jasondavies 的提交内容

Magic Insert: Style-Aware Drag-and-Drop

Diffusion Forcing: Next-Token Prediction Meets Full-Sequence Diffusion

The Mathematical Basics of Diffusion

Segment Any Text

DETRs Beat YOLOs on Real-Time Object Detection

Latent Intrinsics Emerge from Training to Relight

Computational Life: How self-replicating programs emerge from simple interaction

Apple, Microsoft Shrink AI Models to Improve Them

Mip-Splatting: Alias-Free 3D Gaussian Splatting

Connecting the Dots: LLMs Can Infer and Verbalize Latent Structure

Flash Diffusion: Accelerating Any Conditional Diffusion Model

Transcendence: Generative Models Can Outperform the Experts That Train Them

PlatoNeRF: 3D Reconstruction in Plato's Cave via Single-View Two-Bounce Lidar

GGHead: Fast and Generalizable 3D Gaussian Heads

Hallo: Hierarchical Audio-Driven Visual Synthesis for Portrait Image Animation

A Tour of Differentiable Rasterization

Neural Thermodynamic Integration: Free Energies from Diffusion Models

Multimodal Masked Modeling

Understanding Hallucinations in Diffusion Models Through Mode Interpolation

SuperPrimitive: Scene Reconstruction at a Primitive Level

PowerInfer-2: Fast Large Language Model Inference on a Smartphone

Proteus: Real-Time Expressive Generative Humans

Torax: A Fast and Differentiable Tokamak Transport Simulator in Jax

Samba: Simple Hybrid State Space Models

An Image Is Worth 32 Tokens for Reconstruction and Generation

Image Neural Field Diffusion Models

TextGrad: Automatic "Differentiation" via Text

Everything Apple Plans to Show at Its AI-Focused WWDC Event

Apple Intelligence Is Right on Time

Emad Mostaque: "Happy to Announce SchellingAI"