Research & Education

unet.cu is an open-source project that provides a UNet diffusion model implemented entirely in C++/CUDA. Inspired by Andrej Karpathy's llm.c, the goal is to achieve performance comparable to PyTorch implementations, specifically for training unconditional diffusion models. The repository includes benchmarks showing its training speed relative to PyTorch and PyTorch with `torch.compile`. It supports training with sample images from ImageNet 64x64 and allows users to train with their own data. The project emphasizes learning CUDA concepts and provides a detailed breakdown of its architecture, including custom convolution kernels and optimizations to avoid inefficient data transposes.

bili-hardcore is an AI-powered tool designed to automate the process of answering questions for Bilibili's hardcore member exams. Unlike OCR-based solutions, it directly interacts with the Bilibili API, ensuring higher accuracy and efficiency. The tool supports various large language models, including DeepSeek (V3.1) and Gemini (gemini-2.5-flash), with options for custom OpenAI-style APIs like those from Volcengine and SiliconFlow. Users can configure their preferred model and API key, and the tool handles the login via QR code and automatic question answering. It's crucial for users to have a Bilibili account at level 6 or above to participate in the hardcore member trials. The tool also provides guidance on troubleshooting common issues like QR code display problems, low accuracy, or API errors, and emphasizes responsible use in compliance with Bilibili's rules.

Stellon Labs is an AI research lab dedicated to developing powerful, tiny AI models specifically optimized for edge applications. Their focus is on creating 'frontier AI' solutions that can operate efficiently on minimal hardware, making advanced artificial intelligence accessible for devices with limited computational resources. The lab aims to push the boundaries of AI performance in constrained environments, enabling new possibilities for on-device intelligence without requiring extensive infrastructure. Their work is geared towards practical applications where low-power and small-footprint AI is crucial.

Git2Log is an AI-powered changelog generator that simplifies the process of creating polished, user-friendly product updates from technical Git commit messages. Designed for product managers, project managers, and marketing strategists, this tool helps bridge the gap between development and user communication. By leveraging AI, Git2Log ensures that product updates are clear, engaging, and easy for users to understand, enhancing feature adoption and reducing wasted effort on unused features. It is part of the broader ProductLift platform, which offers comprehensive solutions for feedback management, roadmap planning, and changelog generation.

AskVideo.ai is a free tool designed to transform how users interact with YouTube videos. By simply pasting a YouTube URL, the AI processes the video content, allowing users to ask questions and receive instant, accurate answers with corresponding timestamps. This eliminates the need for manual scrubbing through long videos, making learning and information extraction significantly more efficient. It's ideal for academic learning, tutorials, business insights, research, and team collaboration, offering features like AI-powered Q&A, video transcripts, and even a command-line interface for developers. The platform aims to save users hours of time by providing direct access to specific information within video content.

Study Tutor is an AI-powered personal tutor designed to help users master any subject efficiently. It offers instant explanations and practice problems, adapting to individual learning paces to provide personalized support. The tool allows for unlimited questions and delivers patient guidance across a wide range of subjects, making complex topics accessible and understandable. It aims to enhance learning by providing a continuous feedback loop and tailored educational content, ensuring users can build a strong foundation and improve their understanding at their own speed. This makes it an invaluable resource for anyone looking to deepen their knowledge or prepare for exams.

RubyGPT.ai is a user-friendly AI writing assistant designed to streamline content creation for a wide range of needs. It empowers users to effortlessly generate diverse written content, from academic essays and marketing taglines to creative poems and stories. The tool aims to eliminate writer's block and the need for boring content, allowing users to produce beautiful writings quickly. With features like generating ideas, summarizing notes, improving existing text, and explaining complex concepts, RubyGPT.ai serves as a versatile companion for anyone looking to enhance their writing process and create engaging content efficiently.

Quenti is a modern learning platform designed to help students study anything effectively. It offers various study modes including Learn, Flashcards, Test, and Match, all generated from user-created terms. The platform supports importing study sets from Quizlet and understands over 100 languages, including accent help and answer grading. Quenti Cortex, powered by AI, enhances learning through smarter multiple-choice questions and intelligent written response grading, understanding answers by meaning across 27 different languages. It also features a command menu for effortless navigation and keeps track of user progress, allowing them to pick up right where they left off. Quenti is free forever for students.

Pytorch-bert-crf-ner offers a PyTorch implementation for Korean Named Entity Recognition (NER) tagging, leveraging the power of BERT and CRF models. This open-source tool is specifically designed to assist in Korean Natural Language Processing (NLP) tasks and research. It provides functionalities to identify and classify named entities such as persons, locations, organizations, dates, and more within Korean text. The repository includes examples, data utilities, and training scripts, making it suitable for developers and researchers working with Korean language data who need to implement or experiment with NER models.

Pytorch-maddpg offers a PyTorch implementation of the Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm, a key approach in multi-agent reinforcement learning. This open-source project is hosted on GitHub and is designed for researchers and developers working on complex multi-agent systems. The implementation includes a modified Waterworld environment, where agents (evaders, pursuers, poisons) interact under specific physical rules, allowing for experimentation with cooperative behaviors. It supports features like agents bouncing off walls and requiring exact cooperation for rewards, making it a valuable tool for studying multi-agent coordination and policy learning.

XVERSE-13B is a multilingual large language model developed by XVERSE Technology Inc. It features a Decoder-only Transformer network structure with an 8K context length, which is extended to 256K in the XVERSE-13B-256K version for handling extensive input content like literature summaries and report analysis. The model was trained on 3.2 trillion tokens across over 40 languages, with a focus on Chinese and English performance. It utilizes a 100,534-token BPE-based tokenizer that supports multiple languages without requiring additional vocabulary expansion. The project also highlights an efficient training framework with high peak computing power utilization. Quantized models (GGUF, GPTQ) are available for inference on MacOS, Linux, and Windows systems.

mader is an open-source trajectory planner specifically designed for use in multi-agent and dynamic environments. It has been accepted for publication in the IEEE Transactions on Robotics (T-RO), highlighting its academic rigor and practical applicability. The tool facilitates trajectory planning for robotic systems, including single-agent and multi-agent simulations, with features like obstacle avoidance and dynamic environment handling. Users can set up and run simulations using ROS, with options for both Docker and non-Docker installations. It supports backend optimizers like Gurobi and NLOPT, providing flexibility for different computational needs. The project is hosted on GitHub by MIT-ACL, making it accessible for researchers and developers in the robotics community.

Micronet is an open-source library designed for AI model compression and efficient deployment on various hardware platforms. It provides a comprehensive suite of techniques including quantization-aware training (QAT) and post-training quantization (PTQ) for both high-bit and low-bit scenarios, as well as pruning methods like normal, regular, and group convolutional channel pruning. The library also supports batch-normalization fusion for quantization, enhancing model efficiency. For deployment, Micronet integrates with TensorRT, enabling optimized inference in fp32, fp16, and int8 formats with features like op-adapt and dynamic shape support. This makes it an invaluable tool for developers looking to reduce model size and accelerate inference speed.

Dilseheal is an AI-powered hospital recommendation system designed to connect patients with the most suitable healthcare providers. It leverages artificial intelligence to recommend hospitals based on various factors including medical condition, financial status, valid user reviews, and hospital background checks. The platform aims to streamline the process of finding the right treatment by offering information about hospitals near the user. Beyond recommendations, Dilseheal allows users to track health issues through vaccination and monitoring records, and facilitates the creation, editing, and sharing of medical records for controlled collaboration between doctors. It emphasizes verified hospitals, expert consultation, 24/7 support, and a patient community.

MonBotAchat is an AI-powered online shopping assistant designed to simplify the purchasing process for consumers. It offers personalized product recommendations, allowing users to discover innovative products tailored to their preferences. The platform provides intuitive price comparisons across a wide range of categories, including high-tech, fashion, beauty, and leisure, ensuring users find the best deals. MonBotAchat aims to deliver an intuitive and natural shopping experience, offering exceptional offers, current fashion trends, and gift ideas. Its features include easy and fast search capabilities, coupled with personalized advice to enhance the overall shopping journey.

DashPlayer is a specialized video player crafted for English language learners, aiming to enhance language acquisition through immersive video watching. It offers a suite of features tailored for language study, including dual-language subtitles (English/Chinese), the ability to jump between sentences, and an integrated dictionary for quick word lookups by hovering over text. Users can also leverage AI to generate subtitles for videos lacking them and utilize an AI-powered sentence learning function. The player supports adjustable interface sizes, records playback positions, and can be controlled via Bluetooth remotes. It also includes video download and cutting functionalities, making it a comprehensive tool for self-directed English learning.

The serverless-ml-course is an open-source educational resource designed to simplify the development and operation of AI-enabled prediction services. It teaches how to build batch and real-time prediction services using Python, focusing on serverless infrastructure. The course covers essential MLOps fundamentals such as versioning, testing, data validation, and operations, enabling users to deploy features and models, train models, and run inference pipelines. A key differentiator is its emphasis on building a prediction service around a model without needing extensive operations experience, making it accessible for those who can program in Python but are not cloud computing experts. It also guides users on building serverless UIs for their prediction services.

pyannote-audio is an open-source Python toolkit designed for speaker diarization, a process that identifies 'who spoke when' in an audio recording. Built on the PyTorch machine learning framework, it offers robust capabilities for speech activity detection, speaker change detection, and speaker embedding. The toolkit includes pretrained models and pipelines, allowing users to quickly implement and experiment with audio analysis tasks. Furthermore, it supports fine-tuning of these models, enabling users to optimize performance on their specific custom datasets. This makes pyannote-audio a versatile tool for researchers and developers working with audio data.

reddit-ai-trends is an open-source tool designed to help users stay ahead of AI trends by providing automated insights from Reddit. It scans AI-related communities in both English and Chinese, leveraging the Reddit Official API and DeepSeek R1 by OpenRouter for in-depth analysis. The tool summarizes key discussions, tracks trends, and generates daily reports, including hot topics and emerging trends. It supports multimodal content analysis, including image analysis with vision models, YouTube video transcript extraction, and web page content scraping. Reports are generated daily and stored in an organized file structure, with bilingual support and smart caching to minimize API costs.

BashBuddy is an AI assistant designed to simplify command-line interactions by allowing users to write commands in natural language. It translates plain text or half-baked commands into precise shell commands, handling complex syntax and arguments. The tool is context-aware, understanding the shell environment, current directory, and even Git repositories to provide relevant suggestions. BashBuddy is open-source, cross-platform (macOS, Linux, Windows, Bash, Zsh, PowerShell), and works seamlessly across different operating systems and shells. It runs completely locally, ensuring privacy and offline functionality, and leverages GPU acceleration for fast inference performance.

AI Essay Writer: AI Assistant, developed by Quantum4U Lab, is an iOS mobile application designed to empower users with intelligent writing assistance. This tool utilizes advanced AI algorithms to help generate high-quality essays, articles, and various creative content, streamlining the writing process. Beyond basic text generation, the app aims to improve overall writing skills by providing smart suggestions and comprehensive support. It's part of Quantum4U Lab's suite of innovative mobile apps, focusing on user-centric design and cutting-edge technology to simplify tasks and enhance creativity directly from a mobile device.

PATAMDE is an AI-powered learning platform designed to revolutionize exam preparation for students. It enables users to convert any study material, including textbook pages, handwritten notes, PDFs, and text, into interactive quizzes instantly. The platform leverages AI to analyze content, generate personalized quizzes, identify knowledge gaps, and provide targeted study recommendations. PATAMDE supports a wide range of competitive exams in India, such as UPSC, JEE Main & Advanced, NEET, SSC CGL, CAT, and GATE, among others. It also offers AI chat tutors, study guides, and adaptive learning analytics to enhance the learning experience. Available in 11+ Indian languages, PATAMDE aims to make smart study accessible and effective for students across India.

FindU is a free college search application designed to help students discover colleges that align with their interests, goals, and personality. Unlike traditional search tools, FindU utilizes a comprehensive quiz to understand academic interests, career goals, preferred location, and campus culture, then matches students with suitable institutions from a database of over 4,000 colleges and universities across the United States. The app also features a built-in scholarship tracker to help students find opportunities, manage deadlines, and monitor their earnings. Additionally, FindU facilitates collaboration by allowing students to share their college lists, scholarship applications, and deadlines with parents and counselors, ensuring everyone stays informed and organized throughout the college application process.

Lightly-train is a comprehensive, open-source framework designed to streamline the development of state-of-the-art computer vision models. It covers the entire model development lifecycle, from pretraining DINOv2/v3 vision foundation models on unlabeled data to fine-tuning transformer and YOLO models for detection and segmentation tasks. The tool supports various applications such as object detection, panoptic segmentation, instance segmentation, semantic segmentation, and image classification. It also offers distillation capabilities, allowing users to transfer knowledge from DINOv2 or DINOv3 foundation models into their custom architectures, leading to improved performance across different models like YOLOv8, YOLOv11, RT-DETR, and Faster R-CNN. Lightly-train is installable via pip and provides clear usage examples for different vision tasks.