Projects

### Overview: 智能故障诊断和寿命预测期刊（Journals of Intelligent Fault Diagnosis and Remaining Useful Life）旨在提供高质量的学术资源，帮助研究人员快速找到符合研究兴趣的期刊，从而提高论文发表和学术成果的影响力。 ### Key features: 1. **高质量论文推荐**：提供最新、高质量的学术期刊和会议推荐，帮助读者快速定位研究方向。 2. **分区表更新**：定期更新分区表，包括中科院分区、大区升级和降级情况，以及小类分区。 3. **小类详细分区**：提供更详细的期刊分区信息，帮助读者更好地理解期刊的学术影响力。 4. **CCF/CAAI等级**：提供计算机类和软件类学科认可的期刊目录，帮助读者了解学术期刊的学术水平。 5. **预警期刊**：提供中科院预警期刊列表，帮助读者了解学术期刊的学术风险。 6. **强制OA期刊**：推荐一些提供OA（开放获取）的期刊，方便读者获取最新研究成果。 ### Tech stack: - **GitHub**：用于存储和管理项目代码、文档和资源。 ### Why it matters: - 提供高质量的学术资源，帮助研究人员快速找到符合研究兴趣的期刊，提高论文发表和学术成果的影响力。 - 通过分区表和小类分区信息，帮助读者更好地理解期刊的学术影响力和研究方向。 - 提醒读者关注中科院分区、预警期刊和强制OA期刊，帮助他们选择合适的期刊投稿，提高投稿成功率。 ### Repository: liguge/J

### Overview: Slot Attention is an implementation of the Slot Attention model from the paper "Object-Centric Learning with Slot Attention" by Francesco Locatello et al. This model is designed for object-centric learning and is implemented using PyTorch. It can handle varying numbers of slots and is adaptable to different scenarios. ### Key Features: - **Slot Attention Model**: Implements the Slot Attention mechanism from the paper. - **Adaptive Slot Method**: Allows for dynamic adjustment of the number of slots used in the model. - **Flexible Training**: Supports training with varying numbers of slots. - **Integration with PyTorch**: Built on top of PyTorch for seamless integration with existing projects. ### Tech Stack: - **Programming Language**: Python - **Dependencies**: PyTorch ### Why It Matters: Slot Attention is particularly useful in scenarios where the number of objects or entities being tracked can vary, making it adaptable to different use cases and improving model generalization.

### Overview: Seedance 2.5 API is a Python wrapper designed for ByteDance's Seedance 2.5 AI video generator. It offers comprehensive text-to-video, image-to-video, and realistic human face generation capabilities, including 4K and 1080p support. This project is powered by MuAPI, providing high-fidelity video outputs with advanced features like realistic human face generation. ### Key Features: - **Text-to-Video**: Generate cinematic AI videos from text prompts. - **Image-to-Video**: Create videos from static images. - **Realistic Human Face Generation**: Produce high-quality videos with realistic human faces. - **4K and 1080p Support**: Deliver videos in native 4K and 1080p resolutions. - **Consistent Character Generation**: Ensure consistent character generation across different prompts and images. ### Tech Stack: - Python ### Why It Matters: Seedance 2.5 API simplifies the process of generating high-quality AI videos, making it easier for developers to integrate advanced video generation into their projects. The project's comprehensive features and support for various resolutions make it suitable for both professional and educational use cases, including educational videos, marketing campaigns, and more. ### GitHub Stars: 500 ### Related Projects: - [seedance-2.0-watermark-remover](https://github.com/SamurAIGPT/seedance-2.0-watermark-remover) — Removes watermarks from Seedance-generated videos. - [seedance-2-generator

Overview: This repository is a comprehensive paper list dedicated to recent works involving Mamba in Computer Vision, focusing on various applications and techniques. Key features: - A detailed survey of Mamba-based models and methods in Computer Vision. - Includes surveys on Mamba in Remote Sensing, Medical Imaging, and more. - Contains recent papers on Mamba architectures and applications. Tech stack: Python, Markdown, LaTeX, and GitHub Pages. Why it matters: This repository serves as a valuable resource for researchers and practitioners in Computer Vision, providing a clear overview of the latest advancements and methodologies in the field.

### Overview: AI Ethos, Inc. is a 501(c)(3) nonprofit organization dedicated to ethical, inclusive, multilingual AI tutoring for underserved learners worldwide. The organization focuses on privacy-first, culturally relevant, and open-source tutoring that prioritizes the most underserved learners: low-income, rural, multilingual, and neurodiverse students. The primary language used is PHP, and the project has 498 GitHub stars. ### Key Features: - **Localized, Learner-Centered Tutoring**: Personalized academic support tailored for low-income and rural learners, even in low-connectivity settings. - **Multilingual & Culturally Responsive Learning**: Tutoring in native languages that reflect local contexts. - **Adaptive & Supportive Instruction**: Socratic, step-by-step guidance that adapts to each learner's pace and needs. - **Privacy-First Student Care**: Student data is protected by ethical design and never monetized. - **Community-Co-Created Content**: Lessons shaped with educators and local communities for relevance and trust. ### Tech Stack: PHP ### Why it Matters: AI Ethos empowers equitable AI learning globally and ethically, bridging the AI education divide through localized, multilingual, and community-inclusive tutoring that prioritizes the most underserved learners. The organization's focus on privacy-first, cultural relevance, and ethical practices ensures that AI tutoring is accessible and beneficial to all learners, regardless of their background or circumstances.

Overview: LLaVaVision is a simple "Be My Eyes" web app powered by llama.cpp/llava, showcasing a llama.cpp/llava backend and utilizing the SkunkworksAI BakLLaVA-1 model for text description and Web Speech API for narration. Key features: - llama.cpp/llava backend for text description - SkunkworksAI BakLLaVA-1 model for text generation - Web Speech API for narration Tech stack: JavaScript Why it matters: LLaVaVision demonstrates the integration of llama.cpp/llava with a llama.cpp backend, showcasing the capabilities of the SkunkworksAI BakLLaVA-1 model in a practical application.

Overview: SHATTERED is a local-first AI-powered document intelligence platform designed for investigative journalism. It provides a modular and user-friendly interface for analyzing documents and uncovering insights. Key features: - Local-first data management - AI-powered document analysis - Interactive network analysis - Time-based event extraction - Pattern recognition across documents - Credibility assessment of sources - Hybrid semantic/keyword search - Image authenticity analysis Tech stack: Python Why it matters: SHATTERED offers a practical solution for investigative journalists by providing a platform that respects data sovereignty and privacy, making it accessible to those who can't afford expensive cloud services. Repository: mantisfury/ArkhamMirror GitHub stars: 498 GitHub description: Local-first AI-powered document intelligence platform for investigative journalism

### Overview: Explore the latest AI tools, platforms, and resources across various domains, curated by the creator of AI-Catalog. ### Key features: - **Creative AI**: Includes tools for text-to-image, video generation, audio editing, deep voice synthesis, photo editing, and 3D modeling. - **Content & Writing**: Features a summarizer, prompt generator, writing, and copywriting tools. - **Developer Tools**: Offers a code assistant, low-code/no-code tools, SQL tools, builder, website builder, and landing page generator. - **AI Assistants & Chat**: Includes multi-modal, large language models, search engines, chatbots, and chatGPT plugins. - **Education & Learning**: Features education assistants and tools. - **Business & Productivity**: Includes presentation tools, email assistant, start-up tools, AI productivity tools, and AI detection. ### Tech stack: HTML ### Why it matters: One of the first AI-list repos on GitHub, the catalog is a valuable resource for developers and creatives alike, providing a comprehensive overview of AI tools across various domains.

### Overview: ScaleLLM is a high-performance inference system designed for large language models, optimized for use in production environments. It supports a variety of popular models and provides features like tensor parallelism, continuous batching, and advanced caching techniques. ### Key Features: - **High Efficiency**: Utilizes state-of-the-art techniques such as Flash Attention, Paged Attention, and continuous batching to achieve high performance. - **Tensor Parallelism**: Efficiently executes models using tensor parallelism. - **OpenAI-compatible API**: Provides an OpenAI-compatible REST API server for chat and completions. - **Seamless Integration**: Seamless integration with popular Hugging Face models using safetensors. - **Customizable**: Offers flexibility for customization to meet specific needs. - **Production Ready**: Engineered for robust system monitoring and management, ensuring a seamless deployment experience. ### Tech Stack: C++ ### Why It Matters: ScaleLLM is designed to meet the demands of production environments, offering high performance and ease of use. Its support for popular models and advanced features make it a valuable tool for developers looking to enhance their LLM inference capabilities.

### Overview: CS224n-winter-together 是一个旨在提供斯坦福大学 CS224n 课程笔记、感悟与延伸、课程作业与项目分享的开源平台。该项目由微信公众号“夕小瑶的卖萌屋”发起，旨在方便大家参考学习。视频资料来源于 2019 年，其他资料为 2020 年，鼓励使用 Markdown 格式提交内容，以减少 repo 大小。 ### Key Features: - **课程笔记与感悟**：提供官方课程笔记、个人感悟和项目经验。 - **课程作业**：上传个人完成的作业和项目。 - **项目实战**：上传队伍完成的项目实战内容，目前暂未开放。 ### Tech Stack: - JavaScript ### Why it Matters: 本项目为学习者提供了一个共享和沉淀学习资源的平台，便于大家相互学习和交流，同时鼓励使用 Markdown 格式进行提交，以减少 repo 大小，保持项目整洁。

### Overview: KcBERT: Korean comments BERT is a pretrained BERT model and wordpiece tokenizer trained on Korean comments. It has been used for various downstream tasks and has shown significant improvements over existing models. ### Key Features: - **Pretrained BERT Model and Wordpiece Tokenizer**: Trained on Korean comments. - **Extensive Text Dataset**: Includes a publicly available dataset of expanded text data from 2022.3Q. - **Performance Improvements**: Up to 1% performance gains in downstream tasks compared to previous versions. - **Tokenizer and Preprocessing**: Uses a tokenizer and preprocessing code that has been updated to improve performance. - **GitHub Repository**: Publicly available on GitHub with 497 stars and a detailed README. ### Tech Stack: Python, PyTorch, Huggingface Transformers. ### Why It Matters: KcBERT is a powerful tool for natural language processing tasks, particularly in the context of Korean language data, offering improved performance compared to existing models. Its ability to handle unprocessed Korean data makes it highly valuable for researchers and practitioners working with Korean language datasets.

### Overview: Harvestify is a machine learning-based website designed to recommend the best crops to grow, fertilizers to use, and diseases caught by crops. The project is featured on Krish Naik's YouTube Channel and has been featured on Kaggle Notebooks. The dataset used includes custom-built datasets for crop recommendation, fertilizer suggestion, and disease detection. ### Key Features: - Crop Recommendation: Predicts the best crop to grow based on soil data. - Fertilizer Recommendation: Suggests the right fertilizers based on soil and crop type. - Plant Disease Prediction: Identifies and predicts diseases on plant leaves. ### Tech Stack: - Python - HTML - CSS - JavaScript - TensorFlow - Keras - Flask - Git ### Why It Matters: This project presents a practical application of machine learning and deep learning in precision farming, which can significantly improve crop yield and farmer efficiency. By using real-world data, it demonstrates how these technologies can be implemented to enhance agricultural practices.

Overview: Spatial-MLLM is an innovative method designed to enhance the visual-based spatial intelligence of existing video MLLMs, achieving state-of-the-art performance across various spatial reasoning tasks. Key features: - Incorporates a 2D visual encoder - Utilizes a spatial encoder initialized from a feed-forward visual geometry foundation model - Includes a connector to integrate the visual and spatial information - Employs a large language model backbone - Implements a space-aware frame sampling strategy to select spatially informative frames Tech stack: Python Why it matters: Spatial-MLLM significantly improves the visual-based spatial intelligence of MLLMs, making it highly relevant for applications requiring spatial reasoning, such as autonomous driving, augmented reality, and more. Repository: THU-SI/Spatial-MLLM GitHub stars: 497

Overview: Emojinator is a simple emoji classifier designed for humans, helping you recognize and classify different hand emojis. Key features: - Train Machine Learning model - Create Gestures - Emojinator Application Tech stack: Python Why it matters: Emojinator simplifies the process of recognizing and classifying emojis, making it easier for humans to understand and interact with emojis in various contexts. Repository: akshaybahadur21/Emojinator Primary language: Python GitHub stars: 497

Overview: Sacremoses is a Python port of Moses tokenizer, truecaser, and normalizer, designed for natural language processing tasks. Key features: - Tokenizer - Truecaser - Normalizer Tech stack: Python Why it matters: Sacremoses simplifies the process of tokenizing, truecasing, and normalizing text, making it easier for developers to work with language data in Python.

### Overview: Redis is a popular in-memory data structure store that has found extensive use in the AI ecosystem. This curated repository provides a wealth of resources, including code recipes, demos, tutorials, and integrations, to help developers leverage Redis for advanced AI applications. ### Key Features: - Vector Search with RedisVL - RAG from Scratch - Redis RAG Workbench demo - Vector Search with RedisVL - Redis VSS - Simple Streamlit Demo - ArXiv Search ### Tech Stack: Jupyter Notebook ### Why it Matters: Redis is a versatile tool that can significantly enhance the performance and scalability of AI applications. This repository offers practical solutions and examples that developers can use to build efficient and effective AI systems.

Overview: HumanRF is an official repository for High-Fidelity Neural Radiance Fields for Humans in Motion, providing a comprehensive solution for realistic human motion simulation. Key features: - High-fidelity neural radiance fields for realistic human motion - Includes ActorsHQ dataset and data loader - Official installation guide and usage examples Tech stack: Python, GLM, Tiny CUDA NN, ActorsHQ Why it matters: This repository offers a reliable and detailed solution for simulating realistic human motion, making it valuable for researchers and developers in the field of computer graphics and simulation. Repository: synthesiaresearch/humanrf Primary language: Python GitHub stars: 497

### Overview: Medical Diffusion is a project that focuses on developing a denoising diffusion probabilistic model specifically for 3D medical image synthesis. The repository contains the code for the paper "Medical Diffusion: Denoising Diffusion Probabilistic Models for 3D Medical Image Synthesis" (arXiv:2211.03364). ### Key Features: - **Code Repository**: Contains the implementation of the paper's code. - **Dataset Support**: Supports training on the BraTS dataset and can be adapted for datasets of your own. - **Model Training**: Includes training of a three-dimensional VQ-GAN model and a diffusion model in the latent space of the VQ-GAN model. ### Tech Stack: - **Python**: The project is primarily developed using Python v3.8. ### Why It Matters: Medical Diffusion offers a novel approach to 3D medical image synthesis, potentially improving the quality and efficiency of medical imaging by leveraging diffusion models. The project's implementation and training capabilities make it a valuable resource for researchers and developers in the medical imaging field. ### Repository: - **Repository**: FirasGit/medicaldiffusion - **GitHub Stars**: 495 - **Primary Language**: Jupyter Notebook - **GitHub Description**: This repository contains the code to our paper "Medical Diffusion: Denoising Diffusion Probabilistic Models for 3D Medical Image Synthesis" (see https://arxiv.org/abs/2211.03364). ### System Requirements:

### Overview: SDL-MCP is a policy-centered context budget layer for coding agents, transforming sprawling codebases into compact, high-signal context that saves tokens, speeds up workflows, and improves agent output. ### Key features: - Symbol Delta Ledger (SDL-MCP) indexes a repository into a symbol graph and gives coding agents a controlled path from compact metadata to source code. - SDL-MCP turns sprawling codebases into compact, high-signal context that saves tokens, speeds up workflows, and improves agent output. - It supports the Model Context Protocol over stdio or HTTP, running locally and requiring Node.js 24 or later. ### Tech stack: TypeScript ### Why it matters: SDL-MCP is a powerful tool for managing and organizing codebases, making them more efficient and easier to work with.

Overview: Styleformer is a Neural Language Style Transfer framework designed to smoothly transfer natural language text between various fine-grained styles such as formal/casual, active/passive, and more. Created by Prithiviraj Damodaran, it aims to enhance the quality and readability of text across different contexts. Key features: - Transfer text from casual to formal - Transfer text from formal to casual - Transfer text from active to passive - Transfer text from passive to active Tech stack: Python Why it matters: Styleformer provides a powerful tool for text style transfer, making it easier to adjust the tone and style of text in various applications, such as improving the readability of machine-generated content or enhancing the style of paraphrased text. Repository: PrithivirajDamodaran/Styleformer Primary language: Python GitHub stars: 496 GitHub description: A Neural Language Style Transfer framework to transfer natural language text smoothly between fine-grained language styles like formal/casual, active/passive, and many more. Created by Prithiviraj Damodaran. Open to pull requests and other forms of collaboration.