fancybread

@PrestonJClark @SeattleSonics hell yeah brother! kemp was a dawg, i really enjoyed whippin this up.

Bravo @thefancybread! New home office center piece. @SeattleSonics #sonics #seankemp

@karsendaily thank you for exposing all of these dick pics in web3. this community has needed dick pic police for some time now and im glad to see you leading the charge. bring these clowns to justice. #tyfys

just want it publicly documented that patriotism will begin to be “cool” next year and mass adoption will happen 2025. celebs in Americana gear this time next yr if not sooner

@dreamingtulpa i stumbled upon a way to get a similar style consistently without niji, although i bet that could potentially enhance it. "cyber city odeo 808 aesthetic" was the key for me. iirc i never got results i didnt like stylistically. prompts in alts

The last #nijijourney April style for creating "scenic" anime images got released this week. Wanna try it out, add the following to the end of your prompt to activate it 👇 --niji 5 --style scenic Prompts in ALT

diving head first into temporal cohesion!

never posted this on my own feed so here it is: ikuyo. laundry list of ai models and adobe tools. fiverr for the narrator. music is ai generated.

@dreamingtulpa has this been done before? i wouldnt be shocked if it has but i havent seen it on any tutorials! figured id share. formulated it during the @clairesilver contest.

@Render_Artist this is an awesome idea dude!

And an example with cadence, interpolation, and slowmo "matched up" to taste.

I hadn't experimented much with FILM interpolation in #deforum until today. By adjusting the cadence, frame rate, interpolation, and slow motion, I was able to achieve smooth transitions between frames. Now, I'm working on improving the coherence, possibly using image guiding.

@nuclearsamurai @clairesilver that mightve been the sickest thing ive seen in a fuckin long time. holy. shit. 🤜🤛

@clairesilver Okay this is my entry, this process was a learning experience, the piece is titled Consciousness This piece was created using Midjourney, Stable Diffusion, Deforum, Digital ID, Runwal ML, EBSynth, After Effects, ChatGPT + Premiere twitter.com/MutagenSamurai…

🚨#ClaireAIContest 4🚨 Prizes: 1st: 3ETH 2nd: 1.5ETH 3rd: 1 ETH 4th/5th: .5 each & gallery exhibition. Rules: 1 Entry. Post here. Ends 4/14/23. Theme: Collaborative. Make an innovative piece using any combination of tools, as long as AI is a central part. Examples & tools 👇

@clairesilver @midjourney @DiffusionPics @runwayml @boomy here's the secret sauce for now. I believe ai could spawn an entire wave of tiny production studios that never have to film a single shot. thanks for hostin @clairesilver

@clairesilver Tools used: @midjourney @DiffusionPics Big shoutout ControlNet (HED esp) @runwayml Generated the AI audio on @boomy Premiere & AfterEffects WinX HD Vid Converter VLC Photoshop Illustrator ~60 hours. Will post full detailed workflow vid on Monday, busy night/weekend.

home stretch for the @clairesilver contest...buddies in town so hard stop at 9PM EST. heres my fav clip as a teaser.

Pens’ playoff streak is snapped and the Pirates lost 7-0. Yesterday was the worst 412 Day since that clothing shop made Pitt, a Pittsburgh sports team, take numbers off their jerseys

@bilawalsidhu dude such a smart idea to use diff source imagery for diff controlnet tools. multicontrolnet is the major workhorse rn. its the key to the next level imo. cant wait to apply your concept to my workflow! thanks for sharing 🙏🏻

And that's a wrap! I love to share what I'm learning with the open source AI & creator community, so if you found this helpful I'd appreciate it if you: 1. RT the thread below 2. Follow @bilawalsidhu for more 3. Subscribe to get these right to your inbox: creativetechnologydigest.substack.com

Multi ControlNet is a game changer for making an open source video2video pipeline. I spent some time hacking this NeRF2Depth2Image workflow using a combination of ControlNet methods + SD 1.5 + EbSynth. 🧵 Full breakdown of my workflow & detailed tips shared in the thread below ⬇

figured that was beat so i tried to train a LORA. same principle. failed at that worse than the hypernetwork. i dont think i understood controlnet and that was more my issue on both those hiccups^ here i sit now. i have got a smooth clip inspired by wonderwoman thanks to controlnet. a combination of hed, openpose, and canny were the solution, amongst some other creative workarounds. ill share the good clip tmrw, its on my comp and im passin out. also produced the score on @boomy gonna edit that tmrw and hopefully have a mock version of the vid ready by lunch. then its all stylization from there. gn wagwtfc

anyway, my hypernetwork was kinda sick. it didnt work great, but this was the result once applied to this clip from wonderwoman.

been learnin so much stablediffusion the past couple days, i forgot to post some updates on the @clairesilver contest. project: <1min short animated video inspired by icarus. during storyboarding using midjourney, i found a style that i was obsessed with.