The Computational Network Toolkit (CNTK) by Microsoft Research, is a unified deep-learning toolkit that describes neural networks as a series of computational steps via a directed graph.
General purpose gradient boosting on decision trees library with categorical features support out of the box. It is easy to install, contains fast inference implementation and supports CPU and GPU (even multi-GPU) computation.
A deep learning framework developed with cleanliness, readability, and speed in mind. [DEEP LEARNING]
A simple Multi-armed Bandit library. [Deprecated]
A real-time multi-person keypoint detection library for body, face, hands, and foot estimation
VIGRA is a genertic cross-platform C++ computer vision and machine learning library for volumes of arbitrary dimensionality with Python bindings.
A C++ edge video AI engine for RTSP ingestion, CV/VLM inference, visual pipeline orchestration, alarms, and event delivery on Sophon and Rockchip NPUs.
On-device streaming speech recognition runtime with C API for Linux (aarch64/x8664). Based on NVIDIA NeMo FastConformer (80 ms cache-aware lookahead). Companion Silero VAD, wake-word, and 14-command KWS via same runtime.
Ultralytics' YOLOv8 implementation with C++ support for real-time object detection and tracking, optimized for edge devices.
C-based/Cached/Core Computer Vision Library, A Modern Computer Vision Library.
A native C CPU inference engine for running the 284B-A13B DeepSeek-V4-Flash-0731 on one laptop, reaching up to 1.12 token/s on tested hardware while streaming the 167 GB checkpoint from disk with a tested 8 GB memory path and no GPU or Python.
A native C inference engine for running Qwen3.8-27B locally on a single laptop CPU, with direct GGUF loading and a tested 8 GB memory path.
Neural networks framework in pure C: training and inference, no dependencies.
A lightweight C library for ONNX model inference, optimized for performance and portability across platforms.
A lightweight, portable pure C99 onnx inference engine for embedded devices with hardware acceleration support.
An ONNX runtime written in pure C (99) with zero dependencies focused on small embedded devices. Run inference on your machine learning models no matter which framework you train it with. Easy to install and compiles everywhere, even in very old devices.
neonrvm is an open source machine learning library based on RVM technique. It's written in C programming language and comes with Python programming language bindings.
A hybrid recommender system based upon scikit-learn algorithms. [Deprecated]
A C library for product recommendations/suggestions using collaborative filtering (CF).
Darknet is an open source neural network framework written in C and CUDA. It is fast, easy to install, and supports CPU and GPU computation.
Naive Bayesian Classifier implementation in APL. [Deprecated]
MCP server for accessing and managing Kibana dashboards, visualizations, and index patterns through natural language.
🚀 ☁️ - An MCP server application that sends various types of messages to the WeCom group robot.
🎖️ ☁️ - Trade stocks and crypto on common brokerages (Robinhood, ETrade, Coinbase, Kraken) via Trade Agent's MCP server.
The Fastest Gower Distance Implementation for Python. GPU-accelerated similarity matching for mixed data types, 15-25% faster than alternatives with production-ready reliability.
Google Antigravity — IDE нового поколения, где ИИ-агенты автономно планируют, пишут, тестируют и верифицируют код по вашим задачам. Бесплатная превью-версия.
Self-evolving Context Database for AI Agents — unifies Agent Memory, Knowledge RAG, and Skills behind a single interface, with a portable Agent Plugins package format (agent-plugins.org). !GitHub Repo stars
AI-инструмент, который помогает искать работу и адаптировать резюме под вакансии @NeuroRazvedka AI Job Search - пр
AI-агенты для автоматического создания видео @NeuroRazvedka Компания HeyGen открыла исходный код HyperFrames — плат
Магазин Android-приложений на релизах GitHub: находит репозитории с APK, ставит и обновляет их
Рабочий центр для команды AI-агентов: нанимает, ставит задачи и отчитывается о работе
Скил для агента и локальный сервис: убирают следы AI-происхождения из своих текстов и файлов
An open-source terminal agent with sandbox and approval controls.
Ultra-lightweight personal AI assistant framework (~4,000 lines of Python). Supports MCP, 9+ chat channels, and extensible skills system. !GitHub Repo stars
EvoAgentX is building a Self-Evolving Ecosystem of AI Agents, it will give you automated framework for evaluating and evolving agentic workflows. !GitHub Repo stars
Practical principles for building controllable LLM applications around deterministic software.
AI conversations that remember context ☆3,783
Fully open reproduction of DeepSeek-R1
a lean, efficient, and easy-to-hack codebase to research LLMs.
LLM inference in C/C++.
Playwright wrapper for a stealth-patched Firefox 150 binary. Drop-in replacement returning native Playwright Browser objects, fingerprint spoofing in the C++ source. !GitHub Repo stars
Clean, minimal, accessible reproduction of DeepSeek R1-Zero
Cursor Free VIP — сбросить Machine ID в Cursor. Инструмент снимает все лимиты пробного периода и открывает полноценный
Helios: открытая 14B-модель генерирует видео длиной больше 60 секунд со скоростью 19,5 FPS на одной H100 Исследователи из Пекинского университета и ByteDance выпустили Helios — открытую 14B-модель, которая генерирует видео в реальном времени - 19,5 кадров в секунду, длительностью больше минуты.
История одного разочарования.
libraries Еще одна интересная библиотека. В этот раз со всякими интерпретируемыми ML моделями. imodels - sklearn-st
Четвёртого дня зарелизился проект d7vk , в котором реализован Direct3D 7, работающий поверх Vulkan.
Приняли участие в премии Сбера в области R&D для исследователей Сбере прошла церемония награждения лучших команд и со