Hello there and thanks for checking out this workflow!
What's new in v27? : Major rework with new SubGraphs to streamline and compact the workflow even more, removal of outdated/broken nodes + fixes and optimizations all over!
—Purpose—
Built to provide an advanced, versatile and modular workflow for Flux with focus on efficiency, structure and information.
It comes with many notes explaining node settings and recommendations, as well as general guides from instructions to troubleshooting.
—Features—
Convenient loaders for all common versions of Flux and Clip models
Full metadata; recognized by CivitAI
LoRA support
SageAttention, EasyCache + Model Compile acceleration
ControlNet with Union support
Flux Tools LoRAs Canny + Depth (alt. to CNet based on LoRA)
Flux Redux (similar to IPAdapter)
PuLID (SVDQuant version only)
Wildcard prompting
Installation and download guide for models and nodes
multiple passes with optional upscales
— 1st : Detail Daemon + Variation Seed
— 2nd : DD. + Tiled Diffusion / UltimateSDUpscale
— ADetailer with dedicated LoRA Loader
— Inpainting
—Custom Nodes—
ComfyUI-nunchaku — SVDQuant version only
All of which can be installed through the ComfyUI-Manager
—Troubleshooting—
If nodes show up red (failing to load), check the 'Install Missing Custom Nodes' tab of the ComfyUI Manager for the missing node packs and install them.
Please check if all custom node packs load properly after installing, i.e. no
(IMPORT FAILED)messages next to any of them in the console upon ComfyUI startup.Always reload/drag'n'drop the original, downloaded workflow file into ComfyUI to reload an intact version of the workflow.
→ The last opened workflow that appears on startup shows a cached version of the workflow, "remembering" group nodes that failed due to missing nodes as failed, keeping them broken even after having everything installed correctly.
—Thanks—
The workflow would not be possible as is without these custom node packs. If you want to support the custom node creators, give them a ⭐ on their github repos! Thank you!
Feel free to ask questions, share improvements and suggestions in the comment section!
Let me know if you encounter confusing points I can elaborate on in the next update!
Description
v19
— removed Purge VRAM section
— many structural + layout changes
v19.1
— fixed detailer nodes after a breaking update
FAQ
Comments (12)
Hi, i tried follow the steps and it seems to almost work. I get the message that all the nodes are loaded succesfully. But it seems like i cant find the right values to use to make the workflow run or there is something wrong with the python of the math script. Any help would be appriciated :)
I'm using an rtx 2080 ti and windows 11.
Hey there, I believe this issue is caused by a custom node pack that is accidentally found when looking for missing nodes, probably due to having a common internal node name shared with the actual node pack I use.
The culprit would be this pack "ComfyUI-LatentClamp" which is NOT RIGHT and not used
Whereas this node pack "ComfyUI-Allor" is used and has a node called LatentClamp, which does not even make sense as it does not clamp anything and acts as a mere passthrough
So if you have "ComfyUI-LatentClamp" installed, get rid of it and make sure you have the "ComfyUI-Allor" pack 👌
Please report back if that resolves the issue 😊
@redpinkretro Thank you for your quick reply, it works now.
I don't remember "ComfyUI-LatentClamp", i think it got installed when i pressed install missing notes in the manager.
Not sure if it's just me but I updated comfy and my custom nodes and now the Adetailer AIO grouped nodes you created is borked. It looks like a node(s) updated with an additional input and shifted the default workflow values down a position. Nodes in the group are getting input values meant for other nodes which is easily corrected for the visible ones but there are many hidden ones. Love your workflows btw and this one was working greater prior to this
Hey there, did you download the newest version today?
That issue should be fixed in 19.1, and yes it was exactly as you said, affecting both the ADetailer as well as the Inpainting group nodes, because of tiled encoding/decoding options added by the Impact Pack, which I hid, as they don't serve much purpose within this workflow.
Oh, didn't notice the update to 19.1. Yup, all good now. Thanks!
@o0Maikeru0o That's a relief! Thought for a second another breaking update was pushed on the same day 😅
File "D:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Custom-Scripts\py\math_expression.py", line 241, in evaluate r = eval_expr(node) ^^^^^^^^^^^^^^^ File "D:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Custom-Scripts\py\math_expression.py", line 204, in eval_expr raise NameError(f"Name not found: {node.id}") NameError: Name not found: euler
Trying to run darkpictures nymphs aio but getting math expression pysss no name euler.
I just went through installing everything with a fine tooth comb because that demo node wasn't installed, but now this :( plz send help. Same thing happens in v18 but "LatentClamp.op() missing 1 required positional argument: 'samples'".
Hey there :)
I updated the workflow to 19.1 yesterday, as the ComfyUI-Impact-Pack had an update that broke my ADetailer- and Inpainting group nodes, which are fixed now in v19.1
When you have all node packs installed, up to date and loading properly—i.e. no (IMPORT FAILED) messages next to any of the custom node packs in the console log upon ComfyUI startup—and you still get errors, especially anything mentioning "LatentClamp", make sure you DO NOT have the ComfyUI-LatentClamp node pack installed, which is picked up on by accident by the missing-node-search because of a shared internal node name with one of the ComfyUI-Allor node pack that is actually in use.
I also see a NodeNotFoundError for the gguf loader node ID, which uses the ComfyUI-GGUF node pack, so there may or may not be something there as well.
Well I think everything should be running now but instead I got a Pulid faceloader error because my python embed is running 3.12.7 and apparently https://docs.google.com/document/d/1Hjf1LwpEy2KVmKb0TU4cjkzIofdi6tCP7qI7Sr6NtZs/edit?tab=t.0 this says that I can't be running 3.12? But I did get v18 working so there's that. ;-; augh I am so exhausted trying to figure this out, but thank you for helping. I got v19.1 to work by bypassing things, also I realized the alternate clip L was causing nonetype device error.
@EcstaticKalliope Did you replace the pulidflux.py file with the (still -.-) inofficial fix after updating all nodes? I have to do that every time, as the update overwrites the file with the outdated one from the repository.
So we have to go through this annoying procedure every time we dare that "Update All"-leap-of-faith.
This is the file btw. https://github.com/user-attachments/files/18266800/pulidflux.zip the "date modified" should read 12.28., so it was the very same pulidflux you may have used in the last weeks, as there weren't any updates. That's the one that runs for me, my embed's on python 3.10.9 though
Did you get that nonetype error with both selections cuda and cpu? If you get it with cuda, it might have to do with your version or something like a missing PATH variable to the location of it.
@RedPinkRetro Yeah I had to swap out to the 'clip l' I always used prior. Pulid still won't work, unfortunately. I also started getting errors with LLM because I was dumb enough to 'repair' my python install lol. I think I'mn gonna have to give up for now because I've been doing this for hours. Thank you for helping! I am a terrible newbie at this so it's all quite overwhelming. ^^;; I think I should downgrade python but atm I'm worried that will break everything so I might try again another day. <3
















