ACTalker is an end-to-end video diffusion framework for generating realistic speaker head videos. It supports single-signal or multi-signal control, such as audio and facial expressions. The core architecture includes a parallel Mamba structure, utilizing multiple branches to...
SkyReels-A2 is a controllable video generation framework developed by Kunlun Tech. It supports combining any visual elements (such as people, objects, and backgrounds) into a composite video based on text prompts, strictly maintaining consistency with the reference image of each element...
ChildMandarin is a Mandarin speech dataset for children aged 3-5, jointly developed by the Beijing Academy of Artificial Intelligence (BAAI) and the Human Language Technology Lab (HLT Lab) of the School of Computer Science at Nankai University. The dataset contains 41.25 hours of speech...
MagicColor is a multi-instance line art coloring framework developed by the Hong Kong University of Science and Technology, supporting efficient color addition to line art. MagicColor addresses the problem of insufficient multi-instance data by employing a self-supervised training strategy and an instance-guided module...
Step-R1-V-Mini is StepStar's latest multimodal inference model. It supports image and text input and text output, possesses excellent instruction compliance and versatility, and can perform complex inference tasks with high-precision image perception. The model...
SeniorTalk is the world's first Chinese-language dialogue voice dataset for the extremely elderly, jointly developed by the Academy of Artificial Intelligence of Technology (AIAI) and the Human Language Technology Lab (HLT Lab) of the School of Computer Science at Nankai University. The dataset contains 202 individuals aged 75 and above...
HiDream-I1 is an open-source AI image generation model developed by the HiDream.ai team. It boasts 1.7 billion parameters and is licensed under the MIT license. The model excels in image generation quality and cue word adherence, supporting realistic and cartoonish images...
GitHub MCP Server is an official server tool from GitHub based on the Model Context Protocol (MCP). GitHub MCP Server seamlessly integrates with the GitHub API, providing developers with advanced automation...
Nova Sonic is a new generative AI speech model from Amazon. It integrates speech understanding and generation capabilities into a single model, adjusting the generated speech response based on the speaker's intonation, style, and other acoustic context, enabling dialogue...
DeepCoder-14B-Preview is a large-scale code generation model jointly open-sourced by Agentica and Together AI, finely tuned from Deepseek-R1-Distilled-Qwen-14B. DeepCoder-14B-Preview is based on...
Dream-7B is a diffusion-based inference model jointly developed by the University of Hong Kong and Huawei Noah's Ark Lab. It is currently the most powerful open-source diffusion-based large language model. Dream-7B's training data includes text, mathematical data, and code, and pre-training uses 5...
Llama Nemotron is a series of inference models from NVIDIA, focusing on inference and various intelligent agent (agentic AI) tasks. The models are based on the open-source Llama model, post-trained by NVIDIA, and incorporate inference capabilities...
Miracle F1 is a brand-new AI image generation model launched by Meitu WHEE. It boasts powerful image generation capabilities, able to precisely handle light, shadow, space, and material effects like a photographer, architect, and painter, generating realistic images...
Lipsync-2 is the world's first zero-shot lip-sync model from Sync Labs. It requires no pre-training for specific speakers and can learn instantly to generate lip-sync effects that match unique speaking styles.
EasyControl is an open-source, high-efficiency, and flexible control framework based on the Diffusion Transformer (DiT) architecture from Tiamat AI. It independently processes conditional signals through a lightweight conditional injection LoRA module...
OmniSVG is the world's first end-to-end multimodal SVG (Scalable Vector Graphics) generation model, jointly developed by Fudan University and StepFun. Based on a pre-trained Visual Language Model (VLM), it utilizes an innovative SVG tokenization method...
A2A (Agent2Agent) is Google's first open-source standard intelligent agent interaction protocol, enabling AI agents built on different frameworks and from different vendors to collaborate. A2A breaks down system silos based on standardized communication methods, allowing intelligent agents to collaborate securely...
DCEdit is a novel two-layer control image editing method jointly developed by Beijing Jiaotong University and Meitu 2MT Lab. DCEdit is based on the Precise Semantic Localization (PSL) strategy, using visual and text self-attention to optimize cross-attention...
LocAgent is a framework jointly developed by Stanford University, Yale University, and the University of Southern California, focusing on code location tasks. It helps developers quickly and accurately find the parts of the codebase that need modification. LocAgent will...
Multi-SWE-bench is the first multi-language code fixing benchmark open-sourced by ByteDance's Doubao Big Model team. Building upon SWE-bench, it is the first benchmark to cover seven major programming languages besides Python, including Java, TypeScript, etc.
Gemini 2.5 Flash is a high-efficiency, low-latency AI model from Google, built upon the Gemini 2.5 model. Gemini 2.5 Flash introduces thinking capabilities while maintaining low latency and cost-effectiveness.
DevDocs is an open-source technical documentation crawling and processing tool designed for programmers and developers. Based on intelligent web crawling technology, it quickly crawls and organizes technical documents, reducing the time needed to understand them from weeks to hours. DevDocs...
AutoRAG is a fully managed search augmentation generation (RAG) pipeline from Cloudflare that helps developers easily integrate context-aware AI into their applications without managing infrastructure.
MoLing is a dependency-free, local office automation assistant based on an MCP server used by computers and browsers. MoLing interacts with the system using operating system APIs and supports file system operations (such as read/write, merge, ...).