AI研究論文每日精選

Depth Anything V2

An Image is Worth More Than 16x16 Patches: Exploring Transformers on Individual Pixels

Transformers meet Neural Algorithmic Reasoners

OpenVLA: An Open-Source Vision-Language-Action Model

Samba: Simple Hybrid State Space Models for Efficient Unlimited Context Language Modeling

Alleviating Distortion in Image Generation via Multi-Resolution Diffusion Models

Test of Time: A Benchmark for Evaluating LLMs on Temporal Reasoning

DiTFastAttn: Attention Compression for Diffusion Transformer Models

Visual Sketchpad: Sketching as a Visual Chain of Thought for Multimodal Language Models

Interpreting the Weight Space of Customized Diffusion Models

MuirBench: A Comprehensive Benchmark for Robust Multi-image Understanding

HelpSteer2: Open-source dataset for training top-performing reward models

mOSCAR: A Large-scale Multilingual and Multimodal Document-level Corpus

CS-Bench: A Comprehensive Benchmark for Large Language Models towards Computer Science Mastery

4M-21: An Any-to-Any Vision Model for Tens of Tasks and Modalities

EMMA: Your Text-to-Image Diffusion Model Can Secretly Accept Multi-Modal Prompts

Explore the Limits of Omni-modal Pretraining at Scale

Cognitively Inspired Energy-Based World Models

Mistral-C2F: Coarse to Fine Actor for Analytical and Reasoning Enhancement in RLHF and Effective-Merged LLMs

Commonsense-T2I Challenge: Can Text-to-Image Generation Models Understand Commonsense?

TC-Bench: Benchmarking Temporal Compositionality in Text-to-Video and Image-to-Video Generation

Real3D: Scaling Up Large Reconstruction Models with Real-World Images

Estimating the Hallucination Rate of Generative AI

MLKV: Multi-Layer Key-Value Heads for Memory Efficient Transformer Decoding

Language Model Council: Benchmarking Foundation Models on Highly Subjective Tasks by Consensus

CVQA: Culturally-diverse Multilingual Visual Question Answering Benchmark

LRM-Zero: Training Large Reconstruction Models with Synthesized Data

Understanding Hallucinations in Diffusion Models through Mode Interpolation

CMC-Bench: Towards a New Paradigm of Visual Signal Compression

Toffee: Efficient Million-Scale Dataset Construction for Subject-Driven Text-to-Image Generation