取消静音取消静音Alec HelblingX (Twitter)Stochastic and deterministic sampling strategies for diffusion models produce strikingly different…打开视频页面3623在 X 上点赞分享X在 X (Twitter) 上打开
Alec HelblingX (Twitter)Diffusion models leverage a variety of samplers.打开视频页面1385在 X 上点赞分享X在 X (Twitter) 上打开
Alec HelblingX (Twitter)Flow matching produces smooth, deterministic trajectories.打开视频页面3564在 X 上点赞分享X在 X (Twitter) 上打开
Alec HelblingX (Twitter)I made a tool called Diffusion Explorer that lets you to train and visualize simple 2D diffusion and…打开视频页面1071在 X 上点赞分享X在 X (Twitter) 上打开
Alec HelblingX (Twitter)This smalldiffusion project is really fun.打开视频页面444在 X 上点赞分享X在 X (Twitter) 上打开
Tanishq Mathew Abraham, Ph.D.X (Twitter)Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models打开视频页面299在 X 上点赞分享X在 X (Twitter) 上打开
Boyuan ChenX (Twitter)Introducing Diffusion Forcing, which unifies next-token prediction (eg LLMs) and full-seq.打开视频页面1348在 X 上点赞分享X在 X (Twitter) 上打开
Ji-HaX (Twitter)Blog post - Autoregression vs Diffusion - Understanding Sampling via Optimal Transport打开视频页面266在 X 上点赞分享X在 X (Twitter) 上打开
Mark NeumannX (Twitter)This symmetric diffusion paper at ICLR is nice (simple idea in retrospect):打开视频页面238在 X 上点赞分享X在 X (Twitter) 上打开
Xingang PanX (Twitter)Diffusion models are sensitive to small changes in the input noise.打开视频页面407在 X 上点赞分享X在 X (Twitter) 上打开
Mihir PrabhudesaiX (Twitter)1/ Happy to share UniDisc - Unified Multimodal Discrete Diffusion – We train a 1.5 billion parameter…打开视频页面722在 X 上点赞分享X在 X (Twitter) 上打开
Nikos KolotourosX (Twitter)Excited to share our new work "AvatarPopUp"!打开视频页面113在 X 上点赞分享X在 X (Twitter) 上打开
Alex LiX (Twitter)Diffusion models have amazing image creation abilities.打开视频页面385在 X 上点赞分享X在 X (Twitter) 上打开
Arash VahdatX (Twitter)⚡ Video diffusion is slow. PDD changes that.打开视频页面297在 X 上点赞分享X在 X (Twitter) 上打开
NVIDIA AIX (Twitter)Selected as a best paper finalist at #CVPR2026: PixelDiT from NVIDIA Research打开视频页面253在 X 上点赞分享X在 X (Twitter) 上打开
Miguel Angel BautistaX (Twitter)Introducing Manifold Diffusion Fields (MDF), our new work on learning generative models over fields…打开视频页面111在 X 上点赞分享X在 X (Twitter) 上打开
Matthias NiessnerX (Twitter)(1/2) Check out "𝐏𝐨𝐥𝐲𝐃𝐢𝐟𝐟: Generating 3D Polygonal Meshes with Diffusion Models"!打开视频页面548在 X 上点赞分享X在 X (Twitter) 上打开
RisbekX (Twitter)I didn’t get the chance to greet her, but I’m feeling both happy and sad at the same time.打开视频页面875在 X 上点赞分享X在 X (Twitter) 上打开