Coding & Development

Ambly is a specialized ClojureScript REPL (Read-Eval-Print Loop) designed for developers building hybrid applications that combine ClojureScript with native iOS, macOS, and tvOS environments. It achieves this by interfacing with embedded JavaScriptCore, allowing for interactive development and debugging directly within these platforms. The tool includes a ClojureScript REPL implementation alongside Objective-C code for seamless integration. Demo applications for iOS, macOS, and tvOS are provided, making it straightforward to set up and experiment with the REPL. Developers can start the Ambly REPL via `cljs.main` and benefit from features like device auto-discovery and configurable connection options.

Bine is a Go library designed for developers to access and embed Tor clients and servers directly within their Go applications. It provides comprehensive support for the Tor controller API, allowing for advanced control and configuration of Tor functionalities. A key feature is its ability to support statically compiled Tor, enabling developers to embed the entire Tor client or server into their binary, simplifying distribution and deployment. Bine also supports v3 onion services, facilitating the creation of hidden services, and offers support for embedded control sockets in newer Tor versions, enhancing security and reducing the need for external control ports. This makes it a powerful tool for building applications that require secure, anonymous communication.

fvcore is a light-weight core library developed by FAIR (Facebook AI Research) that provides essential and common functionalities shared across various computer vision frameworks. It is specifically designed to support projects like Detectron2, PySlowFast, and ClassyVision. The library emphasizes quality, with all components being type-annotated, thoroughly tested, and benchmarked for reliability. Key features include common PyTorch layers, functions, and losses, a hierarchical per-operator flop counting tool, recursive parameter counting, and a stateless, scale-invariant hyperparameter scheduler. fvcore aims to streamline computer vision research and development workflows by offering robust, shared utilities.

NASLib is a modular and flexible framework designed to facilitate Neural Architecture Search (NAS) research by providing a common codebase to the community. It offers high-level abstractions for designing and reusing search spaces, along with interfaces to various benchmarks and evaluation pipelines. This enables researchers to implement and extend state-of-the-art NAS methods with minimal code. The library's modular nature allows for easy innovation on individual components, such as defining new search spaces while reusing existing optimizers, or proposing new optimizers with current search spaces. Developed by the AutoML Freiburg group, NASLib is continuously updated with new search spaces, optimizers, and benchmarks.

OCEval is a compact JIT interpreter designed for Objective-C, offering the capability to dynamically execute Objective-C code, similar to how `eval()` functions in other languages. This tool supports both iOS and OS X development environments and is entirely written in Objective-C. Its development is driven by unit tests, ensuring reliability and functionality. OCEval extends its utility by supporting various low-level APIs, including blocks and C functions, which allows for more flexible and powerful dynamic code manipulation. Developers can use it to dynamically call Objective-C methods, replace method implementations at runtime, and even theoretically build entire applications that can be delivered and updated over a network.

rust-cpp is a specialized build tool and macro designed to bridge the gap between Rust and C++ programming languages. It allows developers to embed C++ code directly into their Rust projects, enabling seamless interoperability. This functionality is particularly useful for integrating existing legacy C++ codebases into new Rust applications or for leveraging powerful C++ libraries within a Rust environment. By facilitating this interaction, rust-cpp enhances development flexibility and can contribute to performance optimization in projects requiring the strengths of both languages. It simplifies the process of combining these two distinct programming paradigms, making it easier for developers to manage mixed-language projects.

rl-book offers the complete source codes for the book "Reinforcement Learning: Theory and Python Implementation." This resource provides a tutorial approach to reinforcement learning, detailing both theoretical concepts and practical Python implementations. It features one-to-one mapping between theory and code, supporting TensorFlow 2 and PyTorch 1&2. The implementations cover a wide range of algorithms, from classic methods like SARSA and Q-Learning to modern deep reinforcement learning techniques such as PPO, DDPG, and SAC. All codes are designed for compatibility across Windows, Linux, and macOS, and can be run on a laptop without requiring a GPU for most examples. The project also includes supporting content like exercise answers and errata for both English and Chinese versions of the book.

tensor-house offers a comprehensive toolkit for rapid readiness assessment, exploratory data analysis, and prototyping diverse modeling approaches within enterprise AI/ML/data science projects. It includes Jupyter notebooks and demo AI/ML applications tailored for specific business needs such as marketing, pricing, supply chain, and smart manufacturing. This resource is designed to help developers and data scientists quickly build and deploy intelligent applications, manage and compare prompts, and integrate external tools. It also provides features for automating workflows, managing code changes, and securing applications, making it a versatile platform for developing and deploying AI solutions.

TornadoVM is a plugin designed for heterogeneous programming in managed languages, specifically aimed at accelerating Java applications. It enables developers to leverage diverse hardware accelerators, such as GPUs and FPGAs, to significantly boost the performance of their code. The framework offers an efficient way to optimize performance across various computing devices, making it a valuable tool for developers looking to enhance the speed and efficiency of their Java-based projects. By abstracting the complexities of heterogeneous hardware, TornadoVM allows developers to focus on their application logic while still benefiting from specialized hardware acceleration.

Gemini Image to Code is a web application leveraging Google's Gemini Pro Vision model to transform visual designs into functional code. Users can upload an image, and the tool processes it to generate corresponding HTML/CSS code that accurately recreates the design. This capability is particularly useful for developers and designers looking to quickly prototype web pages or convert mockups into code. The platform offers a real-time code preview, allowing for immediate visualization of the generated output and facilitating rapid iterations. While the core functionality is image-to-code conversion, the tool operates within the Hugging Face Spaces ecosystem, which offers various pricing tiers for compute resources and storage, indicating that while the application itself might be free to use, the underlying infrastructure costs can vary.

UNeXt-pytorch is the official PyTorch implementation of UNeXt, an MLP-based network specifically designed for rapid medical image segmentation. This tool is ideal for researchers and developers working on medical imaging tasks, particularly those requiring quick processing for point-of-care applications. Based on a MICCAI 2022 paper, it offers a robust and efficient solution for segmenting medical images. The open-source nature of the project, hosted on GitHub, allows for community contributions and flexible integration into existing workflows, providing a strong foundation for advanced medical image analysis.

UniDet is an open-source object detection tool designed to operate across multiple large-scale datasets with an automatically learned unified label space. It was the winning solution of the ECCV 2020 Robust Vision Challenges. The tool offers state-of-the-art performance on datasets such as COCO, Objects365, OpenImages, and Mapillary. A key feature is its ability to predict class labels within this unified space, allowing it to be directly used for testing on novel datasets not included in its training. The repository also provides state-of-the-art baselines for Objects365 and OpenImages. UniDet is built on detectron2, making its inference API familiar to users of that framework.

whatlanggo is a natural language detection library specifically designed for Go applications. This open-source tool boasts support for 84 different languages and is entirely written in Go, ensuring no external dependencies. It is engineered for speed and accuracy, capable of recognizing not only the language of a given text but also its script (e.g., Latin, Cyrillic). The algorithm is based on trigram language models, a particular case of n-grams, as detailed in the Cavnar and Trenkle '94 whitepaper. It also provides a confidence score and a reliability metric based on unique trigrams and the difference between top detected languages. Developers can easily integrate it into their Go projects and utilize options for blacklisting or whitelisting specific languages.

vscode-browse-lite is an embedded browser extension designed for Visual Studio Code, offering developers a seamless way to preview web pages directly within their IDE. This tool enhances the development workflow with features like faster page refreshing, ensuring immediate feedback on changes. It is dark mode aware and theme-aware, integrating smoothly with the user's VS Code environment. Crucially, it includes built-in devtools support, allowing for direct debugging and inspection of web content. The extension also boasts extendable actions and the ability to re-open pages in a system browser. Notably, vscode-browse-lite is lightweight, significantly smaller than its predecessor, and does not collect telemetry, prioritizing user privacy and performance.

QR Code Maker & QR Scanner is a versatile mobile application designed for both scanning and generating QR codes. It offers robust scanning capabilities, allowing users to quickly and reliably scan QR codes directly from their device's camera or from images stored in their gallery. The app also features a comprehensive QR code generator, enabling users to create custom QR codes for various data types such as vCard, WiFi, SMS, and URLs. Generated codes can be customized with different colors, sizes, and error correction levels, and then saved or shared as images. A convenient scan history feature ensures that all previously scanned codes are saved for easy access and re-scanning. The application supports multiple languages, providing a broad accessibility for users worldwide.

Algorithmic-trading-python is a comprehensive open-source repository designed to accompany freeCodeCamp's YouTube course on algorithmic trading in Python. It offers practical resources for individuals looking to understand and implement algorithmic trading strategies. The repository guides users through fundamental concepts, API basics, and the development of various trading models. Key sections include building an equal-weight S&P 500 index fund, as well as quantitative momentum and value investing strategies. This resource is ideal for students and developers who want to gain hands-on experience in financial programming and automated trading.

Algorithmic Trading with Python is a GitHub repository containing the complete source code for the 2020 book by Chris Conlan. This resource is invaluable for researchers and developers interested in algorithmic trading, providing practical Python implementations of key concepts. It includes stand-alone scripts for performance metrics to evaluate trading strategies, common technical indicators implemented in pure Pandas, and methods for converting these indicators into ternary signals. The repository also features a generic grid search wrapper for numeric optimization, object-oriented building blocks for portfolio simulation, and a generic wrapper for multi-core repeated K-fold cross-validation. Additionally, it offers free-to-use simulated End-of-Day stock data and alternative data streams, making it a comprehensive toolkit for learning and applying algorithmic trading principles.

algotrading is an open-source algorithmic trading framework specifically designed for cryptocurrencies, written in Python. It provides a comprehensive set of tools for building and running trading bots, backtesting strategies, and assisting with trading decisions, including defining stop losses and trailing stop losses. The framework can operate with data directly from crypto exchange APIs, databases, or CSV files, supporting both data-driven and event-driven systems. It offers three operating modes: Realtime for live trading or simulation, Tick-by-tick for detailed strategy testing, and Backtest for evaluating strategies with historical data. Users can define custom entry and exit functions, plot trading data, and log performance for analysis.

awesome-gemini-ai is an open-source repository offering a curated collection of high-performance prompts, use cases, and examples specifically designed for Google's Gemini 1.5 Pro and Ultra models. Sourced from platforms like X (Twitter), Reddit, and top prompt engineers, this resource focuses on maximizing Gemini's capabilities for various tasks. Users can find prompts for web development and coding, UI/UX design generation, creative experiments, and even multilingual applications. The collection emphasizes utilizing Gemini's reasoning for complex applications, such as generating award-winning websites or simulating operating systems, making it a valuable resource for developers and designers looking to push the boundaries of AI-driven creation.

Deepdive-llama3-from-scratch is an open-source project designed to help developers comprehensively understand and implement the Llama3 model from scratch. Building upon an existing project, it offers significant improvements in structural optimization, code annotations, and principle explanations. The tool provides detailed derivations for core concepts like KV-Cache and tracks matrix dimensions throughout calculations, making complex processes easier to grasp. It includes code files in both English and Chinese, ensuring accessibility for a broader audience. Users can follow along to load the Llama3 model, tokenizer, and perform inference step-by-step, with a focus on understanding the underlying mechanisms rather than just execution.

MiniMax-M2 is an open-source, compact, fast, and cost-effective Mixture-of-Experts (MoE) model designed for advanced coding and agentic workflows. With 230 billion total parameters and only 10 billion active parameters, it offers high performance in tasks like multi-file edits, coding-run-fix loops, and test-validated repairs, while maintaining powerful general intelligence. The model is engineered for end-to-end developer workflows and excels in agent performance, planning and executing complex, long-horizon toolchains across shell, browser, retrieval, and code runners. Its efficient design leads to lower latency, lower cost, and higher throughput, making it ideal for interactive agents and batched sampling. MiniMax-M2 is available via API and its weights are open-source for local deployment.

Ask Command functions as a tech blog and command resource, offering a range of articles and tutorials focused on practical technical knowledge. The content covers diverse areas such as Linux user management, understanding `sudo` commands, process termination in terminals, network port checking with `netstat`, and fundamental programming principles like SOLID. It also delves into web development topics, including clean code practices, using Chrome DevTools, GraphQL, and JavaScript debugging. The platform is designed to assist users in navigating common technical challenges and enhancing their programming and system administration skills through clear, instructional content.

GeoParquet is an Open Source specification that defines how to store geospatial vector data, including points, lines, and polygons, within the Apache Parquet columnar storage format. This standardization aims to enhance geospatial interoperability across various tools that utilize Parquet, facilitating advanced cloud-native geospatial workflows. The specification is developed in parallel with GeoArrow to enable cross-language in-memory analytics. It supports multiple spatial reference systems, allows for multiple geometry columns, and offers great compression for smaller file sizes. GeoParquet is particularly well-suited for read-heavy analytic workflows and data partitioning, though it is not ideal for write-heavy interactions. It is in the process of becoming an official OGC standard.

irl-imitation provides a Python/Tensorflow implementation of several Inverse Reinforcement Learning (IRL) algorithms. Key algorithms include Linear Inverse Reinforcement Learning (Ng & Russell, 2000), Maximum Entropy Inverse Reinforcement Learning (Ziebart et al., 2008), and Maximum Entropy Deep Inverse Reinforcement Learning (Wulfmeier et al., 2015). The tool also features implementations for 2D and 1D gridworld Markov Decision Processes (MDPs) and a Value Iteration solver. It's designed for researchers and developers working on reinforcement learning and imitation learning tasks, offering a practical codebase for experimenting with and applying these advanced algorithms.