WBench: A Comprehensive Multi-turn Benchmark for Interactive Video World Model Evaluation

high-performance inference and serving library for interactive autoregressive video and world models

HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D Worlds

From Automated Idea Factory to Realization

Free Motion Capture for Everyone 💀✨

USP: Unified (a.k.a. Hybrid, 2D) Sequence Parallel Attention for Long Context Transformers Model Training and Inference

Steering Forward-Process Reinforcement Learning for Distilled Autoregressive Video Models

[CVPR 2026 Highlight] Reward Forcing: Efficient Streaming Video Generation with Rewarded Distribution Matching Distillation

A light-weight and powerful meta-prompting, context engineering and spec-driven development system for Claude Code by TÂCHES.

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

Helios: Real Real-Time Long Video Generation Model

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflo…

Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)

[ICLR 2026] UniVideo: Unified Understanding, Generation, and Editing for Videos

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

[ICLR 2026] Official Implementation of "reAR: Rethinking Visual Autoregressive Models via Generator-Tokenizer Consistency Regularization"

Codebase for InfoTok: Adaptive Discrete Video Tokenizer via Information-Theoretic Compression

Advancing Open-source World Models

PyTorch implementation of JiT https://arxiv.org/abs/2511.13720

Training API and CLI

Official Python inference and LoRA trainer package for the LTX-2 audio–video generative model.

Official repository for LTX-Video

PyTorch implementation of NEPA

Qwen-Image-Layered: Layered Decomposition for Inherent Editablity

[CVPR 2026] This repository is the code of our paper "VA-Pi: Variational Policy Alignment for Pixel-Aware Autoregressive Generation"

LLM API 管理 & 分发系统，支持 OpenAI、Azure、Anthropic Claude、Google Gemini、DeepSeek、字节豆包、ChatGLM、文心一言、讯飞星火、通义千问、360 智脑、腾讯混元等主流模型，统一 API 适配，可用于 key 管理与二次分发。单可执行文件，提供 Docker 镜像，一键部署，开箱即用。LLM API management & k…

SLA: Beyond Sparsity in Diffusion Transformers via Fine-Tunable Sparse–Linear Attention

Unofficial implementation of Titans, SOTA memory for transformers, in Pytorch

🚀 Efficient implementations for emerging model architectures

[NeurIPS 2025 Oral]Infinity⭐️: Unified Spacetime AutoRegressive Modeling for Visual Generation