[CVPR 2026] OccAny: Generalized Unconstrained Urban 3D Occupancy. The first Unified Framework for Generalized 3D Occupancy Prediction. Supports SAM2/SAM3, MUSt3R & Depth Anything 3.

[CVPR'26] AnomalyVFM -- Transforming Vision Foundation Models into Zero-Shot Anomaly Detectors

[CVPR 2026] This is the official PyTorch implementation of "MoDES: Accelerating Mixture-of-Experts Multimodal Large Language Models via Dynamic Expert Skipping".

[CVPR 2026] Wavelet-based Frame Selection by Detecting Semantic Boundary for Long Video Understanding

[CVPR 2026] Official PyTorch implementation of "Training-Only Heterogeneous Image-Patch-Text Graph Supervision for Advancing Few-Shot Learning Adapters" (TOGA).

CVPR 2026 | Structure-aware label-free test-time calibration for cross-subject EEG-to-image retrieval.

Unofficial PyTorch reproduction for DeltaTok and DeltaWorld generative world modeling with compact delta tokens.

Unofficial PyTorch reproduction for LongVT, long-video reasoning with native tool calling.

[CVPR 2026] StaR-KVQA: Structured Reasoning Traces for Implicit-Knowledge Visual Question Answering

Unofficial PyTorch reproduction for OmniGen2, instruction-aligned multimodal generation and editing.

Unofficial PyTorch reproduction for Demo2Tutorial, multimodal software tutorial generation from human experience.

Differentiable Laplacian Matrix Guided Superpixel Segmentation CVPR 2026

Evaluation Framework for Talking Head Video Generation

Unofficial PyTorch reproduction for SAGE, agentic 3D scene generation for embodied AI.

Unofficial PyTorch reproduction for LLaDA-V, a diffusion-based multimodal large language model with visual instruction tuning.