Freebie Workflow and Detail Of this Tutorial: www.patreon.com/posts/110299136? XLab Custom Node github.com/XLabs-AI/x-flux-comfyui Controlnet-Depth-v3 huggingface.co/XLabs-AI/flux-controlnet-depth-v3 Controlnet-Canny-v3 huggingface.co/XLabs-AI/flux-controlnet-canny-v3 Controlnet-SoftEdge-v3 huggingface.co/XLabs-AI/flux-controlnet-hed-v3 Base Model Flux 1 Dev FP8 huggingface.co/XLabs-AI/flux-dev-fp8/tree/main
🤣 no , i said it on purpose, a normal tutorial is too boring. So sometimes a diss on company, sometime a name for a model. So people do pay attention will know the hidden message.
In the node HED Soft/Edge Lines you have set the resolution to 1024. My image is 1280 x 720, how can I set the resolution of the HED Soft/Edge Lines to 1280 x 720 or to get 1280 x 720 image?
Help ; Error : No module named 'controlnet_aux' (for img2img) (My comfyUI, custom_nodes are updated. I watched different controlnet tutos.) (I'm using Flux1-dev-Q4_K_S + comfyUI + Controlnet) There is a error during the node "Canny edge" : No module named 'controlnet_aux' I installed everything that I need (ComfyUImanager, Xlabs, ControlNet auxiliary models, Controlnet Canny) I don't understand why this doesn't work. If you have a solution :)
@@Nibbarese7 "upsample_bicubic2d_out_frame" not implemented for 'BFloat16' you didnt happen to come across this error did you when it got to the sampler?
@@TheFutureThinker I wish if we will get a Controlnet that goes into the conditioning like before, so we can stack them and also use it in more traditional KSamplers and UltimateSDUpscale, i'm abit confused with this method, I admit that i'm not understanding well these "true_gs" and "timestep to start cfg" parameters yet, but yeah. Well done keep going 👌
Freebie Workflow and Detail Of this Tutorial: www.patreon.com/posts/110299136?
XLab Custom Node
github.com/XLabs-AI/x-flux-comfyui
Controlnet-Depth-v3
huggingface.co/XLabs-AI/flux-controlnet-depth-v3
Controlnet-Canny-v3
huggingface.co/XLabs-AI/flux-controlnet-canny-v3
Controlnet-SoftEdge-v3
huggingface.co/XLabs-AI/flux-controlnet-hed-v3
Base Model Flux 1 Dev FP8
huggingface.co/XLabs-AI/flux-dev-fp8/tree/main
Am I The Only One Listening 'CANDY EDGE' 🔥
🤣 no , i said it on purpose, a normal tutorial is too boring. So sometimes a diss on company, sometime a name for a model. So people do pay attention will know the hidden message.
P s: one of the thumbnail I use a Candy Cake.
"Black Foreign Labs"
thanks! ControlNet V3 looks more clear.
Should be good for your images.
@@TheFutureThinker ok , i just try it. Depth and Canny. Work like before. But hands position pose different, 6 from 10 images, it happen.
Should be. Its not a Open pose or dw pose.
oh yeah! CN for Flux!
Yup its here :)
>comfy UI
Laughingmeme.jpg
>Stable diffusion A111/Forge
>Standingovationofclaps.jpg
Such a great video! I learn so much from your videos. Keep it up!
impossible with RTX 3060Ti, just 8 gb =( error: out of memory
Well , out of memory. 🤷
@@TheFutureThinker 🤣
@@TheFutureThinker resolution 512 x 512 worked well.
GTX 1080 - nice work. Possibly lack of RAM?
Could you please share the link to the t5-xxl-fp8?
In the node HED Soft/Edge Lines you have set the resolution to 1024. My image is 1280 x 720, how can I set the resolution of the HED Soft/Edge Lines to 1280 x 720 or to get 1280 x 720 image?
Can you link a Lora to this workflow?
how long does it take to render?
How to use in Forge?
I don't use Flux in Forge.
How do I install this in Forge UI?
hi great thanks, can you use Flux ControlNet and Lora ?
Yes you can, for Flux LoRA.
why UNET and not checkpoints folder? 😆
Help ; Error : No module named 'controlnet_aux' (for img2img)
(My comfyUI, custom_nodes are updated. I watched different controlnet tutos.)
(I'm using Flux1-dev-Q4_K_S + comfyUI + Controlnet)
There is a error during the node "Canny edge" : No module named 'controlnet_aux'
I installed everything that I need (ComfyUImanager, Xlabs, ControlNet auxiliary models, Controlnet Canny) I don't understand why this doesn't work.
If you have a solution :)
thanks for the video, clear explanation. you deserved a subscribed
Thanks
buddy,how to conver a flux lora to xlabs lora?
Great!
😊
Does anyone know where my face is? I make everything with ControlNet-Union-Pro
Like a cotton bag, the image is messed up...
HELPPP if I add the nodes Load FLux controlnet and flux Loras the list is empty, i have the additional_paths.yaml configured already, what can I do?
needs to be in the xlabs folder
@@Andy.. literally figure out 10 mins ago, anyway it's slow af, i had this image generating at standard settings for 2 hours and it's at 96%
@@Nibbarese7 "upsample_bicubic2d_out_frame" not implemented for 'BFloat16' you didnt happen to come across this error did you when it got to the sampler?
does this work with flux nf4v2 checkpoint?
Saw Reddit, someone said it works.
I tried the Depth ControlNet work with Depth Anything V2.
It does work with depth anything v2?
@@ryansenger937 just the preprocessor.
how to copy face from original? not replaced by ai?
use reactor to swap face
@@supernaturalsubcontinent6282 would love you to create a video of it too.. I subscribe to you man
when I use loras Ive trained they look terrible with depth for some reason.
I tested V2 and it was veeeery slow. Both with fp16 and 8. 3090.
3090, it happens. Its ok.
Will 4090 perform better ?
Cyberpank? LMAO, you mean Cyberpunk?
Cyberspank
Candy? Wth man 😉
Congrats, you are not day dreaming 👍
@@TheFutureThinker I wish if we will get a Controlnet that goes into the conditioning like before, so we can stack them and also use it in more traditional KSamplers and UltimateSDUpscale, i'm abit confused with this method, I admit that i'm not understanding well these "true_gs" and "timestep to start cfg" parameters yet, but yeah. Well done keep going 👌
@@crypt_exe8709 the conditioning can stack like before, sampling node not the same.
hwo to use openpose on this shit?
How to use in Forge?
i tried, it seems dont work yet in forge :S