AI Agents & Automation

Cold-Diffusion-Models offers the official PyTorch implementation of Cold-Diffusion, a novel approach for inverting arbitrary image transformations without the need for traditional noise. Developed by researchers at the University of Maryland, this repository provides comprehensive code to train and test cold diffusion models. It supports a range of image degradations, including Gaussian blur, animorphosis, Gaussian mask, resolution downsampling, image snow, and color desaturation. The implementation is based on lucidrains' denoising diffusion repository and includes pretrained models for CelebA and AFHQ generation. Users can explore both conditional and unconditional generation schedules, with detailed scripts and arguments for training and testing different models and degradation types.

CML (Continuous Machine Learning) is an open-source command-line interface (CLI) tool designed for continuous integration and continuous delivery (CI/CD) within Machine Learning Operations (MLOps). It automates various development workflows, such as machine provisioning, model training, and evaluation. CML enables users to compare ML experiments across project history and monitor changing datasets. It can automatically train and evaluate models, then generate visual reports with results and metrics on every pull request. CML supports GitFlow for data science, allowing management of ML experiments and tracking of model training or data modifications using GitLab or GitHub. It integrates with DVC for codifying data and models and offers functions to package ML workflow outputs into markdown reports for CI systems.

csghub-server is the open-source backend server for CSGHub, a platform designed for managing large model assets. It facilitates the management of models, datasets, and other LLM assets through a robust REST API. Key features include the creation and management of users and organizations, automatic tagging of models and datasets, and comprehensive search functionalities. Users can also preview dataset files online, download individual files including LFS files, and track activity data like downloads and likes. The server supports extensible and customizable architectures, allowing integration with various Git servers and flexible configuration of LFS storage systems. It also enables on-demand content moderation and has a roadmap for supporting more Git servers, Git LFS, dataset online viewers, and model/dataset auto-tagging.

Workmate is an AI scheduling assistant designed to streamline meeting coordination and calendar management. It acts as a proactive assistant, handling scheduling conversations, sharing availability, and sending invites. The tool eliminates back-and-forth emails and ensures no dropped balls with automated follow-ups. Workmate also proactively messages users about conflicts and offers instant scheduling with colleagues. Users can customize their assistant with a human name or email at their domain, and provide free-text instructions for personalized calendar management. It integrates with existing tools and offers free, unlimited booking links for easy meetings, making it suitable for individuals and entire teams.

AI FARM ROBOTICS is a pioneering company dedicated to advancing Cambodia's technological landscape by focusing on robotics and AI. The company specializes in the research and development of core robotic technologies and products, aiming to establish Cambodia as a leader in the robotic industry. Beyond R&D, AI FARM ROBOTICS provides comprehensive system integration and management services tailored for Micro, Small, and Medium Enterprises (MSMEs), facilitating their automation processes. They also offer advanced AI solutions specifically designed for robotics applications and provide Robotics-as-a-Service (RaaS) offerings, making sophisticated robotic capabilities accessible to a wider range of businesses.

Dataherald is an open-source natural language-to-SQL engine designed for enterprise-level question answering over relational data. It enables users to set up an API from their database, allowing them to answer questions in plain English without needing to write SQL. This tool is ideal for business users who need to gain insights from data warehouses without relying on data analysts, for integrating Q+A capabilities from production databases into SaaS applications, and for creating ChatGPT plugins from proprietary data. The platform includes an engine for core NL-to-SQL functionality, an enterprise layer for authentication and business logic, an admin console for GUI-based configuration, and a Slackbot for interactive querying.

DeepBrain Chain is positioned as the world's first public artificial intelligence chain, aiming to create a decentralized AI infrastructure. The platform leverages blockchain technology to address the computational demands of AI by utilizing idle computing resources globally. This approach is designed to offer a more cost-effective solution for AI development while simultaneously enhancing data privacy through the implementation of smart contracts. By decentralizing AI computing, DeepBrain Chain seeks to provide a robust and secure environment for developers and organizations working on AI projects, ensuring both efficiency and data protection.

Optible AI offers an advanced AI-powered platform designed to transform grant management for government departments and foundations. It automates workflows, significantly reducing review times by up to 90% through AI-driven assessment and allocation. The platform ensures fair, accurate, and consistent decisions at scale by screening applications faster and providing highly accurate eligibility screening. Key features include automated setup, real-time document validation to detect fraud, and AI-driven screening that processes thousands of applications in minutes. Optible AI also delivers 300x more data insights through detailed, customizable reports, enabling organizations to track progress, refine policies, and maximize their impact efficiently.

DeepGBM is a deep learning framework specifically designed for online prediction tasks, leveraging the power of Gradient Boosting Decision Trees (GBDT) for distillation. Presented at KDD'2019, this framework aims to significantly improve prediction accuracy in real-time scenarios. It integrates GBDT-based models, specifically LightGBM, with PyTorch-based neural networks. The project includes comprehensive code for data preprocessing, baseline model implementations, and the proposed DeepGBM model. Users can prepare their data in CSV format, process it through encoders, and then load numerical and categorical data for training. The framework supports training GBDT2NN or the full DeepGBM model, offering flexibility for different prediction needs.

Totoy specializes in integrating state-of-the-art AI solutions into existing business processes, focusing on measurable profitability and employee satisfaction. They offer a comprehensive approach starting with a free AI workshop, followed by an in-depth potential analysis where specialists spend a day on-site. The process culminates in AI evaluation and implementation, delivering systems that save time and money. Totoy's solutions are developed and hosted in the EU, ensuring compliance with GDPR and AI Act regulations. They address various use cases including document management, customer support, administration, controlling, quality control, and knowledge management, providing tailored AI agents and systems.

DIAMOND (DIffusion As a Model Of eNvironment Dreams) is a cutting-edge reinforcement learning agent that operates entirely within a diffusion world model. This innovative approach allows the agent to learn and play in an autoregressive imagination of environments like Atari and Counter-Strike: Global Offensive. Recognized as a NeurIPS 2024 Spotlight, DIAMOND offers researchers and developers a powerful tool for exploring diffusion-based world models in reinforcement learning. It provides quick installation for pretrained models, options for launching training runs, and extensive configuration management via Hydra. Users can visualize policy play, human interaction, and dataset replays, making it a versatile platform for advanced AI research.

Dictionariez is a highly customizable, open-source browser extension designed to significantly enhance language learning. It allows users to double-click on any word on a webpage to instantly access its definition, translation, and pronunciation. Supporting over 20 languages, including English, Chinese, Japanese, Korean, German, Spanish, French, Italian, Portuguese, and Swedish, it integrates with more than 1000 dictionaries. Beyond quick lookups, Dictionariez provides text-to-speech functionality, translation services, and seamless integration with Anki for vocabulary retention. The extension also features auto-completion, word history, and keyboard shortcuts, making it a comprehensive tool for students and language enthusiasts alike. It is available on Chrome, Firefox, and Edge, with specialized versions like Ordböcker for Swedish learners and SidePal for a lighter side-panel experience.

Venta AI is an AI sales employee designed for sales teams in Europe, specializing in highly targeted and GDPR-compliant cold outreach. The platform identifies qualified leads from over 70 million companies based on individual criteria, moving beyond traditional job titles to find ideal buying personas. It automates extensive company data research, integrates with CRMs like HubSpot, Salesforce, and Pipedrive, and crafts AI-personalized messages for various channels. Venta AI ensures optimized email deliverability, LinkedIn automation, and even offers physical letter campaigns. It also monitors sales signals such as job postings, website visits, and news articles to identify opportune outreach moments, all while adhering strictly to EU data protection laws.

edm is the official PyTorch implementation of the NeurIPS 2022 paper "Elucidating the Design Space of Diffusion-Based Generative Models." This open-source tool provides a clear framework for understanding and experimenting with diffusion models, separating concrete design choices in sampling and training processes, as well as score network preconditioning. It introduces improvements that lead to state-of-the-art FID scores for CIFAR-10, FFHQ, AFHQv2, and ImageNet, with significantly faster sampling times. The project includes pre-trained models, tools for generating images, calculating Fréchet Inception Distance (FID), and preparing custom datasets. It supports both Linux and Windows, recommending Linux for performance, and requires high-end NVIDIA GPUs for optimal use.

DREAMPlace is an open-source, deep learning toolkit-enabled VLSI placement tool designed for flexibility and efficiency in very large-scale integration (VLSI) design. It supports both CPU and GPU execution, achieving over 30X speedup in global placement and legalization compared to CPU implementations like RePlAce on ISPD 2005 benchmarks with a Nvidia Tesla V100 GPU. The tool integrates ABCDPlace, a GPU-accelerated detailed placer, which provides around 16X speedup on million-size benchmarks over NTUPlace3. Key features include multi-threaded CPU and optional GPU acceleration, net weighting, incremental placement, LEF/DEF support, and Python binding. It also supports timing optimization in global placement, fence regions, and deterministic modes.

gemma.cpp is a lightweight, standalone C++ inference engine specifically designed for Google's Gemma foundation models. It provides a minimalist implementation for Gemma 2-3 and PaliGemma 2 models, prioritizing simplicity and directness over full generality, making it suitable for experimentation and research. The engine supports CPU-only inference, offering features like sampling with TopK and temperature, and a backward pass (VJP) with Adam optimizer for Gemma research. It includes optimizations such as mixed-precision GEMM (fp8, bf16, fp32, fp64 bit), automatic runtime autotuning, and integrated weight compression. The project leverages the Google Highway Library for portable SIMD, ensuring efficient CPU inference. It offers C++ APIs with streaming for single and batched inference, a basic interactive command-line app, and Python bindings. gemma.cpp is designed to be easily embeddable in other projects with minimal dependencies and is highly modifiable, featuring a small core implementation.

Frigate-hass-integration is an open-source project that seamlessly integrates Frigate, an AI-powered Network Video Recorder (NVR), with Home Assistant. This integration enhances smart home surveillance by providing a rich media browser with thumbnails and navigation directly within Home Assistant. Users gain access to various sensor entities, including Camera FPS, Detection FPS, Process FPS, Skipped FPS, and Objects detected, along with binary sensor entities for object motion. It also offers camera entities for live view and object detected snapshots, and switch entities for controlling recording, detection, snapshots, and contrast improvement. Furthermore, the integration provides services for manual events and PTZ control, supporting multiple Frigate instances for comprehensive home security management.

Fregata is a lightweight, super-fast, and large-scale machine learning library designed for Apache Spark. It offers high-level APIs in Scala, enabling developers and data scientists to build and deploy intelligent applications efficiently. A key differentiator is its ability to achieve higher accuracy and significantly faster convergence compared to MLLib, often training Generalized Linear Models in minutes for massive datasets. Fregata utilizes GSA SGD optimization, making it parameter-free by dynamically calculating appropriate learning rates. It supports Spark 1.x and 2.x with Scala 2.10 and 2.11, and includes algorithms like Trillion LR, Trillion SoftMax, and Logistic Regression. Its architecture is designed for seamless integration into existing Spark data processing workflows.

Airia is an enterprise AI platform designed to simplify and secure the deployment and management of AI solutions at scale. It offers a unified solution for orchestration, security, and governance, enabling organizations to build, deploy, and manage AI agents with confidence. Key capabilities include AI discovery, agent constraints, routing engine, and security posture management. The platform also provides robust governance features such as AI inventory management, risk classifications, and compliance reporting, ensuring responsible AI use and regulatory alignment. Airia integrates with thousands of enterprise systems and data sources, allowing AI agents to operate with real-time context and accuracy, making it ideal for large organizations seeking to expand AI adoption without chaos or tool sprawl.

HolmesGPT is an open-source AI agent designed to investigate production incidents and pinpoint root causes across diverse infrastructure stacks, including Kubernetes, VMs, and cloud providers. As a CNCF Sandbox project, it offers robust features like petabyte-scale data handling with server-side filtering and memory-safe execution to prevent OOM kills during large data queries. It boasts deep integrations with popular observability tools such as Prometheus, Grafana, Datadog, and Kubernetes, alongside bidirectional alert integrations with platforms like AlertManager, PagerDuty, and Jira. A key differentiator is its 'Operator Mode,' which allows HolmesGPT to run continuously, detect issues before they impact customers, and even open PRs to fix identified problems, making it a proactive SRE solution.

Holodeck is a high-fidelity simulator designed for reinforcement learning and robotics research, leveraging the power of Unreal Engine 4. It offers a robust platform with over seven rich worlds and numerous scenarios for training AI agents. The simulator supports both Linux and Windows operating systems, allowing for easy extension and modification of training scenarios. A key feature is its ability to train and control multiple agents simultaneously, providing a flexible environment for complex research. It boasts a simple, OpenAI Gym-like Python interface for ease of use and high performance, capable of simulation speeds up to 2x real-time. Holodeck can run headless or with visual feedback, catering to different research needs.

GraphWaveletNeuralNetwork is an open-source PyTorch implementation of the "Graph Wavelet Neural Network" (GWNN) as presented at ICLR 2019. This novel graph convolutional neural network addresses limitations of previous spectral graph CNN methods by utilizing graph wavelet transform, which avoids computationally expensive matrix eigendecomposition. The graph wavelets are sparse and localized, enhancing efficiency and interpretability for graph convolution tasks. The tool is designed for researchers and machine learning engineers working with graph-based semi-supervised classification, demonstrating superior performance on benchmark datasets like Cora, Citeseer, and Pubmed. It includes command-line arguments for easy configuration of training parameters and model options.

Keras-Project-Template is an open-source project template designed to streamline the development and training of deep learning models with Keras. It offers a clear, structured architecture, including predefined folders for models, trainers, data loaders, and configurations, simplifying project organization. The template supports checkpointing and TensorBoard visualization for monitoring training progress. A key feature is its integration with Comet.ml, enabling comprehensive experiment tracking, including hyper-parameters, metrics, and graphs, with real-time updates. This allows developers to easily manage and compare different model iterations and configurations, enhancing the efficiency of deep learning research and development.

Kedro is an open-source Python framework designed for building production-ready data engineering and data science pipelines. It emphasizes software engineering best practices to ensure pipelines are reproducible, maintainable, and modular. Key features include a project template based on Cookiecutter Data Science, a Data Catalog for connecting to various data sources and versioning, and pipeline abstraction for automatic dependency resolution and visualization with Kedro-Viz. Kedro also supports coding standards like test-driven development with pytest and flexible deployment strategies, including integration with Argo, Prefect, Kubeflow, AWS Batch, and Databricks. It aims to address the shortcomings of one-off scripts and Jupyter notebooks by promoting team collaboration and efficiency through modular, reusable analytics code.