Watu wakisoma mambo ya diffussion na osmosis ulikua wapi?😄😄

Have been talking about diffussion LLMs for a long time...wrote this article a while back... medium.com/generative-ai/bey… Mercury 2 has been in my radar for some time now...

Who knew a random trip to MoMa can change your life , Art museums are supposed to inspire you, i didnt get it until i hit this exhibit. Midjourney was popping off, Stable diffussion prompts and optimization all over the feed, chaos everywhere. until i came to the curation of it all (yes im talking about the AI exhibition in '22)

Usa Stable diffussion en local LoRA controlnet, de nada 🥀 (si usas SDGui usa la versión forge, puede hacer que corra en gráficas poco potentes o incompatibles como Iris XE, Geforce serie 1000, Radeon RX)

ngl historically the filipiniana worn by lowland christian groups ended up evolving with very little regional difference. probably due to the increased interaction between these ethnic groups in sp era and mnl being the center of that cultural diffussion

UniRelight is a diffussion based relighting model that has been led by @Kai__He 2/n

Fabienne et 7 autres personnes ont aimé votre réponse.· Une info importante qui mérite une diffussion ,nos autorités sanitaires devraient connaitre le N° de ces lots ,et la secu les gens vaccinés pour faire un suivi et des controles medicaux pour les prendre en charge

Une info importante qui mérite une diffussion ,nos autorités sanitaires devraient connaitre le N° de ces lots ,et la secu les gens vaccinés pour faire un suivi et des controles medicaux pour les prendre en charge .Seulement 4,2 % des lots ont provoqué 71 % d'effets indésirables.

@elonmusk built this for TeraFab to fix thermal gradients, diffussion, and inference. Von Neumann meets Horde Probe drops on Mars rock: Optimus scans—iron, silicon, solar. Chip fires: Real-time inference thermal gradients reroute heat. Hot spots cool, cold boost—no meltdown, no lag. Horde unleashes: 300 agents swarm—mine, smelt, print. Diffusion fix kills noise, bots build without fail. Self-copy: One makes two. Two make four. Horde doubles every hour—raw, chaotic, unstoppable. Data vault feeds Grok: Copyright-free, UPCI-tagged. Grok learns faults, predicts breaks. Merge payoff: Grok runs the horde—reasoning real-time = bots that evolve, no human leash.

it instantly stops being charming if its ai not even bc of the "oh ai bad" principle but its like actively funnier if its made with effort from the devs instead of just typing in a funny prompt into stable diffussion like anyone can do that people do that all day on facebook

what is diffussion approach?

It’s really tough if you want to build a level to get good scale and style consistency. I think @RealAstropulse is the best out there with retro diffussion work. But if you want anything of sophistication and scale, premade tilesets are still best.

One tool is Midjourney. midjourney.com It is very easy to handle but it cost a sub. Stable Diffussion on local is so to say free, but it depends on your hardware how much you can do with it. github.com/lllyasviel/stable… I would recommand to google on it to learn it.

Ok, i hate to be the kind a person that does this, but PLEASE if you like my art and you want to support me in any way just press the RETWEET button. My art is not even reaching my usual moots and followers so the diffussion is greatly appreciated. Again sorry for asking this way

He's admitted to using stable diffussion multiple times in the past. I was told he did have AI/SD in his bio before, but removed it at one point (I assume to start charging for commissions) (2/2) x.com/Redpostit101/status/17…

j'attend comme ça 🧍‍♀️la diffussion des deux BX d'hélé . #Helenabailly

It's not a compression. Stable Diffussion as trained on LAION-5B, weighs less than 4 bits per image in the training set. There simply isn't any meaningful data on a per-image basis in the model.

Dimanche Nov 30 Les Regionales du Peuples de 10hr a 19hr Cameroun, 4hr a 13hr EST. Diffussion Regionales Du Peuple Talk With Mimi Mefo: youtube.com/@MimiMefoInfo-MM… Beyond AfricaPodcast: youtube.com/@BeyondAfricaPod… Nzui Manto: facebook.com/profile.php?id=… Paul Chouta: facebook.com/Paulchouta MMI News: facebook.com/MimiMefoInfo Nfor Hanson: facebook.com/CAMEROONNEWSAGE… Boris Bertolt: facebook.com/borisbertolt1 Rebecca Enonchong: x.com/africatechie Barrister Alice Nkom: facebook.com/alice.nkom.397

CNN are the best ones for detecting cancer, fact It will probably stay that way for a while, it's like the perfect use case possible Diffussion and transformers models (ones used for things like image gens and chatgpt respectively) are better for stuff like developing vaccines

Turns out a gigantic LLM is all you need for image generation. Yet this approach also makes Nano Banana Pro costly to run. Curious to see if @Google can distill Nano Banana into a more efficient diffussion model for the next generation of their Imagen line.