I have been using open art for about 3 months now, and have not watched any videos on making AI art. That being said I have never done any editing and have over 16,000 published pictures. This video was incredible informative and I thank you kindly for making it. 💚🙏🌞🙏💚
Could the next tutorial go into detail about the 'mask blur' and 'mask grow' sliders? With this new version I get a lot of weird artifacts between the area being edited and the unedited areas. Sometimes it actually looks like a 'torn paper' edge especially if the original background is plain. For example, if there is a person against a plain background and I use magic selector to highlight their shirt and change it to a different style there will be a weird artifact ring that extends well beyond the highlighted area. Any possibility of a toggle between this version and the previous?
Hi! Good idea for the tutorial. For the problem you pointed out about plain background, it s a flaw of the current inpainting we found out too. We have a fix we will launch in the upcoming week! Stay tuned. I will ping here once we introduce the fix. Thanks!
Thanks for the feedback! Our team just put in some fix on the issue you raised. Now the "torn paper" edge effect should be minimized. Please try and let us know. Thank you again ❤
The popup that leads to this tutorial fills the entire screen and I couldn't find a way to get rid of it so I can *actually use the website.* If there was an X button on the window, it was way out of frame. I literally had to use an element blocker to get rid of it. So, thanks for that. :/
Sometimes when uploading an image to edit, the two cat tutorial ad, takes over the WHOLE screen no matter what you do. Hope you guys can look into that.
Is there any way to not get downscaled version? Resolution drops, and if I use the upscaler, it just adds to a perfect image to many 'creative' details which i do not want at all
Hi, really really great question, I need to add a note on this! You're using creative upscale. You can use the new ultimate upscale - precise upscale! It doesn't change the image creatively. The editor needs a lot of generation steps, so we downscale to keep it fast.
I am having a major problem here. I have a character whose feet are twisted, and I am trying to untwist his feet using the inpaint tool it's not changing I entered the prompt "cool sneakers," and nothing happens. What am I doing wrong?
This version isn't working out for me, I can only make 4 at once, there is no prompt Prompt Adherence (CFG Scale). there a way for me to go back and use the old version instead of the current one?
I can't find an OpenArt tutorial that shows how to place an object behind another object. For instance, a dog's ear is erroneously placed in front of a man's elbow. I want to move the dog's ear behind the elbow. How do I do that in OpenArt?
Is anyone else running into the issue where after you edit, your image resolution and height and width has been cut in half? Also, upscaling doesn't help because upscaling can only do so much to a low res image.
@@rainyang509 hi, what resolution is your image after its cut in half? does it become ~512x512? If so, it's possible this was related to a bug that should've been fixed!
Struggling to get into this, much prefer the old version, please restore it. And how do I save to png or jpeg like before instead of this annoying webp garbage?
Thanks! Model bias is for sure a big limitation of image models in general. OpenArt focuses on building workflows around great models trained by others. We don't have direct control over this unfortunately, but we believe models will grow to have more and more comprehensive training datasets to better address this in the near future!
@@openart_ai Thanks for responding but there’s a big difference between saying it’s going to be done (future speak) and actually doing it. It has been documented by Joy Buolamwini a Canadian computer scientist in her documentary on Netflix “Coded Bias” that AI doesn’t play fair when it comes to faces/people of color, particularly as it relates to facial recognition. I have been having a dickens of a time trying to generate a black image with short hair. The program consistently rendered an image of a white person with short hair. 🤷🏽♀️ What I can’t understand is why there are countless Caucasian and Asian models, but no Black models? #justsaying
I have been using open art for about 3 months now, and have not watched any videos on making AI art. That being said I have never done any editing and have over 16,000 published pictures. This video was incredible informative and I thank you kindly for making it. 💚🙏🌞🙏💚
@@inner27ofra thank you!! only the best for our users💛
How do I unpublish an image?
@@deborahd.7281 hi, you can go to your profile page and unpublish from there by clicking the 3 dots over the image
I just love listening to her voice!
I just love watching to her boobs!
it's a simple change, but if you take the detail hint from the image and make the original as a link to create a new image it will look more cohesive
Thank you for the video. This really helps figure out new updates.
Wow this is great and gonna make a lot of difference to my creations Thanks
3:56 I understand now why I got some problem with editing! Thx!
Could the next tutorial go into detail about the 'mask blur' and 'mask grow' sliders? With this new version I get a lot of weird artifacts between the area being edited and the unedited areas. Sometimes it actually looks like a 'torn paper' edge especially if the original background is plain. For example, if there is a person against a plain background and I use magic selector to highlight their shirt and change it to a different style there will be a weird artifact ring that extends well beyond the highlighted area. Any possibility of a toggle between this version and the previous?
Hi! Good idea for the tutorial. For the problem you pointed out about plain background, it s a flaw of the current inpainting we found out too. We have a fix we will launch in the upcoming week! Stay tuned. I will ping here once we introduce the fix. Thanks!
Thanks for the feedback! Our team just put in some fix on the issue you raised. Now the "torn paper" edge effect should be minimized. Please try and let us know. Thank you again ❤
Buenisima Herramienta , de gran ayuda
The popup that leads to this tutorial fills the entire screen and I couldn't find a way to get rid of it so I can *actually use the website.* If there was an X button on the window, it was way out of frame. I literally had to use an element blocker to get rid of it. So, thanks for that. :/
Thanks for the feedback! I’ve reported to the devs and it ll be fixed upcoming week. ❤
good job
Sometimes when uploading an image to edit, the two cat tutorial ad, takes over the WHOLE screen no matter what you do. Hope you guys can look into that.
Thanks for reporting! Our team has fixed this issue ❤
Is there any way to not get downscaled version? Resolution drops, and if I use the upscaler, it just adds to a perfect image to many 'creative' details which i do not want at all
Hi, really really great question, I need to add a note on this! You're using creative upscale. You can use the new ultimate upscale - precise upscale! It doesn't change the image creatively.
The editor needs a lot of generation steps, so we downscale to keep it fast.
I am having a major problem here. I have a character whose feet are twisted, and I am trying to untwist his feet using the inpaint tool it's not changing I entered the prompt "cool sneakers," and nothing happens. What am I doing wrong?
the tool is unusuable now, it creates line and downscales part of the image far outside the selected region
This version isn't working out for me, I can only make 4 at once, there is no prompt Prompt Adherence (CFG Scale). there a way for me to go back and use the old version instead of the current one?
thanks for the feedback! yes we're working on those features for the new editor :)
I can't find an OpenArt tutorial that shows how to place an object behind another object. For instance, a dog's ear is erroneously placed in front of a man's elbow. I want to move the dog's ear behind the elbow. How do I do that in OpenArt?
Is anyone else running into the issue where after you edit, your image resolution and height and width has been cut in half? Also, upscaling doesn't help because upscaling can only do so much to a low res image.
@@rainyang509 hi, what resolution is your image after its cut in half? does it become ~512x512?
If so, it's possible this was related to a bug that should've been fixed!
@@openart_ai Hi. The aspect ratio was originally at 16:9 (Cinema 1024 X 576). But after editing, the image went down to 512 X 288
nice ai can used this for poster can you of pixar buy dustya
this girl also ai 😂😂😂😂
You voice just killed me
Does anyone know how to unpublish an image?
i think from your profile, you can hover on the image and click on the 3 dots to unpublish
@@hehe-chilling Thank you!
how to swap to my own face?
You can upload a picture of your face as an image reference with inpainting.
plpi+
dame like
The pop up box to this video stops you from being able to use the app…
are you referring to the box with two cats? if you select or paint an area on the image, the box will go away
Yes, om iPad there is no way to remove it. Clicking it takes you out of the app
@@tempyoutube3399 that's weird, sorry about that!! it might be buggy on ipad - lemme check w team now
@@tempyoutube3399 hi, i tested on my ipad and couldnt reproduce the issue. could you share more info, like the browser you're on? thanks!
It’s safari on the brand new 13 inch version.
what if even her face is ai?
🫣🫣
Struggling to get into this, much prefer the old version, please restore it. And how do I save to png or jpeg like before instead of this annoying webp garbage?
We're bringing back the old inpainting as an option!
the webp thing is a bug. will be fixed soon
@@openart_ai Thank you! Awesome!
@@openart_ai brilliant, thank you :)
@@openart_ai Brilliant! Thank you
Great job - but it does not generate beautiful easy images for black people. This should be addressed.
Thanks! Model bias is for sure a big limitation of image models in general. OpenArt focuses on building workflows around great models trained by others. We don't have direct control over this unfortunately, but we believe models will grow to have more and more comprehensive training datasets to better address this in the near future!
@@openart_ai Thanks for responding but there’s a big difference between saying it’s going to be done (future speak) and actually doing it. It has been documented by Joy Buolamwini a Canadian computer scientist in her documentary on Netflix “Coded Bias” that AI doesn’t play fair when it comes to faces/people of color, particularly as it relates to facial recognition.
I have been having a dickens of a time trying to generate a black image with short hair. The program consistently rendered an image of a white person with short hair. 🤷🏽♀️
What I can’t understand is why there are countless Caucasian and Asian models, but no Black models?
#justsaying
she sounds pedant, why does she have to talk like that?
1:47 I'm but a simple man but option 2 seems like a good choice, maybe just a bit more undercarriage?
Looked quite fine to me though.