Volodymyr Kuleshov 🇺🇦
@volokuleshov
Followers
8K
Following
2K
Statuses
2K
AI Researcher. Prof @Cornell & @Cornell_Tech. Co-Founder @afreshai. PhD @Stanford.
New York, NY
Joined July 2013
✨Simple masked diffusion language models (MDLM) match autoregressive transformer performance within 15% at GPT2 scale for the first time! 📘Paper: 💻Code: 🤖Model: 🌎Blog: [1/n]👇
4
38
193
RT @deedydas: China just dropped a new model. ByteDance Doubao-1.5-pro matches GPT 4o benchmarks at 50x cheaper — $0.022/M cached input to…
0
1K
0
RT @ClementDelangue: Just 10 days after o1's public debut, we’re thrilled to unveil the open-source version of the groundbreaking technique…
0
636
0
Classifier-free guidance enables applying this in the same model, as with Gaussian diffusion:
1
0
8
RT @SchiffYair: Diffusion models produce high quality outputs and have powerful guidance mechanisms. Recently, discrete diffusion has shown…
0
22
0
RT @iScienceLuvr: While the math behind flow matching can be intimidating, it can be implemented in just a few lines of code, as shown in t…
0
43
0
RT @SkyLi0n: Excited to present our #NeurIPS2024 paper: "Simple and Effective Masked Diffusion Language Models" on Thurs at 11:30 a.m. in H…
0
4
0
Another paper similar to VAR (NeurIPS best paper). I’m surprised that these results are coming out just now. The spectral autoregression idea has been around for years. I wonder if we all assumed it was too hard because if it worked someone would have done it already.
Our new paper, "Spectral Image Tokenizer", is on arXiv! We train a tokenizer on DWT coefficients that enables autoregressive coarse-to-fine image generation, w/ applications to multiscale text-to-image, and text-guided editing. w/ @kiamada, @msuhail153
0
0
5
RT @SkyLi0n: Prsenting our #NeurIPS2024 ✨spotlight paper "Diffusion Models with Learened Adaptive Noise" on Fri at 4:30 p.m. in Hall A-C (…
0
5
0
Seen at #neurips2024: authors giving away chocolate candy at their posters. That’s some next-level guerrilla marketing kung fu.
2
0
11
Did you know there is free gourmet coffee right outside #Neurips2024? All you need is to sign up for and go to one of the red coffee carts at the conference center entrance. 🙏 @davederiso
2
1
6
The GAN is dead; long live the GAN!
Can GANs outperform diffusion models in 2024? Yes! #2200: The GAN is dead; long live the GAN! A Modern GAN Baseline! Come join me if you're attending NeurIPS 2024. Thursday evening poster session 4:30-7:30 . #NeurIPSConference with @zonkedNGray @volokuleshov @jtompkin
0
1
13
RT @sedielem: When a bunch of diffusers sit down and talk shop, their flow cannot be matched😎 It's time for the #NeurIPS2024 diffusion cir…
0
21
0