Lyra: Generative 3D Scene Reconstruction via Video Diffusion Model Self-Distillation

Native Multimodal Models are World Learners

Code for "FlashWorld: High-quality 3D Scene Generation within Seconds"

You can easily calculate FVD, PSNR, SSIM, LPIPS for evaluating the quality of generated or predicted videos.

ViPE: Video Pose Engine for Geometric 3D Perception

[CVPR 2024] RoMa: Robust Dense Feature Matching; RoMa is the robust dense feature matcher capable of estimating pixel-dense warps and reliable certainties for almost any image pair.

Official PyTorch Implementation of "Diffusion Transformers with Representation Autoencoders"

Official Code of "VideoCanvas: Unified Video Completion from Arbitrary Spatiotemporal Patches via In-Context Conditioning"

rCM: SOTA Diffusion Distillation & Few-Step Video Generation

LongLive: Real-time Interactive Long Video Generation

Cosmos-Transfer2.5, built on top of Cosmos-Predict2.5, produces high-quality world simulations conditioned on multiple spatial control inputs.

Cosmos-Predict2.5, the latest version of the Cosmos World Foundation Models (WFMs) family, specialized for simulating and predicting the future state of the world in the form of video.

Cosmos-Reason1 models understand the physical common sense and generate appropriate embodied decisions in natural language through long chain-of-thought reasoning processes.

[ICCV 2025 ⭐highlight⭐] Implementation of VMem: Consistent Interactive Video Scene Generation with Surfel-Indexed View Memory

Matrix-Game 2.0: An Open-Source, Real-Time, and Streaming Interactive World Model

Wan: Open and Advanced Large-Scale Video Generative Models

Cameras as Relative Positional Encoding

Long-RL: Scaling RL to Long Sequences (NeurIPS 2025)

[ICCV 2025 Highlights] Large-scale photo-realistic virtual worlds for embodied AI

📖 This is a repository for organizing papers, codes and other resources related to unified multimodal models.

Multimodal Chain-of-Thought Reasoning: A Comprehensive Survey

verl: Volcano Engine Reinforcement Learning for LLMs

[TPAMI'25] PanopticNeRF-360 | [3DV'22] Panoptic NeRF (3D-to-2D Label Transfer in Urban Scenes)

Cosmos-Predict2 is a collection of general-purpose world foundation models for Physical AI that can be fine-tuned into customized world models for downstream applications.

Official repository for the paper "Orientation Matters: Making 3D Generative Models Orientation-Aligned" (NeurIPS 2025)

Official codebase for "Self Forcing: Bridging Training and Inference in Autoregressive Video Diffusion" (NeurIPS 2025 Spotlight)

[CVPR 2025 Highlight] Generative Photography: Scene-Consistent Camera Control for Realistic Text-to-Image Synthesis

[NeurIPS 2025] WorldMem: Long-term Consistent World Simulation with Memory

Open-source unified multimodal model

[ARXIV’25] Learning Video Generation for Robotic Manipulation with Collaborative Trajectory Control