Awesome work from Hao Zhang 's lab! They generalized the idea of consistency distillation in diffusion models to speed up LLM decoding. Lots of potential in this direction.

Diffusion Policies are powerful and widely used. We made them much faster. Consistency Policy bridges consistency distillation techniques to the robotics domain and enables 10-100x faster policy inference with comparable performance. Accepted at #RSS2024

I finally released my new video on YouTube about Diffusion Models / Score-Based Generative Models. youtube.com/watch?v=B4oHJp… Literally planned this for a year and put so much work in. I think this approach to diffusion models is so intuitive and highly recommend giving that a

Time to make your diffusion models one step! Excited to share our recent work on Truncated Consistency Models, a new state-of-the-art consistency model. TCM outperforms a previous SOTA, iCT-deep, using more than 2x smaller networks in both one-step and two-step FIDs. Joint work

Thrilled to share our latest work on consistency models! We simplified the math behind continuous-time consistency models, stabilized their training, and scaled them up to 1.5B parameters. We are now one step closer to real-time multimodal generation!

Mitigating racial bias from LLMs is a lot easier than removing it from humans! Can’t believe this happened at the best AI conference NeurIPS Conference We have ethical reviews for authors, but missed it for invited speakers? 😡

Apparently, this is called “cultural generalization” according to #NeurIPS2024 . Never too old to learn something new.

NeurIPS Conference I don’t see NeurIPS is taking this issue seriously from their post. If they really were serious, they would have avoided using such a machine learning jargon like “implicit biases” and “generalization.” The conference leaders are not taking it seriously at all.

This is awful! This is exactly how your internal racism comes out no matter how much you try. But I am surprised that someone whose main work is pseudoscience-based (journals.sagepub.com/doi/full/10.11…) was invited for a NeurIPS keynote.

Rosalind Picard made gross racist comments at NeurIPS 2024. we urge #mit and the #us Department of Education to investigate her and her claim. there is no place for rasicts in #ml and #ai. don't stop sharing this! help exposure! we need her investigated!

🎥 Frustrated by Sora's credit limits? Still waiting for Veo 2? 🚀 Open-source video DiTs are actually on par. We introduce FastVideo, an open-source stack to support fast video generation for SoTA open models. We have supported Mochi and Hunyuan, 8x faster inference, 720P

Very excited to share our interview with Jascha Sohl-Dickstein on the history of diffusion models — from his original 2015 paper inventing them, to the GAN "ice age", to the resurgence in diffusion starting with DDPM. Enjoy!

This is the most educative post on diffusion models I’ve seen—perfect for beginners. Thanks Slater Stich for the amazing work!

Llama-4's performance has been controversial recently, but how does it perform on playing classic games? Demystified below👇👇👇

Honored to be part of this conversation with Slater Stich! We covered a lot of ground in the history of diffusion — from score matching to consistency models. Hope folks find it useful!

Excited to share our interview with Sander Dieleman! This is Part 3 in our History of Diffusion series. We talk about diffusion as spectral autoregression, diffusion language models, flow matching, and much more. Enjoy!

Thrilled to see DAPS, our new technique for diffusion-based inverse problem solving, selected as an oral at #CVPR2025! Huge kudos to brilliant student collaborators Wenda Chu and Bingliang Zhang. Follow them for more great research, and don’t miss their talk on June 15!

Very excited to see this model released to the open-source community. It's still hard to believe that our latest techniques allow it to be so incredibly powerful yet so remarkably small.

Applications change, but the principles are enduring. After a year's hard work led by Chieh-Hsin (Jesse) Lai, we are really excited to share this deep, systematic dive into the mathematical principles of diffusion models. This is a monograph we always wished we had.

Checking out our latest work in diffusion distillation. Kaiwen Zheng 's 6 month distillation experiments findings and our clean training code are OPEN now.