkpoint handles this prompt best?
- X: your checkpoint loader's
ckpt_name. Tick the models to try. TurnLock seedon so the seed is not also changing.
Comparing LoRAs, case by case
Loras trip people up more than anything else here, because they sit one level deeper than the other settings. Your lora loader has ROWS. An axis changes ONE row. The list you then tick is every lora on your computer, and each tick is one square. So the row you pick is WHERE the change happens, and the ticks are WHAT goes there.
- Just one lora in the loader
- The easy case. Pick
LoRA 1 fileand tick as many loras as you want to see. Each square runs exactly one of them, with nothing else muddying it. - Two or more loras in the loader
- The axis still changes only the row you picked. Every OTHER row that is switched on is applied in every square, which can hide the thing you are comparing. Switch off the ones you are not testing. If one would get in the way, an orange line under the picker tells you which.
- Keeping a lora in every square on purpose
- Sometimes that is exactly what you want: a style lora you always use, while you compare something else. Leave its row on and read the grid as A plus style against B plus style.
- What weight should it be?
- Pick
LoRA 1 strengthand enter a list like0, 0.4, 0.7, 1. The0square is your before shot, because a lora at 0 does nothing at all. - Which lora AND at what weight
- X:
LoRA 1 filewith your candidates. Y:LoRA 1 strengthwith a few weights. Every lora at every weight, in one grid. - Two loras mixed together
- X:
LoRA 1 file. Y:LoRA 2 file. Here you WANT both rows switched on, since every square is a real pairing, so no warning appears. - Three rows all changing at once
- Not possible, since there are only two axes and so at most two settings can change. Keep the third row fixed and switched on, or run a second plot.
- Trigger words
- When you swap the FILE on a LoRA Loader Pixaroma row, that row's ticked trigger words are left out of every square, because they belonged to one particular lora. Put any words you want in all squares into your prompt instead.
- Which entries you get
- With LoRA Loader Pixaroma each row appears as
LoRA 1 fileandLoRA 1 strength, counted from the top of the node, plusLoRA 1 clip strengthwhen model and clip are separate in the gear. The core Load LoRA node and other multi-lora loaders work too.
Entering numbers
- List: values separated by commas, e.g.
5, 6, 7.1, 10. Decimals are kept exactly. - Range: set Start, End and Steps (how many). 5 to 15 in 3 steps gives 5, 10, 15.
- Shorthand inside a list also works:
4-10 (+2)steps by 2, and4-10 [4]gives 4 evenly spaced values.
Buttons and options
- Lock seed
- Keeps the seed the same for every square so the only thing changing is what you're testing. Turns off on its own if you're plotting the seed.
- Draw labels
- Show the value labels and axis names on the grid.
- Save cells
- Also save each square on its own, not just the whole grid.
- Grid: Dark / Light / Mono
- The grid background and label style. Switching re-skins the grid you already have, instantly.
- Reset X / Reset Y
- Clear just that one axis.
- Reset XY
- Clear both axes and all selections, back to a fresh node.
- Save Disk / Save Output / Copy / Open
- Act on the finished grid: save it to your computer or to ComfyUI's output, copy it, or open it in a new tab.
Saving and image size
The grid shown on the node is a preview, capped at 4096 pixels on its long side so it stays light. The Save row picks how big the SAVED file is:
- The full-size grid is built only when you click Save, so choosing a bigger size never slows your runs down.
- Save Disk saves to your computer and Save Output goes to ComfyUI's output folder, both at the Save size. Copy and Open use the smaller on-screen preview.
- The grid sent out of the node's image dot (top right) always stays 4096, to keep anything wired after it fast.
- Want the preview bigger on the canvas? Make the NODE wider and it grows with it (and smaller again when you narrow it). A very tall grid stops growing at a point, so the node stays a size you can still work with - use Open for a proper look at that one.
- 2048 / 4096 / 8192
- Cap the exported grid to that many pixels on its long side.
- Full
- Export at native resolution, every cell at its real size. This makes the largest file.
What you wire in
The settings on the node
What comes out
This is the help that comes with Pixaroma 1.4.212. Inside ComfyUI, select the node and click the orange ? above it to read the same help there.