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

AnyRecon: Arbitrary-View 3D Reconstruction with Video Diffusion Model

SkinTokens: A Learned Compact Representation for Unified Autoregressive Rigging

[SIGGRAPH 2026] AniGen: Unified S^3 Fields for Animatable 3D Asset Generation

JoyAI-Image is the unified multimodal foundation model for image understanding, text-to-image generation, and instruction-guided image editing.

Implementation of "FullCircle: Effortless 3D Reconstruction from Casual 360° Captures"

PoseMaster: A Unified 3D Native Framework for Stylized Pose Generation

Code for SPREAD: Spatial-Physical REasoning via geometry Aware Diffusion

Code of Strips as Tokens: Artist Mesh Generation with Native UV Segmentation. ACM Transactions on Graphics (SIGGRAPH 2026)

[CVPR 2026 Highlight] It's the official repository of "NI-Tex: Non-isometric Image-based Garment Texture Generation".

Official implementation of Kimodo, a kinematic motion diffusion model for high-quality human(oid) motion generation.

Relightable multi-view 3D reconstruction with disentangled illumination and spatially varying PBR materials

[CVPR 2026] LoST: Level of Semantics Tokenization for 3D Shapes

Generate high resolution videos with a custom voice and appearance, based on LTX-2/LTX-2.3 + Identity In-Context LoRA

SegviGen: Repurposing 3D Generative Model for Part Segmentation

Official inference framework for 1-bit LLMs

An open-source long-horizon SuperAgent harness that researches, codes, and creates. With the help of sandboxes, memories, tools, skill, subagents and message gateway, it handles different levels of…

(CVPR 2026) PixARMesh: Autoregressive Mesh-Native Single-View Scene Reconstruction

[CVPR 2026] ZipMap: Linear-Time Stateful 3D Reconstruction via Test-Time Training

OmniTransfer: All-in-one Framework for Spatio-temporal Video Transfer

Official implementation of ImageCritic (CVPR 2026)

Elastic Diffusion Transformer: Accelerating SOTA generation models (e.g., Qwen-Image, Hunyuan3d ) through adaptive computation

Pre-built wheels that erase Flash Attention 3 installation headaches.

[CVPR 2026] LaS-Comp: Zero-shot 3D Completion with Latent–Spatial Consistency

CADAM is the open source text-to-CAD web application