Coding & Development

The SpeechLLM Playbook is a comprehensive resource for exploring SpeechLLMs and neural audio codecs, hosted on Hugging Face Spaces. This application offers in-depth analysis of various speech models, such as Orpheus 3B, LLaSA, and CSM-1B. Users can access visual plots and detailed descriptions of each model's architecture and performance, making it an invaluable tool for researchers and academics in the field of speech technology. Currently a work in progress, it aims to provide a deep dive into the intricacies of these advanced AI models.

ROLO is an open-source recurrent YOLO (You Only Look Once) model designed for simultaneous object detection and tracking. It utilizes the regression capabilities of Long Short-Term Memory (LSTM) networks to interpret visual features and translate them into precise object coordinates. This approach allows ROLO to not only detect objects within a frame but also track their movement over time, making it suitable for applications requiring continuous object monitoring. The project is available on GitHub, indicating its open-source nature and accessibility for developers and researchers.

SpotHero is a mobile application and web platform designed to simplify the process of finding and reserving parking. Users can search for parking spots near their destination, compare prices from various garages, lots, and valets, and pre-pay for hourly, daily, monthly, or event parking. The platform aims to reduce the stress associated with finding parking by offering guaranteed spots and often significant savings compared to drive-up rates. It provides a convenient way to manage reservations, view parking passes, and access customer support directly through the app or website. SpotHero also offers features like business profiles for managing expenses and the ability to earn and use SpotHero credit for future reservations.

awesome-image-captioning is an open-source GitHub repository offering a meticulously curated list of resources focused on image captioning and related fields. It serves as a valuable hub for researchers and practitioners, providing an extensive collection of academic papers categorized by year, from before 2015 up to 2020. The repository also includes information on datasets, image captioning challenges, and popular implementations in frameworks like PyTorch and TensorFlow. Contributions are welcomed via pull requests or email, fostering a collaborative environment for keeping the resource up-to-date and comprehensive.

NIID-Bench is an open-source benchmark designed for experimental studies in federated learning, specifically focusing on scenarios with non-IID (non-independent and identically distributed) data silos. The tool implements four popular federated learning algorithms: FedAvg, FedProx, SCAFFOLD, and FedNova. It supports three types of non-IID settings, including label distribution skew, feature distribution skew, and quantity skew, across nine diverse datasets such as MNIST, Cifar-10, and FEMNIST. Researchers can use NIID-Bench to evaluate and compare the performance of different federated learning algorithms under various challenging data distribution conditions, contributing to advancements in the field. The project also includes follow-up works like FedOV and FedConcat, and hosts a challenge for researchers to test their algorithms.

x-ui-yg is a refined version of the x-ui script, designed to simplify network configuration and management. It incorporates support for the latest Xhttp transport protocol, along with ENC and MLDSA65 encryption for enhanced security. The tool integrates both fixed and temporary Argo tunnels, allowing for co-existence, and includes Psiphon VPN with 30 country options for traffic distribution. It also facilitates the aggregation of various node subscriptions and supports the output of configuration files for popular clients like Sing-box and Clash-Meta, making it a versatile solution for developers and system administrators managing network proxies and VPNs.

MMSkeleton is an open-source toolbox developed by OpenMMLAB, specifically designed for skeleton-based human understanding. It offers a highly extensible framework that systematically organizes code and projects, allowing for adaptation to various tasks and scaling to complex deep models. Key functionalities include 2D and 3D pose estimation, skeleton-based action recognition (like ST-GCN), and action synthesis. The toolbox also supports building custom skeleton-based datasets and creating personalized applications. It is part of the OpenMMLAB project, developed on the ST-GCN research project, and is released under the Apache 2.0 license.

Tonic's Hallucination Space is an AI tool designed to help users identify and mitigate hallucinations in AI-generated content. By providing a piece of context, a related question, and an AI-generated answer, the tool meticulously examines the answer for factual accuracy against the given context. Any portions of the answer that are not supported by the context are clearly marked in red, along with a confidence score indicating the likelihood of a hallucination. This functionality is crucial for developers and QA professionals working with AI models, enabling them to assess the reliability and accuracy of their systems. Hosted on Hugging Face Spaces, it offers an accessible platform for testing and improving AI performance.

Raphtory is an in-memory vectorized graph database engineered in Rust, providing powerful Python APIs for seamless integration. It boasts exceptional speed and scalability, capable of managing hundreds of millions of edges even on a laptop. Users can easily incorporate it into existing pipelines via a simple `pip install`. Key features include time traveling, full-text search, multilayer modeling, and advanced analytics such as automatic risk detection, dynamic scoring, and temporal motifs. Raphtory also supports out-of-memory (on-disk) scaling without performance degradation through its subscription model. It can be run embedded or as a server instance using GraphQL, with a bundled web playground for query experimentation and data visualization.

PyGaze is an open-source, cross-platform Python package designed for the minimal-effort programming of eye tracking experiments. It provides a comprehensive toolbox for researchers in cognitive science and psychology to create and run gaze-contingent or non-gaze-contingent experiments. The tool supports various eye trackers and offers functionalities for data analysis, making it a valuable resource for academic research. PyGaze is freely available to use and modify under the GNU Public License (version 3), emphasizing its commitment to open science and collaborative development within the scientific community.

AI RPG Adventure is an interactive text-based RPG game hosted on Hugging Face Spaces, powered by Pixeltable. Users can choose their genre, character, and actions, and the AI dynamically generates engaging stories and options in response. This creates a personalized and evolving adventure experience, making each playthrough unique. The tool is designed for those who enjoy interactive storytelling and role-playing games, offering a flexible platform to explore various narratives without predefined paths. It leverages AI to provide a responsive and immersive gaming environment.

Deep-reinforcement-learning-with-pytorch is an open-source GitHub repository that offers PyTorch implementations of classic and state-of-the-art deep reinforcement learning algorithms. The project includes implementations of popular methods such as DQN, AC, ACER, A2C, A3C, PG, DDPG, TRPO, PPO, SAC, and TD3. Its primary goal is to provide clear and accessible code, making it easier for individuals to learn and experiment with deep reinforcement learning algorithms. The repository is actively maintained, with plans to add more advanced algorithms and update existing code. It also provides installation instructions and examples for testing the implementations.

JBKenBurns is an iOS library designed to implement the Ken Burns effect, a widely used technique for displaying still photographs in motion pictures. This effect adds dynamic visual interest through slow zooming, panning, and fading transitions between frames. Developers can easily integrate JBKenBurns into their iOS projects to animate an array of UIImage objects or image paths. The library offers control over transition duration, initial delay, looping, and optimization for landscape mode. It also includes a delegate protocol to notify when an image changes or the slideshow ends, providing flexibility for custom interactions within iOS applications.

EJDB2 is an embeddable JSON database engine provided as a C library, designed for applications requiring local JSON data storage and retrieval. It features a simple yet powerful XPath-like query language (JQL) for querying and updating JSON documents, inspired by Unix shell pipes. The engine supports online backups and offers HTTP REST/Websockets endpoints for network access. Built on IOWOW, a persistent key/value storage engine, EJDB2 provides native language bindings and is tested on Linux, macOS, and FreeBSD. Its JQL allows for complex filtering, projections, collection joins, and data modification using JSON patch specifications, including non-standard extensions like increment and add_create.

🇨🇿 BenCzechMark is an AI tool designed for benchmarking AI models by allowing users to submit their model details to a public leaderboard. Users can provide information such as team name, model name, type, parameters, input length, and a link, along with a JSON results file. After validation, the submitted models are added to the leaderboard, facilitating the evaluation and comparison of AI system performance. This platform is hosted on Hugging Face, making it accessible for developers and researchers interested in assessing their AI models against others in a standardized manner. It serves as a valuable resource for understanding the strengths and weaknesses of various AI models.

Bioclip 2 Demo is an interactive application hosted on Hugging Face Spaces, designed for biological research and data exploration. Users can upload images of plants, animals, or other organisms, and the tool will predict their likely taxonomic rank, such as species, genus, or family. This is achieved using a sophisticated large tree-of-life model. The demo also allows users to supply their own taxonomic tree, offering flexibility for specialized research. It serves as a valuable resource for visualization and understanding biodiversity through image analysis, making advanced biological classification accessible.

AstaBench Leaderboard offers a comprehensive platform for viewing and comparing benchmark leaderboards across diverse AI categories. Users can explore performance metrics for models in areas such as literature understanding, code execution, data analysis, and discovery. The tool is hosted on Hugging Face Spaces by AllenAI, providing a centralized location to track and evaluate the advancements in AI model capabilities. It serves as a valuable resource for researchers and developers to assess the effectiveness of different AI systems without requiring any input, simply by browsing the available leaderboards.

Bible AI is a personal Christian AI designed to facilitate exploration of scripture and meaningful conversations about faith, life, and other topics. The tool aims to grow and learn with its users, providing a personalized experience. It offers a platform for engaging with biblical texts and spiritual discussions. Users can sign up for email updates about new features and access an older version of the tool from 2023. The platform also provides information about the team, FAQs, press, media, contact details, terms of service, and privacy policy.

Compare Depth Models is a Hugging Face Space designed for evaluating and comparing different depth estimation models, with a particular focus on Depth Anything and its predecessors. This tool is valuable for AI researchers and computer vision engineers who need to assess the performance and accuracy of various depth models. While the live website currently shows a runtime error, the intention of the tool is to provide a visual comparison of depth outputs from different models, aiding in research and development within the computer vision domain. It serves as a practical demonstration and comparison platform for advanced depth estimation techniques.

WebpageCreator is a user-friendly tool designed to simplify website creation. By leveraging AI, it enables users to generate a complete and functional HTML website with minimal input. Users simply need to provide a brief description of their desired site, specify preferred colors, language, and a company name. The tool then processes this information to deliver a fully designed website, making it ideal for quick prototyping or for individuals and businesses looking to establish an online presence without extensive coding knowledge. It's hosted on Hugging Face Spaces, offering accessibility for various users.

Cad Recode is an innovative AI tool designed to reverse engineer 3D CAD models from mesh files. Users can upload a 3D mesh, and the application processes it by sampling 256 points. From these points, it generates corresponding Python CAD code, effectively translating the visual mesh into programmatic instructions. This generated code can then be used to reconstruct the CAD model, offering a unique approach to converting existing 3D designs into editable, code-based formats. The tool is hosted on Hugging Face Spaces, providing an accessible platform for this specialized conversion task.

angular-youtube-embed is an open-source Angular directive designed to streamline the integration of YouTube video players into web applications. It provides a straightforward way for developers to embed YouTube videos using a simple directive, supporting both video IDs and URLs. The tool offers extensive control over the embedded player, including events for monitoring player state (ready, ended, playing, paused, buffering, queued, error) and functions to manipulate playback (playVideo(), stopVideo()). It also includes utilities like `getIdFromURL` and `getTimeFromURL` for extracting information from YouTube URLs. Developers can customize player parameters, set player dimensions, and implement responsive video layouts, making it a flexible solution for various web development needs.

Brickit is an innovative mobile application designed to help users discover new building possibilities from their existing brick collections. It utilizes advanced scanning technology to identify every piece in a pile of bricks from a photo. The app then generates hundreds of creative project ideas, complete with step-by-step instructions and the exact location of each piece needed. Brickit encourages creativity by allowing users to experiment with different colors and swap out bricks. It also offers a platform for users to submit their own building ideas and share their creations with other enthusiasts. Additionally, Brickit offers a version tailored for schools and camps, called Brickit for Classes, which focuses on hands-on building activities.

Deep-Reinforcement-Learning-Hands-On-Second-Edition is an open-source educational resource published by Packt, designed to help users learn and apply deep reinforcement learning techniques. The GitHub repository provides comprehensive code examples and materials, making it a practical companion for the associated book. It is actively maintained to ensure dependency versions are kept up-to-date, with specific code branches available for major PyTorch versions (e.g., 1.3 and 1.7) to accommodate compatibility needs. The resource includes detailed instructions for setting up a virtual environment using Anaconda, installing PyTorch, and managing other dependencies, making it accessible for hands-on experimentation and learning.