The First AI Art Generator That Can Spell: New FREE Open Source AI Art Generator

Open source AI text to image models are making it possible to create art that spells out words and phrases. We explore the new models and their capabilities. Google announced Muse AI, a powerful and efficient text to image model with editing capabilities. Deepfloyd AI & Stability AI announced IF: An extremely coherent model that can spell very well - it will also be open source! And finally, Karlo, a FREE model that you can try today that spells better than DALL-E 2!

▼ Link(s) From Today’s Video:

-------------------------------------------------

▼ Extra Links of Interest:

-------------------------------------------------

Thanks for watching Matt Video Productions! I make all sorts of videos here on Youtube! Technology, Tutorials, and Reviews! Enjoy Your stay here, and subscribe!

All Suggestions, Thoughts And Comments Are Greatly Appreciated… Because I Actually Read Them.

-------------------------------------------------

AI Art That Can Spell! New FREE Open Source AI Art Generators - AI NEWS

[MattVidPro]

What do you guys think?? Share Karlo results in my discord!! Who else is super excited to try Deepfloyd's Open Source IF Model!!!

[ninjuhdelic]

were getting closer and closer to bringing the psychedelic experience to life. In etherial states you can just manifest whatever you want. for free and with no effort. im ready for it.

[CMatt]

I can't wait for Karlo to get integrated into colabs and I'm able to customize it to my liking.

[ThatMan.]

Nice video! I’ve never heard of deep floyd before!

[tonioinverness]

8:05 You buried the lede: It's not about Marilyn's sign... it's about those HANDS! Pretty decent.

[Adreitz]

Note that the band Pink Floyd have a song called If, which is referenced by the model name, the poem, and the images based on the poem. Many of the images (the dog wearing a shirt, the alien graffiti, the plant-based text coming out of a book, and the fountain pen with writing) riff on released demo images from Imagen, Parti, eDiff-I, and Muse.

[Anti-HyperLink]

I typically only use AI art generation to generate my fictional characters, so I feel like they'd only be cool to me.

[Magius]

Finally! I was literally using stable diffusion

[holdthetruthhostage]

This Has Potential but we need Ai that you can insert your own artwork & have it modify it but stay faithful to your Art Style

[sperziebn]

7:13 the SD 2 glass of wine being below half full is probably a reference to how everyone is still sticking mostly with SD 1.5 models, because SD 2.x models got such a mixed/poor response as they seem nerfed in comparison

[tearlach]

I wonder if these models improving at text accuracy using language models will also be better at interpreting spacial relativity or plurality. Things like "X to the left of Y" or "5 Xs and 2 Ys inside Z"

[tosvus]

with Karlo using the prompt: "a cute koala holding a sign that says hello", I actually got 3/6 perfectly written "Hello", one had the sign going out of frame but the first two letters were right. The two remaining ones, 1 was also going out of frame but first letter was off, and the last one actually said Hello, but the H also doubled, so it was like they mixed H and E to make a new letter, then the rest was correct.

[JimLove]

Does Google every release anything? All I seem to see are things it might release some time soon.

[davidjacobs]

If you ask midjourney to draw a famous person, it creates picture of a person that looks somewhat like that famous person, but not quite. almost like actors portraying that famous person in a biopic movie or something. I'm not sure if this was done due to copy right issues, or they are intentionally making the picture so that it doesn't look like the real person, or it's the limitation of the AI. I suspect it is intentional, not limitation of the technology.

[asterlofts]

Hahaha, you have Axolot plush! XD
Amazing video, by the way.

[russelllukenbill]

3B looks like when I'm trying to read a book in my dreams.

[rivergrrrl]

Hi👋🏻 I grabbed your invite and signed up for discord but I’m lost looking for deep Floyd —navigation tips video?

[samybean]

9:07 Maybe they mean "Be raised to discover",

[Yipper]

I one time told dallE to make a meme. And it came up with frodded Freds. I don't know what it means but it's funny.
So task completed successfully.

[jasonrist]

until then, a good photo editor can add the text