//=time() ?>
You can now follow me on Telegram ! 💕
On my channel ;
🌿 Mlice 🍻 - Art & Updates ! ✏️
https://t.co/HyeZAwnNWW
Still hard for me to optimize my communication and medias, at least I'm trying !!! 😤
(Very old doodle to illustrate) 🙃
学習用Cool Japan Diffusion 2.0のOptimizerを削ったバージョンに対するAUTOMATIC1111用のColabのノートブックを公開します。
Colab標準ランタイムでも動くようになりました。ただ、画像が以前よりちょっとボケているような気がします。
https://t.co/qw8ykFqKre https://t.co/SJrQ86RCnv
Oh right, here's the HD version of my last art piece. I figured out how to optimize it for Twitter so it doesn't look like trash lol
Very sorry for the abrupt ending to the stream! Guess Warzone wanted me to stop playing lmao. Hopefully we'll be able to play it again next time once its a bit more well optimized.
ORZ
Until then, looking forward to other games later this week! 🥰
Art by @/istinuu https://t.co/hVqaBSMzna
最近の研究では、事前に「超勾配」の式を手動で導出するで、勾配降下法によってステップ サイズをオンラインで「学習」できることが示されていて、バックプロパゲーションを変更して、超勾配を自動計算する方法
Gradient Descent: The Ultimate Optimizer
(NeurIPS 2022)
https://t.co/segT8iSyXE
Working on the Among Us VR Crewmate avatar for VRChat a bit more, I got all the hats and am making them toggles like the player colors. Triangle count totals to 17,089 so it's still very optimized for both PC and Quest use.
optimized the #stablediffusion Unet for Neural Engine. this time the whole thing, not just self-attention.
I think CoreML is still running it on GPU instead though (GPU-only benchmark gets same speeds).
anybody know how to peer into the black box?
https://t.co/JH1bdG8Htj
ELON # 4 (TESLA 2.0) 🎨
Only .01 ETH 🪙
Please Like & RT If You Approve 🔥
"It is going to be highly optimized for autonomy—meaning it will not have a steering wheel or pedals."
Elon Musk
#NFT #NFTs
https://t.co/55OzXP2D8K
I'm interested in making something in a balanced way, but I'm also interested in making something that's highly optimized at one point, very specific. it will be your problem, me. Swrks - 2022
GN.🌕🐺☕️
it's never a bad day to mint a runner (only .005 eth, with optimized gas!) ⚡️
They beat me 3-0 on the 10th Floor but I could tell this person was celestial. Those 3 losses were a gift. It showed me that there are way more optimized situational punishes than I thought and it made me want to learn them.
got #stablediffusion's Unet compiled to CoreML, targeting all compute units (including Neural Engine).
replaced self-attention with the MultiHeadedAttention that Apple optimized for Neural Engine.
not faster yet (need cross-attention too).
https://t.co/9tdKdGIK7X
@NortonsMarigold i will stop bullying hamlet when they optimize him to look as pretty as his promo art 😔 got bamboozled by this false advertising
"Cotton Candy" was burned and reminted in the @vcaresidency @objktcom collection.
Used the occasion to add another level of complexity. Now we have another plane of entirely #generative lightnings in the background, optimized .png layers and a better synced ambient light.
link👇 https://t.co/TCv7ABtjmP
@Selen_Tatsuki Hey Selen! I'm a 3D Artist working in the games industry! I like stylized art and can make optimized engine ready props, environments, and characters.
https://t.co/hWLSDa6hk4
7/9
Few notes while you enjoy a slow version:
- The process isn't optimized at all, it was a way to try stuff.
- Sometimes I was happy with a prompt/settings combo and I just wanted some variation. So I would launch multiple renders with a different starting frame