Code-as-CoT for Physically-Consistent Video Generation via an Agentic Dual-Engine System

One Discrete Word for Visual Reasoning Overtakes Agentic and Latent Methods

The first Interleaved framework for textual reasoning within the visual generation process

Are Video Models Ready as Zero-shot Reasoners?

ULMEvalKit: One-Stop Eval ToolKit for Image Generation

A framework for unified personalized model, achieving mutual enhancement between personalized understanding and generation. Demonstrating the potential of cross-task information transfer in persona…

Official implementation of UnifiedReward & [NeurIPS 2025] UnifiedReward-Think & UnifiedReward-Flex

Official repository for the paper "TIIF-Bench: How Does Your T2I Model Follow Your Instructions?".

[NeurIPS 2025] MINT-CoT: Enabling Interleaved Visual Tokens in Mathematical Chain-of-Thought Reasoning

Fast-in-Slow: A Dual-System Foundation Model Unifying Fast Manipulation within Slow Reasoning

EchoTraffic: Enhancing Traffic Anomaly Understanding with Audio-Visual Insights (CVPR 2025)

[NeurIPS 2025] T2I-R1: Reinforcing Image Generation with Collaborative Semantic-level and Token-level CoT

MAGI-1: Autoregressive Video Generation at Scale

[CVPR 2025] The First Investigation of CoT Reasoning (RL, TTS, Reflection) in Image Generation

[ICLR 2025] The First Multimodal Seach Engine Pipeline and Benchmark for LMMs

The Most Faithful Implementation of Segment Anything (SAM) in 3D

[ECCV 2024] Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems?

[CVPR 2024] OneLLM: One Framework to Align All Modalities with Language

[CVPR2024 Hightlight] No Time to Train: Empowering Non-Parametric Networks for Few-shot 3D Scene Segmentation

Align 3D Point Cloud with Multi-modalities for Large Language Models

Personalize Segment Anything Model (SAM) with 1 shot in 10 seconds

[ICCV 2023] Code for "Not All Features Matter: Enhancing Few-shot CLIP with Adaptive Prior Refinement"

(ICCV2023) Official implementation of 'ViewRefer: Grasp the Multi-view Knowledge for 3D Visual Grounding with GPT and Prototype Guidance'

[ICLR 2024] Fine-tuning LLaMA to follow Instructions within 1 Hour and 1.2M Parameters

[CVPR 2023] Parameter is Not All You Need: Starting from Non-Parametric Networks for 3D Point Cloud Analysis

Official pytorch implementation of "DSPoint: Dual-scale Point Cloud Recognition with High-frequency Fusion"

Generic PyTorch dataset implementation to load and augment VIDEOS for deep learning training loops.

[CVPR 2023] Prompt, Generate, then Cache: Cascade of Foundation Models makes Strong Few-shot Learners