The other day I was looking for free programs that can do this generative fill that Adobe came out with a month ago. Where it can fill in things that aren't shown. But people can get messed up.
In a vid I saw about this photopea program that's just like photoshop but free.. has a few ads on the right. To get to the generative fill, you go in the band aid tool.. it's this magic lasso where it'll fill in stuff. You have a bar you can input text. A few model types and couple advanced option sliders. Can also use "remove" where it'll take something out.
I was testing it out on the weekend. Took a pic of people beside each other outside. Selected one side and tested out all kinds of combinations of the advanced sliders and models. You can have it more so it goes towards what you type. People's faces get messed up and look retarded. They'll have 3 arms. Try taking out sunglasses on somebody. lol. You can keep pushing it to try different things. You can also go to smaller areas and keep trying to fix fuck ups.
This photopea though, I was trying to get a top border filled in and all at once it snafued. Had to go smaller bits at a time. If you see lines that aren't merging smoothly can select that and try to remove. It did the sides ok though. I thought of turnning it 90 degrees so the top was the sides and it was probably not enough area.
I don't know.. kinda neat though how it was free and could export with no watermark. Could also keep on trying new ones. Didn't have to make an account. This other free program after a few tries it said oh buy credits. So screw that other one. Then you got the Adobe firefly where you have to make an account and a guy was trying to take out a chick's sunglasses and it couldn't figure out the eyes. With photopea I was trying that same pic and it got the eyes ok. So.. even Adobe is fucked up with this AI shit. lol. Like "why".. right? If this shit is so smart you'd think it'd realize they're fuckin wonky. Buildings too.. it'll be lazy and doesn't do shit straight.
Like WTF with this AI shit for images and them trying it with frames for video. That's like when this first began where it'd be able to take paintings and do images in that style. When they tried video there was problems with it realizing frames before and after so it'd be more smooth. I don't know if they even got proper shit of this AI image stuff for "video".
The other day I was looking for free programs that can do this generative fill that Adobe came out with a month ago. Where it can fill in things that aren't shown. But people can get messed up.
In a vid I saw about this photopea program that's just like photoshop but free.. has a few ads on the right. To get to the generative fill, you go in the band aid tool.. it's this magic lasso where it'll fill in stuff. You have a bar you can input text. A few model types and couple advanced option sliders. Can also use "remove" where it'll take something out.
I was testing it out on the weekend. Took a pic of people beside each other outside. Selected one side and tested out all kinds of combinations of the advanced sliders and models. You can have it more so it goes towards what you type. People's faces get messed up and look retarded. They'll have 3 arms. Try taking out sunglasses on somebody. lol. You can keep pushing it to try different things. You can also go to smaller areas and keep trying to fix fuck ups.
This photopea though, I was trying to get a top border filled in and all at once it snafued. Had to go smaller bits at a time. If you see lines that aren't merging smoothly can select that and try to remove. It did the sides ok though. I thought of turnning it 90 degrees so the top was the sides and it was probably not enough area.
I don't know.. kinda neat though how it was free and could export with no watermark. Could also keep on trying new ones. Didn't have to make an account. This other free program after a few tries it said oh buy credits. So screw that other one. Then you got the Adobe firefly where you have to make an account and a guy was trying to take out a chick's sunglasses and it couldn't figure out the eyes. With photopea I was trying that same pic and it got the eyes ok. So.. even Adobe is fucked up with this AI shit. lol. Like "why".. right? If this shit is so smart you'd think it'd realize they're fuckin wonky. Buildings too.. it'll be lazy and doesn't do shit straight.
Like WTF with this AI shit for images and them trying it with frames for video. That's like when this first began where it'd be able to take paintings and do images in that style. When they tried video there was problems with it realizing frames before and after so it'd be more smooth. I don't know if they even got proper shit of this AI image stuff for "video".