Great, you explain the workflow so well, thank u! I was doing this manually jumping between 3D and SD. This tool makes it so much easier and interactive.
Suggestions: as soon as possible add the ability to use Loras and a set default positions for a set number of cams. Training your own Loras for it would greatly help with fast and or constant projections.
In the previous versions, the black color meant "erase the mask" and the white was "restore the mask". But nowadays we see the big - and + on the vertical ribbon :)
Weird question, would it be possible to convert a diffuse map into a normal map or height map / bump map? And I dont mean dark colors = lower and light colors = higher, but like intelligently, sort of like AI assisted bump from the diffuse where it understands lighting so it wont make simple mistakes like turning light and shading into raised bumps
It does contain SF3D which generates 3d models from 2d backgrounds. But it's experimental, install is rather lengthy and not all people got it working yet. Have a look in the 3D panel --> SF3D tab
You can either use Symbolic link (folder shortcut) , to point towards your folder where the models are. or if you mean that you already have Forge or A1111 webui you can use it: see #custom-webui in our discord. For symbolic link you need to point the stable-diffusion-webui-forge/model/stable-diffusion and also folders like /lora and /VAE towards your existing folders with them
is this still strictly SD 1.5 or is it able to take SDXL models if the correct controlnets are paired? I know controlnets on SDXL arent the best but the SDXL resolution is better. amazing work
Yes, have a look in our dirscord. People got it working with Bottles and Wine. You could also run webui stable diffusion on a separate linux machine and connect via IP+Port from windows (bottom right corner in StableProjectorz)
when trying to use this it said it has an update for 2.0.2. but when i download it my browser blocks it as dangerous. is version 2.0.2. legit? i havent seen any youtube videos about it
Yes, that's because I didn't publish it on a large store, but distributing myself. Virus checks are fine, see here: www.virustotal.com/gui/file/e6d83685743690ca04d9511a70eedda07df69a0cd4484cfaf233cbb1d850cc90
Click the Dropdown and on the top it will show CivitAI link. Get a neural net, then paste it into stable-diffusion-webui-forge/webui/models/stable-diffusion folder. Or into LoRA folder if your network is LoRA. For SDXL neural nets see #sdxl in our discord
Очень круто! Благодарю за урок и твою работу. Скажи пожалуйста, а есть ли в планах возможность загружать свою модель с текстурой и на основе ее(img2img) делать улучшения текстуры или repaint. Как это можно сейчас делать в 3d cms ai - ruclips.net/video/14WmIV2J9co/видео.html
Технология уже создана. Одна из моих задач - обеспечить бесплатный доступ. Для нас, искуственный интеллект существенно упрощает работу, позволяя выполнять более масштабные проекты. Например, создать модели целого города, вместо дома.
It's beautiful, but it will remains useless if you don't remove all these baked shadows and light. The render is cool, but it's useless for any production. Just useful to lie on your portfolio.
This requires extraction of Albedo and Specular maps. Currently it's diffuse, but it's already used by most users to make hand-painted assets, like on website. So the Albedo+Specular is on the roadmap, once I manage to tackle the UDIMs and a few other things :)
Great, you explain the workflow so well, thank u! I was doing this manually jumping between 3D and SD. This tool makes it so much easier and interactive.
Please don't get acquired by some 3D software company.
I love that this is so user friendly and uses Forge! ❤❤❤
Я с хабра. Просто вау. Продолжайте работу 😊
Now even if you are bad at drawing, you can make a cool looking game.
Just sculpt an arrow and make it look like a spaceship... Very cool !
Who has the best ideas makes the best game - not the guy who is best in drawing.
@@mowax74 ah yes, ideaman
Incredible work you've done so far, going to try this out soon! Good eye for UX too, really well done!
Awesome to see the progress
Awesome work! This is a great soft
Nice I can't wait to try the once the SD3 Fine tunes are out.
Супер, спасибо большое за твою работу!!!
tyvm for the amazing work!
Great work bro :)
Great job!
Suggestions: as soon as possible add the ability to use Loras and a set default positions for a set number of cams.
Training your own Loras for it would greatly help with fast and or constant projections.
You can already use lora
See #lora in our discord
Это гениально, молодец 👏
oh my god , this is amazing
Wow that is a great tool!! 😮❤
Amazing
Welcome to the future
Can you create 3D printable files.
Great!
Incredible work with this! I cannot wait to use Stable Projectorz to help me make unity games.
I didn't get the difference between that "white color" / "black color" thing 🤔
In the previous versions, the black color meant "erase the mask" and the white was "restore the mask". But nowadays we see the big - and + on the vertical ribbon :)
It looks amazing! May I use this if I use Easy diffusion? Thanks
I think this will only work with Automatic1111 or Forge webui. But someone already made a bridge for Comfy, as a github repo
Weird question, would it be possible to convert a diffuse map into a normal map or height map / bump map?
And I dont mean dark colors = lower and light colors = higher, but like intelligently, sort of like AI assisted bump from the diffuse where it understands lighting so it wont make simple mistakes like turning light and shading into raised bumps
It's not in the current version of StableProjectorz, though I think Alchemist Designer had this feature. It should be completely doable by AI
Does it create 3d models from image/text input - or is it only for texturing existing 3d models?
It does contain SF3D which generates 3d models from 2d backgrounds. But it's experimental, install is rather lengthy and not all people got it working yet. Have a look in the 3D panel --> SF3D tab
@@IgorAherne Thanks! Viewed several videos but none explained exactly what this beast can do :)
is anyone else making more videos?
Верно ли я понял - создается только диффузная текстура ? Все остальное - карта нормалей, roughness и т.д. -отсутствует?
да, пока что только дифуз. Возможно добавлю альбедо и дополнительные текстуры, для PBR
I have so many checkpoint models already download, how do I get this to use my local installs?
You can either use Symbolic link (folder shortcut) , to point towards your folder where the models are. or if you mean that you already have Forge or A1111 webui you can use it: see #custom-webui in our discord.
For symbolic link you need to point the stable-diffusion-webui-forge/model/stable-diffusion and also folders like /lora and /VAE towards your existing folders with them
@@IgorAherne I used command line parameters. I don't like symLinks.
Thanks for making this! One dumb question I have is, how do you export the textured OBJ file at the end?
You can use the same obj you imported into the StableProjectorz.
So keep using your original obj + use the .png produced by the StableProjectorz
will there be a Linux version ?
is this still strictly SD 1.5 or is it able to take SDXL models if the correct controlnets are paired? I know controlnets on SDXL arent the best but the SDXL resolution is better.
amazing work
Yes, it works very well, we had it since around version 1.1 and fixed a few bugs related to it. Have a look at #sdxl in the discord
Я так понял программа делает только basecolor. Можно ли таким образом получить PBR текстурки ?
Пока что только диффуз, но скоро буду работать над извлечением альбедо, металлик и тд, для pbr.
Сейчас добавляю поддержку udim
@@IgorAherne Это пять 👍
верю ваши планы осуществятся. И также надеюсь, что это будет как для стилизации так и для реалистичного жанра.
I have problem in first step installing process, it has error in this step : receiving objects
please post on our discord + attach a screenshot.
I'll help or someone from community can
@@IgorAherne send your discord link plz
@@armandadvar6462 discord.gg/aWbnX2qan2
❤
Does it work with AMD graphic cards?
@@Kapthos yes, but you need a GPU with 10+ gb vram. Have a look in our discord, in #amd-gpu
Does this work on linux as well?
Yes, have a look in our dirscord. People got it working with Bottles and Wine. You could also run webui stable diffusion on a separate linux machine and connect via IP+Port from windows (bottom right corner in StableProjectorz)
when trying to use this it said it has an update for 2.0.2. but when i download it my browser blocks it as dangerous. is version 2.0.2. legit? i havent seen any youtube videos about it
Yes, that's because I didn't publish it on a large store, but distributing myself. Virus checks are fine, see here:
www.virustotal.com/gui/file/e6d83685743690ca04d9511a70eedda07df69a0cd4484cfaf233cbb1d850cc90
How do I add a new AI model?
Click the Dropdown and on the top it will show CivitAI link. Get a neural net, then paste it into stable-diffusion-webui-forge/webui/models/stable-diffusion folder.
Or into LoRA folder if your network is LoRA. For SDXL neural nets see #sdxl in our discord
@@IgorAherne Thanks for a quick response!
@@IgorAherne Ok. Now I have another question, how do I resolve "Not Connected yet. Check Black Window"?
I resolved it by starting the app as Administrator. Thanks to Discord channel :D
how to add custom image models
Top of Art panel add custom 2d image.
Or to add custom neural net, click More on the left panel.
⭐⭐⭐
Очень круто! Благодарю за урок и твою работу. Скажи пожалуйста, а есть ли в планах возможность загружать свою модель с текстурой и на основе ее(img2img) делать улучшения текстуры или repaint. Как это можно сейчас делать в 3d cms ai - ruclips.net/video/14WmIV2J9co/видео.html
Конечно, это можно уже!
Наверху панели Art, выбирай кнопку Custom Texture, затем делай как в видео, Step 3 или Step 4
@@IgorAherne Обалдеть! Благодарю за быстрый ответ =) Сейчас попробую.
Очень неприятно, когда твой труд, а возможно и дело всей жизни - удешевляется.
Технология уже создана. Одна из моих задач - обеспечить бесплатный доступ. Для нас, искуственный интеллект существенно упрощает работу, позволяя выполнять более масштабные проекты. Например, создать модели целого города, вместо дома.
It's beautiful, but it will remains useless if you don't remove all these baked shadows and light. The render is cool, but it's useless for any production. Just useful to lie on your portfolio.
This requires extraction of Albedo and Specular maps. Currently it's diffuse, but it's already used by most users to make hand-painted assets, like on website. So the Albedo+Specular is on the roadmap, once I manage to tackle the UDIMs and a few other things :)
Looked neat at first, but way too complicated.
We're not yet at the point that we can just magically ask for a high quality texture and get it, it still takes some work.