Adding to Cart…
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.You currently have no notifications.
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
I think the hands are wonky in every image I've tried. But it's been a fun test.
Just trying some sci-fy pics,looks really good to me.
Some of mine.
You can do it one of two ways. You can upload the image to an outside server, something you can get a link to the image or you can do as you did here. Once it is posted, click on the image and copy the URL. Once you do that go to your text box and select the icon in the header of the image, just next to the flag on the right side. Click on that past your URL in it and adjust the size so the with is no bigger than 800 pxl. The easiest way is to do the off-site host and then you just start with doing your post and then select that letter icon and add it.
Not the greatest, also not the worst -- after all, it is a beta and I am quite sure there will be many more features and improvements to come. Hopefully we'll be able to use the Daz content we already own in it without having to buy everything all over in LORA format or whatever to get it working. We also really, really need inpainting -- I do not know what happened to this poor kitty but he seriously needs that tail fixed. Haven't seen anything like it in ages, probably because in other AI apps I have control over negative prompts and other settings that give better though admittedly not always 100% perfect results.
Playing around a bit :)
photo of a woman, slim, blonde shaggy pixie cut, pale skin, white t-shirt, gray hoodie, blue jeans, standing on a new york rooftop, looking down on the busy city, rim lighting,
Prompt:: A spaceship approaches a purple planet shrouded in apricot-coloured mist
It's fun to play with this, I like dragons ;)
theme iridescent
That's a very clever idea I might have use for as well
Hands are definitely a continuing challenge. lol
Using one of my Nightcafe prompts - pity you can't add the negative ones too.
prompt:
dark skinned woman, wearing a pricess gown covered with perls and diamonds sparkling in the moonlight, the black curly hair is formed in an voluminous updo with a diadem and perls, she stands in a park with flowers blooming in the night
I wanted to see how well it get's dark skin, it decided on it's own that she needs to wear a white dress
this one is the best I think:
Another recycled Nightcafe prompt:
Those are beeeeeeutiful Linwelly... I love how the artwork came out even with words mispelled... Princess has an N in it.. and we in America spell peArls with an A. Pretty smart that AI ;) well done.. thanks oodles for sharing. I love wedding style dresses... I'm such a girl girl ;)
oh you're right, me and my typos I had to laugh, but honestly that is pretty neat, I had another one where it managed to read the misspelled word correctly
I still don't like how it handles hands, but I think that's a problem with any AI generator. To me, it does quite well with portrait types and landscapes.
This one you prompted the AI to deliver is really impressive.
It's one of my favorites of the hundreds (probably) that I've generated. I think the prompt was "close-up rain-soaked woman, jungle, raining, sunset".
A really stunning image. Well done. you might want to delete the prompt.
Tried some crowded scenes and started each prompt with "The central park of New York"
Not a single one has the right amount of arms and fingers.
Amazing.
I don't think it really matters. I used the same prompt to generate 20 images and got 20 different pictures, some of them quite beautiful, some not so much. The only thing, as I understand it, that would generate the same image would be using the same seed. Even then, it will only generate one image that would be very close to mine (I've noticed only very, very minor differences using a seed), and any others would be vastly different.
Fair enough. I noticed that the Daz gallery now has dedicated section for AI generated art. You should add it there.
Yuck. (A little girl with blonde pigtails in a pink dress holds a pink balloon in a park in the summer using V9)
They still come out looking like renders vs Midjourney.
Prompt: Portrait of cute little girl with freckles and blonde pigtails and a pink dress with portrait studio lighting.
Also when you zoom in on many images that look decent posted in the forums, all sorts of discrepancies and distortions are visible. It's still nowhere near Midjourney. Also there's no way to make variations to selected sections so you can't fix hands or on this case the hair. Daz AI still has a lot of work to do to catch up. I'd rather do an actual render then bring the renders into selfie phone apps!
That's so funny... next it will know what we are thinking! erm.. maybe not then... Wasn't there a horror movie about that? We humans will never learn I suppose... lol
This one came out better except for the eye. You need to zoom in to see all the flaws and still need good Photoshop skills to fix these.