UniRL is a Framework for Unified Multimodal Model Reinforcement Learning

A Curated List of Vision-Language-Action (VLA) and World Action Models (WAM) Research and Beyond

【三年面试五年模拟】AIGC/LLM/AI Agent算法工程师面试秘籍。涵盖AIGC、LLM大模型、AI Agent、传统深度学习、自动驾驶、机器学习、计算机视觉、自然语言处理、强化学习、大数据挖掘、具身智能、元宇宙、AGI等AI行业面试笔试干货经验与核心知识。

Codec-Aligned Sparsity as a Foundational Principle for Multimodal Intelligence

Implementation for "The Scalability of Simplicity: Empirical Analysis of Vision-Language Learning with a Single Transformer"

Trainable fast and memory-efficient sparse attention

Video-Panda: Parameter-efficient Alignment for Encoder-free Video-Language Models [CVPR 2025]

MTRefSeg: An Open-Source Benchmark and Baseline for Multi-temporal Referring Segmentation

ICML2025

Unofficial PyTorch reproduction of DeepSeek's Thinking with Visual Primitives.

Efficient Universal Perception Encoder: a single on-device vision encoder with versatile representations that match or exceed specialized experts across multiple task domains.

Awesome Unified Multimodal Models

[ICLR & NeurIPS 2025] Repository for Show-o series, One Single Transformer to Unify Multimodal Understanding and Generation.

Repository for Meta Chameleon, a mixed-modal early-fusion foundation model from FAIR.

PyTorch denoising diffusion demo

[ICLR 2024 Spotlight] DreamLLM: Synergistic Multimodal Comprehension and Creation

PyTorch implementation of MAR+DiffLoss https://arxiv.org/abs/2406.11838

Janus-Series: Unified Multimodal Understanding and Generation Models

Official repo for "Let ViT Speak: Generative Language-Image Pre-training"

Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe

Hy3 preview (295B A21B), a leading reasoning and agent model in its size, with great cost efficiency

SenseNova-U series: Native Unified Paradigm with NEO-unify from the First Principles

Official implementation of Tuna-2: Pixel Embeddings Beat Vision Encoders for Unified Understanding and Generation

VisionFoundry: Teaching VLMs Visual Perception with Synthetic Images

Vero: An Open RL Recipe for General Visual Reasoning