Johnowhitaker_Art

I'm @johnowhitaker, working with generative art and AI. he/him :)

[Johnowhitaker_Art @JohnowhitakerA]

My wall needed a dinosaur :)

[Johnowhitaker_Art @JohnowhitakerA]

A tale in 2 parts... Last day at the beach the sun finally got me 😂

[Johnowhitaker_Art @JohnowhitakerA]

🦊

[Johnowhitaker_Art @JohnowhitakerA]

Light painting :D

[Johnowhitaker_Art @JohnowhitakerA]

Embracing the spooky season 💀 Feels like half the yards in our area have some plastic bones in em! Not a thing in Zimbabwe really, I'm enjoying the novelty :)

[Johnowhitaker_Art @JohnowhitakerA]

Family band :)

[Jonathan Whitaker @johnowhitaker]

Video2Video models have come a long way since 2021 when I made this dancing video (https://t.co/4ZKs2AwsOc) and turned myself into a chicken (https://t.co/8gZQvwclQF, still one of my favourite creations) 😃

[Jonathan Whitaker @johnowhitaker]

I had fun making the assets for this video - why just have a talking head when it could be a marble statue! Using my 'Coherence Guided' InstructPix2Pix notebook (https://t.co/JYptu7mDAh) for all clips unless stated otherwise.

[Jonathan Whitaker @johnowhitaker]

New YouTube video, 'Stylizing Video with Diffusion Models': https://t.co/11a45FJ93j
In it, I talk through the challenges, show some cool new approaches and review the GEN1 paper from @runwayml - plus share how you can get similar results yourself! Shares + feedback appreciated <3

[Johnowhitaker_Art @JohnowhitakerA]

While joking with my in-laws about VeggieTales, I decided to imagine a version from a slightly darker timeline. A few years ago I wouldn't have had a way to share the vision on our group chat, today it's easy - hooray for generative models!

[Johnowhitaker_Art @JohnowhitakerA]

Space bears :)

[Johnowhitaker_Art @JohnowhitakerA]

Video nonsense :)

[Johnowhitaker_Art @JohnowhitakerA]

SDelfie (depth-conditioned SD-2 fed in my default face pic and a ceramic-related prompt)

[Johnowhitaker_Art @JohnowhitakerA]

"3d render of a simple robot with friendly simplified eyes, Pixar style, HD" (SD1.5)
When I next feel like picking up Blender and doing some 3D stuff I'm definitely generating my inspiration/reference!

[Johnowhitaker_Art @JohnowhitakerA]

Higher resolution, crazier style.

[Johnowhitaker_Art @JohnowhitakerA]

Stylizing video with Stable Diffusion. This test takes ~10s per keyframe and ~1s for other frames on a t4 GPU. Now to try at high res and work on consistency - although I do like the transitions synced to the beat!

[Johnowhitaker_Art @JohnowhitakerA]

"A happy round furry blob-creature holding medieval weapons at night, digital art, cute and fluffy"
I get the whole 'your dall-e gens are like your kids and nobody else cares' thing but LOOK AT THAT FACE!
10/10 - would let them invade

[Johnowhitaker_Art @JohnowhitakerA]

How is @playground_ai giving people free access to Dalle-2? Just a ton of VC money? Anyway, really fun to finally get to play with that model on a lazy Sunday morning :)

[Johnowhitaker_Art @JohnowhitakerA]

Fed a pic into https://t.co/9yC9BHTXBi and then used Stable Diffusion with the resulting description. Can you tell which is the real me?

[Johnowhitaker_Art @JohnowhitakerA]

Aaaand there goes our electricity and wifi, so calling this a wrap earlier than expected. As always when I actually find time to play with these models, I can see why it's addictive! Lots of weird results but enough gems that you keep trying to think of new ideas to test.
/end

[Johnowhitaker_Art @JohnowhitakerA]

Mostly avoiding the computer for the weekend but I couldn't resist sitting down this pm to try actually making some images - something I've done fairly little recently as I focused on tools! This will be a long spammy thread of my favourite outputs :)

[Johnowhitaker_Art @JohnowhitakerA]

I enjoy painting, and I can see myself using this a bunch to get nice references. Doesn't matter if an eye is a little weird, the important thing is lighting reference and getting inspiration for colours/composition.