For the lazy ones like me, here are the starting settings shown : Check ENABLE ControlNet Check PIXEL PERFECT PREPROCESSOR : Invert MODEL : QRCode Monster CONTROL WEIGHT : 1.25 STARTING CONTROL STEP : 0.15 ENDING CONTROL STEP : 0.75 Thanks OP for sharing this video ! :)
I show Motionleap in this video: th-cam.com/video/faETMOvFUq8/w-d-xo.html but only briefly since it is very easy to use. The effect I used in CapCut is called "3d zoom" (you can find it under "style")
I'm trying to run this and I'm getting "IndexError: pop from empty list" underneath the generate. If I unselect a model under ControlNet it works, but the text doesn't show through. Any ideas?
That’s the only hit I found for your specific error: github.com/AUTOMATIC1111/stable-diffusion-webui/pull/1273 Is it possible that you’re also using special characters in the prompts?
what PC Specs do you use? I use rtx 3060 laptop (6gb vram) ram 16gb processor i7 11th, and the outputs is "CUDA out of memory..." , when I check Low VRAM, generated images looks bad, very bad
I'm running this on a 4090 (24GB) and i9. Before I upgraded, I used a GPU with 8GB of RAM and it was running out of memory too every now and then. You can try lowering the resolution of the images and then upscale them later
If you're missing the preprocessors then you should do a clean install of controlnet. If you're only missing the models, then you can download them at huggingface.co/lllyasviel/ControlNet-v1-1/tree/main
same happens to the picture if the "invert" preprocessor is used. Everything except the text will be very dark "influenced by the black background of the text picture" while the text will be super bright.
@@AiVOICETUTOR What does the denoiser in text2image do? I thought it only applies to hiresFix. Regardless, any setting produces the same "error" kind of frustrating.
Yeah you shouldn't need to change it but I thought its worth a shot with what you described. I've once had an issue with A1111 where nothing was working anymore as it should have and I ended up reinstalling it. If everything else is working for you in A1111 though then it might just be some random bug with Controlnet
For the lazy ones like me, here are the starting settings shown :
Check ENABLE ControlNet
Check PIXEL PERFECT
PREPROCESSOR : Invert
MODEL : QRCode Monster
CONTROL WEIGHT : 1.25
STARTING CONTROL STEP : 0.15
ENDING CONTROL STEP : 0.75
Thanks OP for sharing this video ! :)
Thanks for posting this summary :)
*great ! got to follow it !*
Helped alot, thanks!
Awesome. Thats great to hear!
Really insane man. ❤❤❤
Thank you ❤️
Very cool... I figure that using a mask produces even more crazy fonts transfers
Thanks. Not quite sure how you'd use the mask. Could you elaborate that a bit?
Really nice - thank you for sharing!
Glad you enjoyed it!
wonderfull,,,,,is there a video teaching how to make animated image?
I show Motionleap in this video: th-cam.com/video/faETMOvFUq8/w-d-xo.html but only briefly since it is very easy to use. The effect I used in CapCut is called "3d zoom" (you can find it under "style")
I'm trying to run this and I'm getting "IndexError: pop from empty list" underneath the generate. If I unselect a model under ControlNet it works, but the text doesn't show through. Any ideas?
That’s the only hit I found for your specific error: github.com/AUTOMATIC1111/stable-diffusion-webui/pull/1273 Is it possible that you’re also using special characters in the prompts?
Is there a way to use qrmonster with a existing image. Kinda like inpainting but getting the qrmonster result on a image you choose.
I have tried a couple of things so far like adding another controlnet but without any success. Someone will figure out a way to do it eventually
In ComfyUI?
I plan to make a video about comfy and controlnet soon
what PC Specs do you use?
I use rtx 3060 laptop (6gb vram) ram 16gb processor i7 11th, and the outputs is "CUDA out of memory..." , when I check Low VRAM, generated images looks bad, very bad
I'm running this on a 4090 (24GB) and i9. Before I upgraded, I used a GPU with 8GB of RAM and it was running out of memory too every now and then. You can try lowering the resolution of the images and then upscale them later
@@AiVOICETUTOR ah I see, looks like VRAM under 12GB is not enough for playing or training serious AI this day.
Yeah can't have enough VRAM these days. If you look at chatbots, not even 24gb is enough to run the largest models.
How do I get my controlnet to have all those options because I don't
If you're missing the preprocessors then you should do a clean install of controlnet. If you're only missing the models, then you can download them at huggingface.co/lllyasviel/ControlNet-v1-1/tree/main
@@AiVOICETUTOR do I download all of this?
You only need to download the ones you actually want to use
@@AiVOICETUTOR how do I know which ones I need?
Well for this tutorial you only need the one model that I'm using (huggingface.co/monster-labs/control_v1p_sd15_qrcode_monster/tree/main)
When I try it, my "white" background in the controlnet picture basically affects the whole picture, making it whiter and blurrier. Very strange.
same happens to the picture if the "invert" preprocessor is used. Everything except the text will be very dark "influenced by the black background of the text picture" while the text will be super bright.
That sounds like maybe you have a different value set for the Denoiser? If not maybe try to increase the Denoiser a bit and see if that helps
@@AiVOICETUTOR What does the denoiser in text2image do? I thought it only applies to hiresFix. Regardless, any setting produces the same "error" kind of frustrating.
Yeah you shouldn't need to change it but I thought its worth a shot with what you described. I've once had an issue with A1111 where nothing was working anymore as it should have and I ended up reinstalling it. If everything else is working for you in A1111 though then it might just be some random bug with Controlnet
not working for me :/
Sorry to hear that. If you tell us what exactly isn't working, we might be able to help you
😑😑my face the whole video
May I ask what you didn't like about the video?
@@AiVOICETUTOR I mean I squint my eyes to read the "hidden" texts
Haha right. I deffo need to up my emoji reading skills. Thanks for clarifying
Same here 😑😑😑😑😑😑😑😑😑