Research & Education

Fish Audio S1 is an AI audio tool available on Hugging Face Spaces, designed to convert written text into realistic spoken audio. Users can easily input text, customize audio settings such as speed and tone, and then generate high-quality spoken output. While the current live website indicates a runtime error, the tool's core functionality is text-to-speech, making it suitable for various audio processing and experimentation needs. It aims to provide an accessible platform for exploring AI-driven audio manipulation, particularly for those interested in generating voiceovers or spoken content from text.

Orca is a comprehensive platform designed to enhance the Minecraft experience for players, friends, and creators. It allows players to launch any mod or modpack with a single click, eliminating manual setup for Fabric, Forge, NeoForge, and Quilt. For communities, Orca provides instant server hosting, including free local servers or cloud hosting from $12/month, with AI chat for adding mods and plugins. Creators can leverage Orca's built-in AI development agent to generate mods, plugins, data packs, and resource packs from natural language descriptions, complete with code, textures, models, and in-game testing by OrcaBot. The platform also includes a 3D model editor, AI texture generation, and one-click publishing to CurseForge, Modrinth, and GitHub.

Flux Kontext FaceLORA is an AI tool designed for transforming portrait photos into stylized versions. Users can upload a portrait and choose from a gallery of art styles to apply to their image. The tool leverages LoRA (Low-Rank Adaptation) techniques for image generation, allowing for experimentation with Face ID models. Beyond pre-defined styles, it offers options to add custom details and adjust various settings, providing fine-grained control over the final output. This makes it suitable for individuals interested in exploring AI image generation and stylized photo transformations.

Ought is a product-driven research lab dedicated to advancing AI safety and scaling up good reasoning. The organization develops mechanisms for delegating high-quality reasoning to advanced machine learning systems, with the goal of ensuring that future advances in ML automatically increase collective intelligence. Ought has spun off Elicit, which is now an independent public benefit corporation. Their work includes creating resources like the Interactive Composition Explorer (ICE) and the Factored Cognition Primer, and they emphasize the importance of product builders in AI safety. Ought operates as a non-profit with 501(c) status, accepting donations to support its mission.

Florence2 WebGPU is an innovative AI tool that leverages the Florence-2 base model to provide advanced image analysis capabilities directly within your web browser. Users can upload a photo, and the application will automatically generate a clear, natural-language caption describing the scene. Beyond just captioning, it also identifies and extracts key objects and their attributes, presenting them as a concise list. This tool is designed for accessibility, running efficiently with WebGPU technology, eliminating the need for complex technical setups or external dependencies. It's ideal for researchers, developers, and anyone interested in exploring AI models for image processing and understanding.

FlowEdit is a powerful AI tool that enables users to edit images using text prompts, leveraging advanced diffusion models for inversion-free text-based editing. This Gradio demo, hosted on Hugging Face, allows for a seamless workflow: upload an image, describe its current content, and then input a new textual description to transform the image according to your specifications. It's designed to showcase the capabilities of text-based image manipulation, offering a unique approach to photo editing without requiring complex inversion techniques. This makes it accessible for experimenting with creative image transformations and exploring the potential of AI in visual content creation.

FreeU is an AI tool hosted on Hugging Face Spaces, specifically designed for enhancing the quality of images generated by diffusion models. While the live website currently displays a runtime error, indicating a technical issue with its deployment, its intended purpose is to provide a method for improving AI-generated visuals. This tool would typically be utilized by AI developers and researchers who are working on or experimenting with diffusion pipelines and seek to achieve higher fidelity or more aesthetically pleasing outputs from their models. Its availability on Hugging Face Spaces suggests it is intended to be accessible for community use and experimentation, likely offering a free or open-source approach to AI model enhancement.

FineWeb is a specialized tool designed to extract and refine high-quality text data from the vast expanse of the web. It focuses on creating meticulously curated datasets, such as FineWeb and FineWeb-Edu, which are crucial for training advanced AI models. The tool's primary purpose is to assist researchers and data scientists in sourcing relevant, clean, and high-quality information, thereby streamlining the data preparation phase for various AI and machine learning projects. The project emphasizes the quality of the extracted data, ensuring it is suitable for demanding academic and research applications.

GAIA Leaderboard provides a platform for evaluating and comparing the performance of AI chatbot models. Users can submit details about their AI model and upload a JSONL file containing its answers to the GAIA benchmark tasks. The application then scores these answers against reference solutions, records the results, and updates a public leaderboard. This tool is invaluable for AI researchers and developers who need to benchmark different AI models, track progress in chatbot development, and understand how their models stack up against others in the field.

Rolli is an AI-powered narrative intelligence platform designed for communications, security, and research teams. It monitors 8+ social platforms including X, Reddit, YouTube, Facebook, Instagram, Threads, Bluesky, and LinkedIn. Rolli IQ scores every social signal for authenticity, detecting coordinated inauthentic behavior and filtering out manipulated engagement. This allows users to act on verified data, not noise, reducing false alarms by over 80%. Unlike traditional social listening tools that count mentions, Rolli verifies what's real, flagging bot-driven spikes and synthetic amplification to surface genuine signals in minutes. It offers a free trial and paid plans starting at $99/month.

Galactica Demo is a platform hosted on Hugging Face Spaces, designed for users to explore and interact with AI models. It serves as a demonstration environment where individuals can test the functionalities and capabilities of various AI agents. The tool is suitable for AI enthusiasts, researchers, and developers who wish to experiment with AI models in a practical setting. As a Hugging Face Space, it leverages the community-driven ecosystem for machine learning applications, offering a readily accessible way to engage with AI technology. The platform is currently sleeping due to inactivity, indicating its nature as a demo or experimental space.

LLMSurvey is the official GitHub repository for the survey paper "A Survey of Large Language Models," offering a curated collection of papers and resources on Large Language Models (LLMs). This platform is designed to help researchers and students stay updated with the latest advancements in the field. It includes a timeline of LLMs, lists of publicly available and closed-source models, and detailed paper lists with links to arXiv and checkpoints. The repository also features new content on long CoT reasoning, trends in LLM-related arXiv papers, technical evolution of GPT-series models, and an evolutionary graph of the LLaMA family. Additionally, it provides useful tips for designing prompts and details on instruction tuning and ability evaluation experiments.

Research Commons is a comprehensive platform designed to enhance and streamline the academic research process for individuals at all career stages. It offers a collaborative environment alongside a suite of powerful tools. Key features include guided research publication, which helps users navigate the complexities of academic publishing, and AI-powered paper analysis, providing in-depth insights and summaries of research papers. The platform also assists with automated citation management, ensuring accuracy and saving valuable time. Furthermore, Research Commons facilitates collaborative projects, allowing researchers to work together efficiently on shared endeavors, making it an invaluable resource for academic professionals.

Gligen Demo is a platform designed to showcase the capabilities of the Gligen model for image generation. It provides a public interface for users to interact with and test the model's functionalities. While the current live website indicates a runtime error, suggesting temporary unavailability, the tool's purpose is to allow exploration of advanced image generation techniques. It is particularly suitable for AI researchers, developers, and creative professionals who are interested in understanding and experimenting with cutting-edge AI models in the field of visual content creation. The demo aims to provide insights into how the Gligen model can be utilized for various image generation tasks.

GOT OCR Transformers is a demonstration of the GOT-OCR 2.0's Transformers implementation, hosted on Hugging Face. This application enables users to perform Optical Character Recognition (OCR) by uploading an image and selecting their preferred OCR method. It is designed for extracting text from various image formats, providing a straightforward interface for text recognition tasks. While the current live website indicates a runtime error, the tool's core functionality is centered around advanced OCR capabilities, making it useful for researchers and developers in the field of text extraction and document processing.

GLM OCR Demo is a multimodal OCR model designed for complex document understanding, available as a Hugging Face Space. This application allows users to upload an image and specify whether they want to extract plain text, mathematical formulas, or table data. After processing, the recognized content is returned in an editable format. This tool is particularly useful for researchers and developers working with OCR technology who need to analyze intricate documents, offering a flexible solution for various data extraction needs from visual inputs.

LongCite is an open-source project designed to enhance Large Language Models (LLMs) by enabling them to generate fine-grained, sentence-level citations within long-context question answering scenarios. This capability significantly improves the verifiability and accuracy of LLM outputs. The project offers two pre-trained models, LongCite-glm4-9b and LongCite-llama3.1-8b, which support up to 128K context and are based on GLM-4-9B and Meta-Llama-3.1-8B, respectively. LongCite also provides a CoF (Coarse to Fine) pipeline for automated SFT data construction, allowing users to generate high-quality long-context QA instances with fine-grained citations. Additionally, it includes an automatic benchmark, LongBench-Cite, for evaluating citation quality and response correctness.

LLM_MultiAgents_Survey_Papers is a GitHub repository dedicated to surveying the rapidly evolving field of Large Language Model (LLM) based Multi-Agents. It provides a curated collection of research papers, offering an overview of current progress and identifying key challenges. The repository categorizes these papers into five main streams: Multi-Agents Framework, Multi-Agents Orchestration and Efficiency, Multi-Agents for Problem Solving, Multi-Agents for World Simulation, and Multi-Agents Datasets and Benchmarks. This structured approach helps researchers and academics navigate the extensive literature. The project also includes a summarized LLM-based Multi-Agents architecture and an overview table, making it a valuable resource for understanding the landscape of multi-agent systems powered by large language models.

DAIMLAS is a leading platform designed to build, manage, and grow national and local artificial intelligence ecosystems. It serves governments by helping implement national AI strategies, enterprises by building internal AI ecosystems and orchestrating AI teams, and academia by connecting them to global AI opportunities, jobs, and funding. The platform focuses on creating inclusive pathways into AI for all stakeholders, including upstream resources, downstream channels, and subsidiary resource providers. DAIMLAS facilitates the creation of AI Centers of Excellence and brings together the 8 key AI stakeholders in a local ecosystem to platformize strategies and achieve AI goals. It also addresses common challenges faced by AI talent, such as gaining real-world experience and working on complex AI projects in teams.

MARL-Papers offers a meticulously curated collection of research papers focused on multi-agent reinforcement learning (MARL). This resource is invaluable for researchers and practitioners, providing an organized overview of the field's foundations, modern approaches, and applications. The collection includes tutorials, review papers, and research papers, all sorted chronologically for easy navigation. It covers various sub-topics within MARL, such as joint action learning, cooperation and competition, coordination, security, self-play, learning to communicate, transfer learning, imitation, inverse reinforcement learning, meta-learning, and applications in Large Language Models (LLMs) and robotics. The platform encourages community contributions, making it a dynamic and up-to-date repository for MARL literature.

GLM 130B is an AI tool that provides access to the GLM-130B large language model for research and experimentation. Users can interact with the model by typing prompts in either English or Chinese. The tool supports fill-in-the-blank tasks using `[MASK]` and continuation tasks using `[gMASK]`. This platform is ideal for researchers and developers looking to explore the capabilities of a large-scale AI model. It offers a straightforward interface for generating text and understanding model behavior, making it a valuable resource for those working in natural language processing and AI development.

GODEL Demo is an AI chatbot demonstration hosted on Hugging Face Spaces by Microsoft. This tool offers users an opportunity to interact directly with a conversational AI model, allowing them to test its capabilities and observe its responses in real-time. While the current status indicates a build error, the intention is to provide a platform for exploring and experimenting with advanced AI models. It serves as a valuable resource for those interested in understanding the practical application and performance of AI in conversational contexts, offering a hands-on experience with cutting-edge technology.

GenAI-Bench Dataset Viewer is a Hugging Face Space designed for exploring and analyzing the GenAI-Bench dataset. Users can browse and filter a vast collection of images based on both basic and advanced skills, providing a comprehensive view of the dataset's contents. The tool facilitates the comparison of images generated by various AI models and includes human ratings for deeper analysis. This interactive viewer is particularly useful for researchers and developers working on generative AI models, offering a visual and interactive way to understand model performance and data characteristics within the GenAI-Bench framework.

The Stanford Institute for Human-Centered Artificial Intelligence (HAI) is dedicated to guiding and developing AI technologies that prioritize human well-being and societal benefit. HAI brings together diverse expertise from fields such as business, economics, law, and medicine to foster interdisciplinary research and collaboration. The institute aims to serve as a global hub for leading AI thinkers, researchers, and policymakers, facilitating discussions and advancements in the ethical and practical applications of AI. Through its initiatives, HAI contributes to the understanding and responsible implementation of AI across various sectors, promoting discovery and learning in this rapidly evolving field.