Thanks for watching!

@@AiVOICETUTOR thanks for sharing. Seriously

Glad you liked it!

You can find the code for Automatic1111 on GitHub: github.com/AUTOMATIC1111/stable-diffusion-webui

Thank you! Well in theory it should work by putting the existing image into a second ControlNet unit. Control weight set to 0.5 and use the depth preprocessor and depth model. Sadly I have not managed to get this to work properly though.

@@AiVOICETUTOR Thank you! Same here, just searching to solve this issue :)

Thanks I'm glad you like it. I'm not aware of any alternatives but you can use Motionleap on MacBook by installing an Android emulator (like BlueStacks: www.bluestacks.com) and then downloading Motionleap from the Play Store.

@@AiVOICETUTOR that's a cool trick, unfortunately not for me 😞. I am unable to install bluestacks on my Mac as it has new Apple silicon processor and it's not supported by Bluestack

Ahh sorry I didn't realize that Bluestacks X is something completely different. You could always try setting up Android Studio (Developer platform for Android) if you think that's worth the effort. Can find more info here: www.reddit.com/r/summonerswar/comments/u7gi4a/how_to_use_an_android_emulator_on_mac_with_m1_chip/

Thanks. Glad you like it

Maybe i have a Job for you. Can I contact you with a personal message?

Thanks bro! Glad you like it

Sure. You can contact me at aivoicetutor@gmail.com

Thank you!

Glad you like the video. Could you please explain what you mean by dynamic qr code? Do you mean that the code changes over time?

You can definitely run multiple controlnet models on collab. Not sure if that’s your issue but check the settings of controlnet (within the stable diffusion settings) to make sure the max number of models is higher than 1. If you don’t have that setting, you might need to update the webUI.

@@AiVOICETUTOR ok thanks for ur help

Hello, you’re right. They changed the inpainting in v1.60. Sometimes I need to drag n drop the image into the WebUI more than once before i can see the entire image. Instead of using a white background in Paint (or in Irfanview) to expand the image, you can use a black one and you will see the masking pen. If you do it like that, you don’t have to resize the image. Resizing is a different technique, just like using an outpainting script. Hope this helps

Hi, thank you very much! In theory it should work by using an additional controlnet instance where you upload the image to. I am not sure which settings to use though or if that’s even the right way to do it.

@@AiVOICETUTOR Thanks for answering. I have tried that condition but I still can't find the parameters to make it work. Greetings

I had another try at it but also couldn't find a solution. One day we'll figure it out :)

@@AiVOICETUTOR Thanks for trying. I saw an alternative on a discord channel. Although it is not the solution sought, it is an option. For example a Teddy bear holding a painting. pose achieved with openpose, and then make the generated qr individually. To achieve color consistency with the entire image, the color grid option can be used in controlnet. Then join the qr to the image with an image editor or with impaint. It looks good but it is far from the sought solution. Any news I will let you know here. kind regards.

Thank you very much for sharing that technique! I'll keep you up to date on it as well. There's an alternative to Controlnet now (Unicontrol) that I haven't looked into yet. Maybe combined with SDXL we'll get there? Cheers

Yes check this out: github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Installation-on-Apple-Silicon

Hi the codes scan on my iPhone with the default camera app. Third party apps work even better. On Android there seem to be major differences between the devices. Try reducing the distance to the screen when scanning.

Hi you need a working installation of Stable Diffusion github.com/AUTOMATIC1111/stable-diffusion-webui and Controlnet (kzfaq.info/get/bejne/j5eijLef2ta-m2g.html). Hope this helps

thank you. will try first. @@AiVOICETUTOR

Well in theory it should work by putting the full screen image into a second ControlNet unit. Control weight set to 0.5 and use the depth preprocessor and depth model. Sadly I have not managed to get this to work properly though

Download the model from huggingface.co/lllyasviel/ControlNet-v1-1/tree/main and put it in stable-diffusion-webui/models/ControlNet (and then refresh the list of models or restart the webui)

Hey glad you liked it. I use Motionleap to animate the codes 😀

thank you for your answer 😊

I‘m using A1111 and it has Inpainting, as you can see in the video. I haven’t used Thinkdiffusion myself but from a quick google search, it seems like it should support Inpainting too. Maybe you can update your instance to get it to show up?

@@AiVOICETUTOR Thanks for the reply ! I figured it out i didn't click img to img lol. I have another question. How can we inpaint the logo in a way that it will be transparent ? Because now when i do inpaint , i have letters under the logo also, there is a white background.

Not sure about the transparency but habe you tried putting „text“ and „letters“ in the negative prompt?

@@AiVOICETUTOR No i mean that my logo has text under. Anyway i created the qr without the logo and i put it after.

Sorry I missed your reply. You could have added some words such as “text“ or “letters“ to the negative prompt. But sounds like you figured out a workaround 👍

Unfortunately, from what I read, Samsung seem to be among the phones that have the worst QR code recognition. Maybe try a third party QR scanner from the Google Play Store.

Thanks for the feedback! It's interesting, that while they work across all iOS devices, on Android there seem to be many differences between the brands and their devices.