AI Trending

Project directory

Projects

Open-source AI projects tracked by the leaderboard, ordered by GitHub stars.

Go 389 stars

crillab/gophersat

gophersat, a SAT solver in Go

Overview: Gophersat is a high-performance SAT and pseudo-boolean solver written in Go, developed by the CRIL (Centre de Recherche en Informatique de Lens) at the Artois University & CNRS. It is released under the MIT license and is known for its efficiency, especially on typical SAT benchmarks. Key features: - Cutting-planes solving strategy - Support for MAXSAT problems and pseudo-boolean decision and optimization problems Tech stack: Go Why it matters: Gophersat is a powerful tool for solving complex logical and optimization problems, making it highly relevant in fields such as artificial intelligence, computer science, and engineering. Its efficient performance and ability to handle large-scale problems make it a valuable asset in various applications.
C++ 389 stars

zilliztech/knowhere

Vector search engine inside Milvus, integrating FAISS, HNSW, DiskANN.

### Overview: Knowhere is a vector search engine developed as part of Milvus, integrating FAISS, HNSW, and DiskANN. It is written in C++ and serves as the core of Milvus. ### Key Features: - Integrates FAISS, HNSW, and DiskANN for efficient vector search. - Built as part of Milvus, it is designed to be used within the Milvus ecosystem. - Supports both CPU and GPU builds, with GPU support requiring additional configuration. ### Tech Stack: C++ ### Why It Matters: Knowhere is crucial for Milvus, enabling efficient vector search capabilities that are essential for Milvus's core functionalities.
Python 388 stars

BolisettySujith/J.A.R.V.I.S

A voice assistant 🗣️ which can be used to interact with your computer 💻 and controls your pc operations 🎛️

Overview: JARVIS is a versatile voice assistant designed to interact with your computer and perform various tasks, including tracking Covid-19 cases, screen recording, voice and video recording, accessing mobile and web cameras, and much more. It's a practical tool that can be used in various scenarios, including Iron Man movies. Key features: - Tracks Covid-19 cases for each state in India - Screen recording and voice recording - Voice recording - Access to mobile camera - Access to web camera - Finds phone number location - Reads PDFs - Works as a telephone dictionary - Generates QR codes - Checks Internet speed - Provides IP address - Latest news - System condition check - Sends emails - Sends WhatsApp messages - Plays YouTube songs - Downloads YouTube songs - Downloads Instagram profiles - Finds current location - Takes screenshots - Displays current time - Displays current day - Tells programming jokes - Shows schedule for each day - Silences for a specified duration - Searches Wikipedia and provides brief information - Searches web for specific information - Controls system volumes - Controls system power activities - Plays music file from a specific directory Tech stack: Python Why it matters: JARVIS is a practical tool that can be used in various scenarios, including Iron Man movies, making it a useful addition to any developer's toolkit.
Python 388 stars

KittenCN/stock_prediction

基于神经网络的通用股票预测模型 A general stock prediction model based on neural networks

Overview: 这个项目是一个基于神经网络的通用股票预测模型实验项目，使用 PyTorch 实现，包含行情获取、特征处理、训练、推理和模型评估等功能。 Key features: - LSTM - Transformer - Hybrid - PTFT/VSSM - Diffusion - Graph Tech stack: Python Why it matters: 该项目为股票预测提供了一个通用的模型框架，通过神经网络技术进行时间序列预测，旨在为投资者提供有价值的投资建议。 Repository: KittenCN/stock_prediction GitHub stars: 388 GitHub description: 基于 PyTorch 的股票时间序列预测实验项目，包含行情获取、特征处理、训练、推理和模型评估。
Python 388 stars

jayparks/tf-seq2seq

Sequence to sequence learning using TensorFlow.

### Overview: This project is a comprehensive implementation of sequence-to-sequence learning using TensorFlow, focusing on the Encoder-Decoder architecture with Attention mechanism. The package supports various advanced features such as multi-layer GRU/LSTM, residual connections, dropout, attention, and beam search decoding. It is designed for efficient training and decoding of sequence data, making it suitable for tasks like machine translation. ### Key Features: - **RNN Encoder-Decoder Architecture**: Utilizes recurrent neural networks for both encoding and decoding. - **Attention Mechanism**: Implements attention to improve the model's ability to focus on relevant parts of the input sequence. - **Advanced Features**: - **Multi-layer GRU/LSTM**: Supports multiple layers of GRU/LSTM for better sequence modeling. - **Residual Connections**: Enhances the model's ability to learn long-range dependencies. - **Dropout**: Reduces overfitting by randomly dropping units during training. - **Attention and Input Feeding**: Improves the model's ability to attend to relevant parts of the input sequence. - **Beam Search Decoding**: Allows for beam search decoding, which is useful for tasks like translation where multiple hypotheses are considered. ### Tech Stack: - **Python**: The primary programming language used for implementing the project. - **TensorFlow**: The deep learning framework used for building and training the model. ### Why It Matters: This project is essential for researchers and developers working on sequence-to-sequence learning tasks, particularly in the field of machine translation. The implementation includes advanced features that significantly improve the model's performance
Python 388 stars

neuralmagic/sparsezoo

Neural network model repository for highly sparse and sparse-quantized models with matching sparsification recipes

Overview: SparseZoo is a repository for neural network models that are highly sparse and sparse-quantized, featuring matching sparsification recipes. Key features: - Highly sparse and sparse-quantized models - Matching sparsification recipes Tech stack: Python Why it matters: SparseZoo provides a community-driven repository for developers to access and experiment with highly efficient neural network models, making it easier to implement sparse and quantized techniques in their projects.
Python 388 stars

polm/cutlet

Japanese to romaji converter in Python

Overview: Cutlet is a Japanese to romaji converter tool written in Python, designed to be simple and efficient for quick conversions. Key features: - Support for Modified Hepburn, Kunrei-shiki, and Nihon-shiki systems - Custom overrides for individual mappings and specific words - Built-in exceptions list for Tokyo, Osaka, etc. - Uses foreign spelling when available in UniDic - Proper nouns are capitalized - Slug mode for URL generation Tech stack: Python Why it matters: Cutlet is useful for quickly converting Japanese text to Romanized versions, making it ideal for translations, casual communication, or any situation where you need to simplify Japanese text for easier reading or sharing. Repository: polm/cutlet Primary language: Python GitHub stars: 388 GitHub description: Japanese to romaji converter in Python README excerpt: [![Open in Streamlit](https://static.streamlit.io/badges/streamlit_badge_black_white.svg)](https://polm-cutlet-demo-demo-0tur8v.streamlit.app/) [![Current PyPI packages](https://badge.fury.io/py/cutlet.svg)](https://pypi.org/project/cutlet/)
Kotlin 387 stars

GetStream/gemini-android

✨ Gemini Android demonstrates Google's Generative AI on Android with Stream Chat SDK for Compose.

### Overview: Gemini Android showcases Google's Generative AI capabilities on Android using the Stream Chat SDK for Compose. This project demonstrates how to integrate Generative AI with Android applications, implementing Jetpack Compose for the UI, and integrating Stream Chat for real-time messaging and event handling. ### Key Features: - Demonstrates Gemini API integration for Android. - Implements entire UI elements with Jetpack Compose. - Uses Android architecture components with Hilt and AppStartup. - Performs background tasks using Kotlin Coroutines. - Integrates Stream Chat Compose SDK for real-time chat functionality. ### Tech Stack: Kotlin ### Why It Matters: This project provides a practical example of how to leverage Generative AI and Stream Chat SDKs to build powerful and real-time chat applications on Android, showcasing the integration of AI with traditional Android development practices.
Jupyter Notebook 387 stars

guyernest/advanced-rag

Jupyter Notebooks for Mastering LLM with Advanced RAG Course

### Overview: This repository provides a comprehensive set of Jupyter Notebooks designed to deepen understanding and practice in the advanced realm of Large Language Models (LLM) and Retrieval Augmented Generation (RAG). It serves as an essential resource for developers and researchers aiming to master the intricacies of RAG systems, particularly in enterprise environments. ### Key Features: - **Simple RAG (01_simple_rag.ipynb)**: Introduces the fundamental concepts and implementation of Retrieval Augmented Generation (RAG). - **Embedding Model (02_embedding_model.ipynb)**: Examines the role of embedding models in RAG systems, including their applications and limitations. - **Semantic Chunking (03_semantic_chunking.ipynb)**: Investigates the process of semantic chunking in RAG systems, focusing on its significance in information retrieval. - **Contextual Retrieval (04_contextual_retrieval.ipynb)**: Delves into the strategies for contextual retrieval in RAG systems, emphasizing the effective handling of numerical data and tables. - **Reverse Hyde (05_reverse_hyde.ipynb)**: Explores the Reverse Hyde technique in RAG systems, highlighting its importance in addressing contextual relevance and query ambiguity. - **Hybrid Search (06_hybrid_search.ipynb)**: Discusses the integration of hybrid search capabilities in RAG systems, encompassing both temporal relevance and multi-lingual support. - **Reranking (07_reranking.ipynb)**: Focuses on the
Python 387 stars

hailanyi/VirConv

Virtual Sparse Convolution for Multimodal 3D Object Detection

Overview: This repository contains the official code for the paper "Virtual Sparse Convolution for Multimodal 3D Object Detection" by [Hailan Yu](https://github.com/hailanyi). The code is based on OpenPCDet and includes three models: VirConv-L, VirConv-T, and VirConv-S, designed for 3D object detection tasks. Key features: - VirConv-L: A light-weight multimodal 3D detector based on Virtual Sparse Convolution. - VirConv-T: An improved multimodal 3D detector using Virtual Sparse Convolution and transformed refinement scheme. - VirConv-S: A semi-supervised version of VirConv-T, trained on both labeled and unlabeled data. Tech stack: Python Why it matters: This repository provides comprehensive code for the paper, which is essential for researchers and developers working on 3D object detection tasks. The models are designed to be efficient and effective, making them suitable for various applications in computer vision and robotics. GitHub stars: 385
Python 387 stars

istupakov/onnx-asr

A lightweight Python package for Automatic Speech Recognition using ONNX models

### Overview: onnx-asr is a lightweight Python package designed for Automatic Speech Recognition (ASR) using ONNX models. It simplifies the process by providing a pure Python interface with minimal dependencies, making it suitable for both developers and researchers. ### Key Features: - **Model Support**: Supports a wide range of modern ASR models from various providers. - **Device Compatibility**: Runs efficiently on a variety of devices, including small IoT/edge devices to powerful GPUs. - **Cross-Platform**: Works seamlessly across Windows, Linux, and macOS on x86 and Arm CPUs. - **Dependencies**: Minimal requirements, no need for PyTorch, Transformers, or FFmpeg. - **Versatile Models**: Supports models from Hugging Face and local directories, including quantized versions. - **File Handling**: Accepts WAV files or NumPy arrays, with built-in file reading and resampling. - **Custom Models**: Allows loading custom models with detailed documentation. - **Batch Processing**: Capable of batch processing for improved efficiency. - **Voice Activity Detection (VAD)**: Supports VAD for long-form recognition. - **Timestamps and Probabilities**: Returns token-level timestamps and log probabilities. - **CLI Interface**: Provides a simple command-line interface for easy usage. ### Tech Stack: - **Python**: The primary language used for development. - **ONNX**: Used for model conversion and inference. ### Why It Matters: onnx-asr simplifies the ASR process by providing a lightweight, efficient, and easy-to-use solution. It supports a wide range of models
Python 387 stars

kyegomez/ScreenAI

Implementation of the ScreenAI model from the paper: "A Vision-Language Model for UI and Infographics Understanding"

Overview: This project implements the ScreenAI model, a vision-language model designed for understanding user interfaces and infographics. It follows the structure described in the paper "A Vision-Language Model for UI and Infographics Understanding". Key features: - Implementation of the ScreenAI model - Use of PyTorch for the technical stack - The model processes image and text inputs to generate outputs Tech stack: PyTorch Why it matters: This project provides a practical implementation of the ScreenAI model, which can be used for various applications related to UI and infographics understanding, potentially aiding in the automation of design tasks and enhancing user experience analysis.
Java 387 stars

labsai/EDDI

Config-driven engine that turns JSON into production-grade AI agents. Multi-agent orchestration, 19+ LLM providers, MCP/A2A protocols, RAG, persistent memory, and enterprise compliance (EU AI Act, GDPR, HIPAA). Built on Quarkus.

Overview: EDDI is a production-grade, config-driven multi-agent orchestration middleware designed for conversational AI. It enables intelligent routing, persistent memory, and API orchestration without requiring code. Key features: - Config-driven engine for JSON to AI agent conversion - Multi-agent orchestration - 19+ LLM providers - MCP/A2A protocols - RAG support - Persistent memory - Enterprise compliance (EU AI Act, GDPR, HIPAA) - Built on Java 25 and Quarkus Tech stack: Java 25, Quarkus Why it matters: EDDI simplifies the development and deployment of conversational AI solutions, making it easier to integrate with various LLM providers and protocols, while ensuring compliance with industry standards and regulations. Repository: labsai/EDDI Primary language: Java GitHub stars: 387 GitHub description: Config-driven engine that turns JSON into production-grade AI agents. Multi-agent orchestration, 19+ LLM providers, MCP/A2A protocols, RAG, persistent memory, and enterprise compliance (EU AI Act, GDPR, HIPAA). Built on Quarkus.
Unknown 387 stars

lfai/lfai-landscape

🌄 Open Source AI & Data Landscape - provides overview of top tier projects in the open source AI and Data ecosystem, shows projects through GitHub data, funding or market cap, first and last commits, contributor count and much other information.

Overview: LF AI & Data Landscape is an open-source project that provides a visual map of top-tier projects in the open source AI and data ecosystem. It includes information such as GitHub stars, funding, market cap, first and last commits, contributor count, and more. Key features: - Visual representation of open source projects in the AI and data domains - Data from GitHub, Crunchbase, and other sources - Includes LF AI & Data Foundation members and projects - Supports local build and installation using the landscape2 tool Tech stack: JavaScript, Node.js, and various open-source libraries for data visualization and image generation Why it matters: LF AI & Data Landscape serves as a valuable resource for developers and researchers looking to explore and understand the landscape of open source projects in AI and data. It provides comprehensive information on projects, funding, contributors, and more, making it an essential tool for anyone working in the field.
Python 386 stars

AnkitNayak-dev/EpsteinFiles-RAG

A RAG pipeline implementation built on the 'Epstein Files 20K' dataset from Hugging Face (Teyler).

Overview: This project implements a Rich Answers Generator (RAG) pipeline using the 'Epstein Files 20K' dataset from Hugging Face (Teyler). It includes features such as no hallucinations, intelligent retrieval, fast processing, semantic understanding, REST API, interactive UI, scalability, and production-readiness. Key features: - No Hallucinations - Intelligent Retrieval - Fast Processing - Semantic Understanding - REST API - Interactive UI - Scalable - Production-Ready Tech stack: Python Why it matters: This project provides a robust solution for generating accurate, contextually grounded answers from large volumes of unstructured text data, making it highly valuable for applications requiring high-quality information retrieval and processing. Repository: AnkitNayak-dev/EpsteinFiles-RAG GitHub stars: 386 GitHub description: A RAG pipeline implementation built on the 'Epstein Files 20K' dataset from Hugging Face (Teyler).
Python 386 stars

graykode/commit-autosuggestions

A tool that AI automatically recommends commit messages.

Overview: This project is an implementation of CommitBERT, a model that automatically generates commit messages based on pre-trained programming language models. It leverages the capabilities of CodeBERT to create commit messages for code changes, addressing the common issue of writing commit messages when adding or deleting code. Key features: - Automatically generates commit messages for code changes - Supports Python, JavaScript, and a few other programming languages - Uses a pre-trained model for efficient processing Tech stack: Python Why it matters: This project addresses a common challenge in software development by providing a tool that automatically generates commit messages, making it easier for developers to manage their code changes and improve collaboration within the team.
Go 386 stars

mb-14/gomarkov

Markov chains in golang

### Overview: gomarkov is a Go implementation of Markov chains designed for textual data analysis. It provides tools for creating and manipulating Markov chains, generating new text based on given seeds, and even serializing the chain for later use. ### Key features: - Markov chains for textual data analysis. - JSON serializable for easy integration and persistence. - Tools for generating new text based on initial seeds. - Training data can be added to the chain for more accurate predictions. ### Tech stack: Go ### Why it matters: gomarkov is a valuable tool for anyone working with natural language processing, text generation, or Markov chains. Its Go implementation makes it easy to integrate into Go-based projects, and its JSON serialization allows for easy storage and retrieval of Markov chains.
Jupyter Notebook 386 stars

pierpaolo28/Artificial-Intelligence-Projects

Collection of Artificial Intelligence projects.

Overview: This repository contains a collection of Artificial Intelligence projects created by the author. The projects cover various topics such as probability distributions, augmented and virtual reality, machine learning deployment, and more. Key features: - Probability Distributions In Data Science - Getting Started With Augmented And Virtual Reality - How To Deploy Machine Learning Models On Mobile And Embedded Devices - Stochastic Processes Analysis - Online Machine Learning With Tensorflow.js - Deep Learning Analysis Using Large Model Support - Gpu Accelerated Data Analytics & Machine Learning - Big Data Analysis: Spark And Hadoop - Understanding Cancer Using Machine Learning - Flask And Heroku For Online Machine Learning Deployment - Computational Biology - SQL For Data Science - Reinforcement Learning For Anyone Open AI Gym And Ray - Introduction To Evolutionary Algorithms Tech stack: Jupyter Notebook Why it matters: The author has shared several projects on their blog and has received 386 GitHub stars, indicating that the projects are valuable and relevant to the developer community.
Go 386 stars

shixzie/nlp

[UNMANTEINED] Extract values from strings and fill your structs with nlp.

Overview: `nlp` is a general-purpose Natural Language Processor that extracts values from strings and fills structs with NLP data. Key features: - Registers models for specific types of data (e.g., `int`, `string`, `time.Time`). - Parses text and fills structs with extracted data. - Supports various data types including integers, floats, strings, and timestamps. - Uses Naive Bayes algorithm for learning models from samples. Tech stack: Go Why it matters: `nlp` is useful for developers who need to extract specific data from text and fill structs with NLP results, making it easier to work with structured data in Go. Repository: shixzie/nlp GitHub stars: 386 GitHub description: [UNMANTEINED] Extract values from strings and fill your structs with nlp.
Python 386 stars

thunlp/TensorFlow-Summarization

### Overview: This project implements a sequence-to-sequence model using a bidirectional GRU encoder and a GRU decoder, specifically for abstractive short text summarization. The model aims to help users quickly start working on summarization tasks and potentially also for machine translation. ### Key Features: - Bidirectional GRU Encoder - GRU Decoder - Sequence-to-Sequence Model - Supports DUC2003, DUC2004, and Giga datasets - Pre-trained models available for download - Easy-to-use script for training, testing, and generating summaries ### Tech Stack: Python ### Why It Matters: This project provides a practical solution for abstractive text summarization, making it accessible for users without deep knowledge of TensorFlow. It also demonstrates the effectiveness of sequence-to-sequence models in natural language processing tasks. ### Repository: thunlp/TensorFlow-Summarization - GitHub Stars: 386 - Primary Language: Python - GitHub Description: No description provided. - GitHub Repository: [thunlp/TensorFlow-Summarization](https://github.com/thunlp/TensorFlow-Summarization) ### Usage: - **Setup Environment:** - **With GPU:** ```bash pip3 install -U tensorflow-gpu==1.1 ``` - **Without GPU:** ```bash pip3 install -U tensorflow==1.1 ``` - **Model and Data:** - Files should be organized