Sabitlenmiş Tweet
Next Gen Typography Tools are going to be soo crazy!! 🤯
Combined Touch type by @schultzschultz_ with LCM + sdxl
@peterom
@HelloCivitai
@StableDiffusion
#aianimation #aivideo #stablediffusion #SDXL #typography #LCM
English
Mjay
1.5K posts
Mjay
@mjayai
Award Winning CG Artist with over a Decade of Dev Experience | Founder @ Dyzen Labs | prev. Head of AI, MLE and CVPR Reviewer
Bangalore, India Katılım Aralık 2018
1.1K Takip Edilen526 Takipçiler
Had a great time working on this film, congrats to the team 🙌👏
Ram Gopal Varma@RGVzoomin
Check out the highly intriguing trailer of GAME of LOANS directed by @abhishek_leslie and wonderfully photographed by @sabarivgnes youtu.be/p7qjkE0kZEQ All the best to team @actor_nivas @esternoronha5 @abinay_kinger @AthvikJalandar @SuryaSid09 @joe_costa @DPradeepJenifer @shiva3721 @vinoo @jeevaeditor @saregamasouth #Gameofloans
English
Mjay retweetledi
Just when you thought we were done for the day... 🎬
Introducing Vortek – the multi-color solution you've been waiting for. Up to 7 materials, zero purge waste, wireless reliability. Finally, multi-color printing that actually makes sense.
Coming Q4 2025! ✨
#Bambulab #Vortek #MultiColor3DPrinting #ZeroWaste #3dprinting #Innovation
English
Career update 🚀
I'm thrilled to join Microsoft Research, Bangalore, as a Postdoctoral Researcher!
I want to express heartfelt gratitude to everyone who's guided, supported, and uplifted me through this journey 🙏
🧵
English
@jfischoff Yes, playground 2.5 is the best open source model out there... but too bad there is no supported controlnets for it yet.
English
Stable Cascade isn't a flashy model, like Playground 2.5 which nearly always has very aesthetic colorful dynamic compositions, but it makes very predictable nice consistent compositions.
English
@lizaledwards Although, nothing beats a CG Artists currently but being a cg artist and ai researcher, i think we could achieve the same thing for free, thanks to open source models like instant mesh and auto-rig pro blender addon pipeline then finally mixamo for animation.
English
Mjay retweetledi
Hostile Mug. Copper mug designed in @sidefx Houdini. Inspired by anti-human hostile architecture. Go away. #3Dprinting #Houdini #electroplating
English
@jfischoff 🤔Isnt that just vector DB with RAG where we retrive via current character description similarity and recieve character arc template?
English
Do story embedding models exist? Like get embedding that represents the characters, arc etc. Bonus points for interpretable directions.
English
That thing scares me, the way it stands is really creepy, I imagined it as being made of liquid metal or something like that @BostonDynamics #ai #stablediffusion #AnimateDiff #animation #aiar #VFX #cgi
English
Mjay retweetledi
Mjay retweetledi
@EMostaque @mrityunjay_99 But I think fully open models may be necessary for the community to embrace - but SD3 will be powerful/impactful w/o significant community engagement but think it may not experience that capability gain that e.g. SD1.5 or AD did
English
English
@thibaudz @mrityunjay_99 They released an SD1.5 version that performs nicely but is limited by the fact that it's SD1.5
@DataPlusEngine and his team are working on an SDXL version (and a few other other teams I believe)
English
SD3 is incredible but few will extend it due to Stability's approach to licensing & it'll be surpassed by SDXL as a result - specifically, an open source ELLA (attached) will beat SD3's prompt comprehension.
Their business model discourages contributions. The vibes are off.
Stability AI@StabilityAI
Today, we are pleased to announce the availability of Stable Diffusion 3 and Stable Diffusion 3 Turbo on the Stability AI Developer Platform API. We have partnered with @FireworksAI_HQ , the fastest and most reliable API platform in the market, to deliver these models. In keeping with our commitment to open generative AI, we aim to make the model weights available for self-hosting with a Stability AI Membership in the near future. You can get started and learn more here: bit.ly/3W43FjY Prompt: Awesome artwork of a wizard on the top of a mountain, he's creating the big text "Stable Diffusion 3 API" with magic, magic text, at dawn, sunrise.
English
Experts of Bregman divergences, I need help for a reference
Who first proposed the concept of generalized Bregman divergence D_f(x;a)=f(x)+f*(a)-<a,x> ?
I found Gordon COLT'99, see image, but it must be an older idea. Maybe Amari proposed it first?
English
Mjay retweetledi
Average person vs normal person vs shader compiler error person
Español
Mjay retweetledi
Depth requires time.
English
@jfischoff yea, that makes sense, but i still think that diffusion based super resolution will have much greater quality outputs then just NFs or U-net with alt. obj fns.
English
@mrityunjay_99 It just seems convoluted to do super-resolution with diffusion and a CN. You have this extra noised input and the lq. It seems more straightforward to go lq->hq
English
Has anyone looked into the benefits of using an ReferenceNet vs a ControlNet? Could they be combined? For instance, a ReferenceNet that combines attention but also has zero-convolutions for features.
Also, why does the CN only adapt the half the unet? Why not the whole thing?
English