Zhuang Liu

DepthLM: Metric Depth from Vision Language Models

SoFlow: Solution Flow Models for One-Step Generative Modeling

AutoCode: LLMs as Problem Setters for Competitive Programming

UEval: A Real-World Benchmark for Unified Multimodal Generation

Generative Modeling of Weights: Generalization or Memorization?

AgentVQA: A Unified Benchmark for Agentic Visual Understanding

A Decade's Battle on Dataset Bias: Are We There Yet?

Meta CLIP 2: A Worldwide Scaling Recipe

LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding

Deconstructing Denoising Diffusion Models for Self-Supervised Learning

A Coefficient Makes SVRG Effective

LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding

Idiosyncrasies in Large Language Models

Deep Neural Networks without Normalization