Coding & Development

Prompt-patterns is an Open Source GitHub repository dedicated to prompt engineering patterns, offering a structured approach to designing effective prompts for AI models. The resource categorizes patterns into foundational types like specific instruction, instruction template, proxy, and demonstration, along with various sub-patterns. It aims to help users understand how to frame their thinking when interacting with AI, drawing analogies from traditional software design patterns. The project encourages community contributions through issue reporting, chapter writing, translation, and sharing, fostering a collaborative environment for improving prompt engineering practices.

skrub (formerly dirty_cat) is a powerful open-source Python library designed to streamline machine learning tasks when working with dataframes. It offers a comprehensive suite of functionalities for data cleaning, preprocessing, and wrangling, making it easier to prepare tabular data for machine learning models. The library is built to handle common challenges associated with dirty or inconsistent data, ensuring that data scientists and developers can focus on model building rather than tedious data preparation. skrub integrates seamlessly into existing Python-based machine learning pipelines, providing efficient and robust solutions for data preparation.

tf_unet is an open-source project offering a generic U-Net implementation developed with TensorFlow, specifically designed for image segmentation tasks. Originally used for Radio Frequency Interference mitigation, this tool is highly adaptable and can be applied to diverse imaging data, from detecting circles in noisy images to identifying galaxies and stars in wide-field imaging. The project provides Jupyter notebooks for toy problems and RFI mitigation, making it accessible for both learning and practical applications. While the project is discontinued in favor of a TensorFlow 2 compatible version, it remains a valuable resource for understanding and implementing U-Net architectures.

SF Tensor, also known as The San Francisco Tensor Company, is dedicated to reinventing the software and infrastructure stack for modern AI and High-Performance Computing (HPC). The platform provides automatic kernel optimization and cross-cloud, cross-vendor compute capabilities, ensuring code runs faster, cheaper, and is portable across various platforms. It supports a heterogeneous future where CPUs, GPUs, TPUs, and domain-specific accelerators are treated as first-class citizens. SF Tensor offers two main options: Tensor Cloud for experiments and medium-scale training jobs, and Forward-Deployed for scaling training runs with dedicated infrastructure support. Pricing is aligned with the savings delivered to customers.

SciKit-Learn Laboratory (SKLL) is a Python package designed to make machine learning experimentation easier, particularly with scikit-learn. It offers command-line utilities like `run_experiment` to configure and execute a series of learners on datasets using a simple configuration file, eliminating the need for extensive custom code. SKLL supports various tasks such as cross-validation, evaluation, prediction, and training, with options for hyperparameter tuning via grid search. Beyond the command-line interface, it provides a Python API for those who prefer programmatic control, supporting pandas DataFrames. The tool also includes utilities for converting between machine learning toolkit formats, filtering, and joining feature files, making it a comprehensive solution for managing machine learning workflows.

DeepRL is an open-source platform dedicated to Deep Reinforcement Learning (DRL), aiming to make DRL technology accessible and enjoyable for everyone. It functions as a comprehensive lab, offering a wealth of resources including books, papers, courses, and open-source frameworks. The platform guides users through the fundamental principles, cutting-edge algorithms, and practical applications of DRL. It also covers industry trends, competition analysis, and paper sharing, fostering a community for DRL researchers and enthusiasts. DeepRL is designed to support learning and research in various DRL applications, from autonomous driving and robotics to natural language processing and financial systems.

skops is a Python library designed to facilitate the sharing and deployment of scikit-learn based models, enabling their integration into production environments. A key feature is `skops.io`, which provides secure persistence for scikit-learn estimators and other objects, offering a safer alternative to traditional pickle serialization. Additionally, `skops.card` allows users to create comprehensive model cards, explaining the model's purpose and usage. These model cards can be stored on the Hugging Face Hub, complete with pre-populated metadata for better understanding and discoverability. The library aims to streamline the process of managing and distributing machine learning models effectively, with detailed documentation and contributing guidelines available for users and developers.

Safeguard Applied Innovation LTD provides an advanced AI safety platform designed to prevent workplace accidents. The platform simplifies safety management, making it smart and user-friendly. It offers tools and insights for real-time risk identification and reduction, leveraging advanced sensing and monitoring technologies. Safeguard's unique approach breaks down safety management into individual tasks with control mechanisms, while also aggregating real-time data to create a comprehensive safety status through its 'Compass' metric. Key features include personnel competency control, subcontractor competency control, equipment competency control, and proactive safety management, ensuring a data-driven approach to loss prevention.

aiToolHub serves as a comprehensive platform dedicated to the discovery and adoption of artificial intelligence tools. It enables users to efficiently search, compare, and select from a wide array of AI solutions available in the market. The platform is designed to assist both businesses and individual professionals in quickly identifying the most suitable AI tools to meet their specific requirements, streamlining the process of integrating AI into their workflows. By providing a centralized hub for AI tool exploration, aiToolHub aims to simplify decision-making and accelerate the adoption of innovative AI technologies.

reasoning-gym is a Python library designed for training reasoning models using reinforcement learning. It offers a comprehensive set of dataset generators and reasoning environments, allowing users to create and manage training data with adjustable complexity. The tool provides access to over 100 distinct tasks, covering a wide range of reasoning challenges. This makes it a valuable resource for researchers and developers focused on advancing AI's reasoning capabilities, particularly those working with reinforcement learning approaches. While the provided content is from GitHub's pricing page, it indicates that the underlying project is likely open-source or free to use, given its presence on GitHub and the lack of specific pricing for the 'reasoning-gym' itself, suggesting it's a development framework rather than a commercial product.

Real3DPortrait is an open-source project providing a PyTorch implementation for one-shot realistic 3D talking portrait synthesis. It allows users to generate high-quality talking face videos from a single source image and a driving audio or video. The tool supports both audio-driven and video-driven methods for generating expressive 3D portraits. Key features include the ability to control mouth amplitude, map initial poses, and provide custom background images. It offers a command-line interface, a Gradio WebUI, and a Google Colab notebook for inference, making it accessible for various users. The project also provides training code for its audio-to-motion and image-to-plane models.

Recruitment Workflow is an open-source tool designed to automate and streamline various tasks involved in the hiring process. Leveraging the CrewAI framework, it orchestrates AI agents to build and execute real-world recruitment applications. This tool helps recruiters and HR professionals by automating initial candidate screening, managing candidate communication, and optimizing workflow efficiency. It is particularly useful for those looking to integrate AI into their talent acquisition strategies to reduce manual effort and improve the speed and quality of hiring. The open-source nature allows for customization and integration into existing HR systems.

PaddleSlim is an open-source library designed for deep model compression and architecture search, offering a comprehensive suite of strategies to optimize machine learning models. It supports techniques such as low-bit quantization, knowledge distillation, and various sparsity methods, enabling developers to significantly reduce the size and improve the efficiency of their models. Key features include automated compression, which allows direct loading of ONNX and Paddle models for tasks like offline quantization (PTQ), quantization-aware training (QAT), and sparse training. The library also provides tools for performance estimation on various ARM CPU devices and supports deployment with Paddle Inference and Paddle Lite. PaddleSlim is particularly useful for optimizing models for deployment on resource-constrained environments like mobile devices.

Python-Machine-Learning-Second-Edition is a comprehensive code repository accompanying the second edition of the book published by Packt. This resource is designed to support readers in their journey to learn and implement machine learning models. It includes all the necessary project files, allowing users to follow along with the book's examples and exercises. The content specifically focuses on practical applications of machine learning using popular libraries such as TensorFlow and scikit-learn, making it an invaluable asset for those looking to gain hands-on experience in the field. It serves as a practical companion for understanding and applying machine learning concepts.

IDWise is an AI-based identity verification solution designed to help businesses streamline customer onboarding, prevent fraud, and ensure compliance with e-KYC and AML regulations. The platform supports over 13,000 ID documents across 200+ countries and territories, with a strong focus on emerging markets. Key features include AI-based identity document recognition and validation, facial verification with liveness detection, and comprehensive AML screening against global watchlists. IDWise prides itself on its truly AI-based, in-house developed technology, offering up to 50 security checks per ID document in seconds. It provides a seamless integration experience through Mobile and Web SDKs, APIs, and low-code/no-code options, aiming to deliver a superior user experience and dramatically accelerate customer conversion.

v2ray-SSR-Clash-Verge-Shadowrocke is an open-source repository offering free, high-speed server nodes for popular protocols like v2ray, SS, sing-box, Clash, Verge, SSR, and Shadowrocket. This tool is designed to help users bypass internet restrictions and access geo-blocked content on platforms such as YouTube, Netflix, TikTok, ChatGPT, and bilibili. It provides comprehensive subscription guides for setting up these nodes across a wide range of devices, including Windows, Mac, Linux, iOS, Android, and routers. The repository also includes VPN reviews and is compatible with various client applications like Clash, V2ray, and sing-box, making it a versatile solution for scientific internet access.

pytorch-original-transformer offers a PyTorch implementation of the original transformer model by Vaswani et al., designed to facilitate learning and experimentation with transformers. The repository includes a `playground.py` file with visualizations for complex concepts like positional encodings and custom learning rate schedules, making them easier to grasp. It also provides pretrained models on the IWSLT dataset for English-German machine translation, demonstrating practical application. The tool supports training new models and inference, with well-commented code and setup instructions for a smooth user experience. It's an excellent resource for anyone looking to understand and work with the foundational transformer architecture.

PoseEstimation-CoreML is an open-source project designed for inferencing pose estimation on iOS devices utilizing Apple's Core ML framework. This tool allows developers to estimate body poses from still images and real-time video feeds captured by the device's camera. Key features include visualizing poses as heatmaps or lines and points, and the ability to capture and match poses. It supports various models like cpm and hourglass, providing performance metrics across different iPhone models. The project offers clear instructions for integrating models, handling camera permissions, and performing inferences using the Vision framework, making it a valuable resource for iOS machine learning development.

yolov3-channel-and-layer-pruning is an open-source project built upon ultralytics/yolov3, designed to optimize YOLOv3 and YOLOv4 object detection models. It leverages the principles of Network Slimming (ICCV 2017) by pruning channels based on BN layer Gamma coefficients, and also incorporates layer pruning. This approach significantly reduces model parameters and computational load, leading to faster inference times. The project offers various channel pruning strategies, including those that handle shortcut connections, and introduces layer pruning to compress model depth. Additionally, it integrates knowledge distillation strategies to help recover or even improve model accuracy after aggressive pruning. The tool supports sparse training, fine-tuning, and offers different sparsity strategies to balance compression and accuracy.

RewardBench is an open-source benchmark and evaluation tool specifically designed for assessing the capabilities and safety of reward models, including those utilizing Direct Preference Optimization (DPO). The repository offers common inference code compatible with various reward models such as Starling, PairRM, OpenAssistant, and DPO. It ensures fair evaluation through standardized dataset formatting and testing procedures. Additionally, RewardBench includes robust analysis and visualization tools to help researchers and developers interpret results effectively. It supports quick evaluation of any reward model on any preference set, with features for logging model outputs and accuracy scores, and options for generative models (LLM-as-judge) and DPO models. The platform also facilitates contributing models to a public leaderboard and offers offline ensemble testing.

Phala Cloud offers a hardware-secured compute platform designed for confidential AI, ensuring verifiable AI with enterprise-grade privacy. It allows users to deploy confidential AI models with Trusted Execution Environment (TEE) protection quickly. The platform supports various pre-configured confidential AI models from providers like MoonshotAI, Qwen, and DeepSeek, ready for deployment on hardware-secured GPU servers. Phala Cloud provides an all-in-one confidential compute platform for AI workloads, offering nearly native performance with 100% privacy. It is built for enterprise security and regulatory requirements, being SOC 2 Type I certified and HIPAA compliant, with ISO 27001 in progress. The platform supports popular AI frameworks like TensorFlow, PyTorch, and Hugging Face, and offers per-minute billing with no minimums or hidden fees.

SGX-Full-OrderBook-Tick-Data-Trading-Strategy is an open-source project designed for developing and implementing high-frequency trading (HFT) strategies. It leverages data science and machine learning techniques to analyze full order book tick data, providing insights into market microstructure. The framework is built to capture the intricate dynamics of high-frequency limit order books, which is crucial for HFT. Key features include methods for feature extraction, such as Rise Ratio and Depth Ratio, enabling users to derive meaningful signals from raw tick data. This project is ideal for quantitative researchers and traders looking to backtest and deploy sophisticated trading algorithms.

RL-Factory is an open-source framework designed for efficient reinforcement learning (RL) post-training in Agentic Learning. It significantly simplifies the process by decoupling the environment from RL post-training, allowing users to train agents with only a tool configuration and a reward function. A key differentiator is its support for asynchronous tool-calling, which makes RL post-training up to 2x faster than existing frameworks. The platform natively supports one-click DeepSearch training, multi-turn tool-calling, model judge reward mechanisms, and training for various models, including Qwen3. Future updates aim to introduce a WebUI for data processing, environment definition, and project management, alongside support for more models and multimodal agentic learning.

schnetpack is an open-source toolbox designed for researchers and developers working with atomistic systems. It provides a robust framework for developing and applying deep neural networks to predict various properties of molecules and materials, such as potential energy surfaces and quantum-chemical characteristics. The tool includes fundamental building blocks for atomistic neural networks, simplifying the process of conducting simulations and making accurate property predictions. Its open-source nature, hosted on GitHub, encourages community contributions and provides transparent access to its codebase, making it a valuable resource for academic and industrial research in computational chemistry and materials science.