DeepClaude is a high-performance open-source AI tool based on a deep integration of the DeepSeek R1 and Claude models, combining inference, creativity, and code generation capabilities. DeepClaude provides zero-latency, instantaneous responses and supports...
AuraFusion360 is a novel reference-based method for 360° borderless scene restoration, primarily used in fields such as virtual reality and architectural visualization. It achieves high-quality restoration of 3D scenes represented by Gaussian scattering...
The RWKV-7-2.9B model (RWKV-7-World-2.9B-V3) is an advanced RNN large-language model launched by the RWKV Foundation. Trained on the RWKV World V3 dataset, it has 2.9 billion parameters and supports all languages in the world. The model...
Data Formulator is an open-source, AI-driven data visualization tool from Microsoft Research that helps users quickly create rich data visualizations through simple interactions and commands. It combines a graphical user interface (GUI) with...
Magic 1-For-1 is a high-efficiency video generation model developed by Peking University, Hedra Inc., and Nvidia. It rapidly generates high-quality video clips by optimizing memory consumption and inference latency. The model transforms complex text into video...
Praison AI is a production-grade multi-agent framework that uses low-code solutions to create and manage AI agents, automating and solving tasks ranging from simple to complex. Praison AI supports various process types, including sequential, delimited, and sequential workflows...
Goedel-Prover is an open-source large-scale language model (LLM) developed by Princeton University, Tsinghua University, and other institutions. It is used to automate the generation of formal proofs for mathematical problems. Based on natural language processing...
CineMaster is a text-to-video generation framework launched by Kuaishou, featuring 3D perception capabilities. Similar to a video version of ControlNet, it allows users to precisely control the position of objects and camera movement in videos using various control signals. ...
GAS (Generative Avatar Synthesis from a Single Image) is a technique proposed by researchers at Carnegie Mellon University, the Shanghai Artificial Intelligence Laboratory, and Stanford University to generate high-quality, viewpoint-oriented avatars from a single image...
OpenThinker-32B is an open-source inference model jointly developed by Stanford University, UC Berkeley, the University of Washington, and other institutions. It boasts 32.8 billion parameters and supports a context length of 16,000 tokens. The model uses only 114k of data...
Social Media Agent is an intelligent social media content management tool that generates Twitter and LinkedIn posts based on input URLs. It uses a human-computer interaction process to handle social media platform authentication...
Matrix-Zero is a world model developed by Kunlun Tech, comprising two sub-models: a large-scale 3D scene generation model and a large-scale interactive video generation model. Matrix-Zero can transform user-input images into freely explorable, realistic 3D models...
Sa2VA is a multimodal large language model jointly developed by ByteDance, the University of California, Merced, Wuhan University, and Peking University. It combines SAM2 and LLaVA to achieve dense, fine-grained understanding of images and videos.
WebLI-100B is a massive dataset from Google DeepMind containing 100 billion image-text pairs, used for pre-training visual language models (VLMs). WebLI-100B is an expanded version of the WebLI dataset, based on data from the WebLI dataset...
BAG (Body-Aligned 3D Wearable Asset Generation) is an innovative 3D wearable asset generation technology jointly proposed by the Chinese University of Hong Kong and Tencent. It combines a multi-view image diffusion model and a control network (Control...
Tonggu Model is an AI language model developed by the Deep Learning and Visual Computing Laboratory (SCUT-DLVCLab) at South China University of Technology, focusing on the processing of classical Chinese texts. It is incrementally pre-trained based on Baichuan 2-7B-Base and uses...
ENEL (Exploring the Potential of Encoder-free Architectures in 3D LMMs) is an innovative encoder-free 3D large-scale multimodal model (3D LMM) that addresses the limitations of traditional encoder architectures in 3D understanding tasks...
AnyCharV is a character-controlled video generation framework jointly developed by the Chinese University of Hong Kong, Tsinghua University Shenzhen Graduate School, and the University of Hong Kong. It can combine arbitrary reference character images with target-driven video to generate high-quality character-controlled videos...
NobodyWho is a plugin designed for the Godot game engine that enables interactive novel creation through a locally running Large Language Model (LLM). It supports Windows, Linux, and macOS platforms, and users can use it without an internet connection...
PIKE-RAG (Simplified Knowledge and Rational Augmented Generation) is a search-enhanced generation framework developed by Microsoft Research Asia. It addresses the limitations of traditional RAG systems in complex industrial applications...
Collaborative Gym (Co-Gym) is a framework focused on human-agent collaboration, supporting real-time interaction and collaboration between humans and AI agents. Through both simulated and real-world experimental conditions,...
TIGER (Time-frequency Interleaved Gain Extraction and Reconstruction Network) is a lightweight speech separation model proposed by a research team at Tsinghua University. It utilizes a time-frequency interleaved modeling strategy...
NPOA is a powerful, open-source public opinion monitoring tool that helps users monitor online public opinion in real time and understand public sentiment towards specific topics or brands. Based on advanced natural language processing technology, the system can analyze large amounts of online data...
Long-VITA is an open-source multimodal model developed by Tencent YouTu Lab, Nanjing University, and Xiamen University. It can handle long text inputs with over 1 million tokens and performs exceptionally well on short text tasks. Long-VITA is based on staged training...