Projects

Overview: tf2onnx is a tool that converts TensorFlow, Keras, Tensorflow.js, and Tflite models to ONNX format. Key features: - Converts TensorFlow, Keras, Tensorflow.js, and Tflite models to ONNX format - Supports ONNX opset versions 14 to 18 - Provides command-line and Python API options - Supports ONNX Runtime for runtime inference Tech stack: Python Why it matters: tf2onnx simplifies the conversion process for developers and researchers working with various TensorFlow-based models, making it easier to integrate these models into the ONNX ecosystem for further analysis, optimization, or deployment.

### Overview hCaptcha Challenger is a project designed to help developers face hCaptcha challenges with a multimodal large language model. It provides a seamless interface to face these challenges without relying on any tampermonkey scripts or third-party anti-captcha services. The project is built using Python and is open-source on GitHub, with a high number of stars indicating its popularity and usefulness. ### Key Features - **Image Labeling (Binary)**: Uses ResNet for binary classification. - **Image Labeling (Area Select: Point)**: Utilizes YOLOv8 for point detection. - **Image Labeling (Area Select: Bounding Box)**: Implements YOLOv8 for bounding box segmentation. - **Image Labeling (Multiple Choice)**: Applies ViT for zero-shot motion detection. - **Image Drag and Drop**: Uses Spatial Chain-of-Thought for this task. - **Ranking Strategy**: Supports nested-model-zoo for advanced ranking. - **Self-Supervised Challenge**: Implements CLIP-ViT for a self-supervised challenge. - **Agentic Workflow**: Employs AIOps Multimodal Large Language model for a more efficient workflow. ### Tech Stack The project is built using Python, leveraging the capabilities of a large language model to face hCaptcha challenges effectively. ### Why It Matters hCaptcha Challenger simplifies the process of facing hCaptcha challenges, making it easier for developers to integrate this functionality into their projects. By using a multimodal large language model, the project ensures that the challenges are handled efficiently and effectively,

### Overview: OFA (OFA-Sys) is an official repository of the OFA (Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework) project. This project is designed to unify modalities and tasks through a simple sequence-to-sequence learning framework, supporting English and Chinese. OFA is used for tasks such as image captioning, visual grounding, visual question answering, text-to-image generation, text classification, text generation, image classification, and more. The repository provides step-by-step instructions for pretraining and finetuning, as well as corresponding checkpoints. ### Key Features: - **Image Captioning**: Supports English and Chinese. - **Visual Grounding**: Supports English and Chinese. - **Visual Question Answering**: Supports English and Chinese. - **Text-to-Image Generation**: Supports English and Chinese. - **Generic Interface**: Supports English and Chinese. - **Chinese OCR**: Supports Chinese text recognition. ### Tech Stack: - Python ### Why It Matters: OFA is a unified sequence-to-sequence pretrained model that can be used for a wide range of tasks and modalities. It simplifies the process of pretraining and finetuning by providing step-by-step instructions and corresponding checkpoints, making it easier for developers to integrate into their projects. ### Repository: - Repository: OFA-Sys/OFA - GitHub Stars: 2553 - GitHub Description: Official repository of OFA (ICML 2022). Paper: OFA: Unifying Architectures, Tasks, and Modal

Overview: DeepDetect is a deep learning runtime, command-line tool, and REST server designed for practical model operations, including creating services, training models, running predictions, monitoring jobs, and managing model repositories on the filesystem. Key features: - Supports images, text, CSV/tabular data, time series, sparse/SVM-style data, object-detection boxes, and segmentation masks through a single API surface. - Provides a Python wheel that embeds the DeepDetect runtime in the current Python environment and a CLI for repeatable model workflows. - Supports training and inference using the Torch and TensorRT frameworks. - Focuses on practical model operations, making it suitable for real-world applications. Tech stack: C++ and Python. Why it matters: DeepDetect simplifies the process of deploying and managing deep learning models, making it easier to create, train, and run models without the need for extensive knowledge of deep learning frameworks and environments. Repository: jolibrain/deepdetect GitHub stars: 2551

### Overview: Hivemind is a decentralized deep learning library built on PyTorch, designed to train large models across thousands of computers from various sources. This project aims to facilitate collaborative training of neural networks without a central authority, ensuring fault tolerance and efficient parameter aggregation. ### Key features: - Distributed training without a master node. - Fault-tolerant backpropagation. - Decentralized parameter averaging. - Training of neural networks of arbitrary size. ### Tech stack: Python ### Why it matters: Hivemind enables decentralized deep learning, allowing models to be trained on a global network of volunteers, making it ideal for distributed computing and enabling collaborative research without the constraints of a central authority.

Overview: AingDesk is a simple and easy-to-use AI assistant that supports knowledge bases, model APIs, sharing, web search, and intelligent agents, growing rapidly. Key features: - One-click deployment of local AI models and mainstream model APIs - Local knowledge base - Intelligent agent creation - Can be shared online for others to use - Supports web search - Supports server-side deployment Tech stack: TypeScript Why it matters: AingDesk is an easy-to-use AI assistant that supports local AI models, APIs, and knowledge base setup, growing rapidly.

### Overview: Auto-PyTorch is an automatic architecture search and hyperparameter optimization framework designed specifically for PyTorch, focusing on the optimization of neural network architectures and training hyperparameters. This tool aims to automate deep learning tasks, particularly for tabular and time series data, by leveraging the strengths of both traditional AutoML and neural architecture search. ### Key features: - **Automatic Architecture Search**: Automatically identifies the most suitable neural network architecture for a given dataset. - **Hyperparameter Optimization**: Optimizes the hyperparameters of the selected architecture to improve model performance. - **Efficient Training**: Utilizes efficient training techniques to speed up the training process. ### Tech stack: Auto-PyTorch uses SMAC (Scalable Model-based Algorithm Configuration) as its optimization package to ensure robust and efficient performance. ### Why it matters: Auto-PyTorch simplifies the process of automating deep learning tasks, making it accessible to researchers and practitioners. By automating the selection of neural network architectures and the optimization of hyperparameters, it reduces the need for extensive manual tuning, thereby accelerating the development and deployment of machine learning models.

Overview: 动手学Ollama，快速实现大模型本地部署，让CPU也能玩转大模型推理。 Key features: - Ollama介绍、安装和配置 - 自定义导入模型 - Ollama REST API使用指南 - Ollama在LangChain中的使用 - Ollama可视化界面部署和应用案例 Tech stack: Jupyter Notebook Why it matters: 本教程提供清晰的步骤和实用的技巧，帮助学习者快速上手Ollama，实现本地部署大模型以及相关应用，让大模型赋能千行百业。

Overview: uzu is a high-performance inference engine designed for deploying AI models directly in applications, offering zero latency, full data privacy, and no inference costs. Key features: - Simple, high-level API - Unified model configurations - Traceable computations - Utilizes unified memory on Apple devices - Supports a broad range of models Tech stack: Rust Why it matters: uzu simplifies the deployment of AI models by providing a straightforward API, making it easy to integrate AI into applications without compromising performance or data security. Repository: trymirai/uzu GitHub stars: 2384 GitHub description: A high-performance inference engine for AI models README excerpt: [Link to README](https://github.com/trymirai/uzu/blob/main/README.md)

Overview: mPLUG-Owl is a powerful multi-modal large language model family that revolutionizes the field of natural language processing by integrating multimodal information. Key features: - mPLUG-Owl: Modularization Empowers Large Language Models with Multimodality - mPLUG-Owl2: Revolutionizing Multi-modal Large Language Model with Modality Collaboration - mPLUG-Owl3: Towards Long Image-Sequence Understanding in Multi-Modal Large Language Models Tech stack: Python Why it matters: mPLUG-Owl represents a significant advancement in the field of large language models, particularly in the integration of multimodal information, making it a valuable tool for applications requiring comprehensive understanding of both text and images.

Overview: GluonNLP is a comprehensive toolkit designed to simplify the development of natural language processing (NLP) applications using deep learning techniques. It provides a user-friendly interface for loading, processing, and training NLP models, making it accessible to both engineers and researchers. Key features: - Easy-to-use Text Processing Tools and Modular APIs - Pretrained Model Zoo - Write Models with Numpy-like API - Fast Inference via Apache TVM (experimental) - AWS Integration via SageMaker Tech stack: Python Why it matters: GluonNLP simplifies the complex process of NLP development by providing a streamlined approach to text processing, model training, and inference, making it an essential tool for anyone working with NLP projects.

### Overview: TranslateBooksWithLLMs is a desktop application that translates full-length books, subtitles, and documents using AI, preserving formatting and context. It supports EPUB, SRT, DOCX, PDF, and TXT formats and allows for seamless translation without size limits. ### Key Features: - **Formats Support:** Supports EPUB, SRT, DOCX, PDF, and TXT formats. - **No Size Limit:** Processes documents of any length, from a single page to thousand-page novels. - **Preserve Formatting:** Documents come out exactly as they went in, preserving EPUB formatting, styles, and structure. - **Resume Anytime:** Resumes translation where you left off, with checkpoint system. - **Reusable Writing Styles:** Extract a style preset from sample books or write one manually and apply it to every chunk. - **Auto Glossary & Style:** No glossary or preset ready? Pick Auto in the dropdown and the app derives both from the document being translated. - **PDF Support:** Translates text PDFs into a new, reflowed PDF, preserving headings, paragraphs, lists, and images but not the original layout. - **OCR Support:** Not supported for scanned PDFs, multi-column layouts are best effort, and the refinement pass is not available for PDF yet. ### Tech Stack: Python ### Why it Matters: TranslateBooksWithLLMs is a powerful tool for translators, allowing them to translate books and documents efficiently and accurately, while preserving the original formatting and context. It supports a wide range of formats and is designed to be

### Overview: This repository is a curated list of awesome deep learning-based papers on text detection and recognition, focusing on optical character recognition (OCR) using deep learning methods. It includes papers sorted by publication date, with scores for localization tasks provided in a leaderboard format. The repository also includes official codes and models for various tasks. ### Key Features: - Papers sorted by published date. - IC stands for ICDAR. - Score is F1-score for localization task. * (L) stands for score in the leader-board. * If the reported score in the leader-board is somewhat different from the paper, (L) is provided. - Official codes and models for various tasks. ### Tech Stack: - Official codes and models for various tasks. ### Why It Matters: This repository provides a comprehensive list of papers and models for text detection and recognition, making it an essential resource for researchers and practitioners in the field of computer vision and deep learning.

Overview: skfolio is a Python library designed for portfolio optimization, factor model construction, and risk management. Built on scikit-learn, it offers a unified interface and compatible tools for scikit-learn users. Key features: - Portfolio optimization - Factor model construction - Risk management Tech stack: Python Why it matters: skfolio simplifies portfolio optimization by providing a unified interface and compatible tools with scikit-learn, making it accessible to users familiar with scikit-learn. It is open-source under the 3-Clause BSD license and supports enterprise-level support and SLAs from Skfolio Labs. Repository: skfolio/skfolio Primary language: Python GitHub stars: 2526 GitHub description: Python library for portfolio optimization built on top of scikit-learn Downloads: 2526 CI/CD: GitHub Actions Python version: 3.10, 3.11, 3.12, 3.13, 3.14 PyPI: skfolio Codecov: codecov.io/gh/skfolio/skfolio/graph/badge.svg?token=KJ0SE4LHPV DOI: https://zenodo.org/badge/731792488.svg Contribution: https://github.com/skfolio/skfolio/blob/main/CONTRIBUTING.md Website: https://skfolio.org JupyterLite: https://jupyterlite.rtfd.io/en/latest/_static/badge.svg Discord

### Overview: Nanocoder is an open-source AI coding assistant built by a community collective, allowing users to bring their own models and keep their code on their machine without any external dependencies. It offers privacy-respecting, local-first, and open-source features. ### Key features: - **Model Choice**: Users can choose their preferred model via Ollama or any OpenAI-compatible API like OpenRouter, Requesty, Anthropic, and Google. - **Code Management**: Code is stored locally on the user's machine, ensuring privacy and security. - **Open Source**: The project is built by a community collective, not a company, making it accessible and free of any paid tiers. ### Tech stack: TypeScript ### Why it matters: Nanocoder empowers users to code locally and securely, fostering a more open and private coding environment. It is built by a community collective, ensuring it remains free of closed-source features and proprietary dependencies.

**Overview:** HyperLearn is a cutting-edge machine learning library designed to accelerate and optimize machine learning algorithms, significantly reducing computation time and memory usage. It is compatible with all hardware, including modern and older systems, making it an indispensable tool for developers and researchers alike. **Key Features:** - **Faster Algorithms:** Achieves 2-2000x faster training times and 50% less memory usage compared to traditional methods. - **Versatile Hardware Support:** Works seamlessly with new and old hardware, ensuring compatibility across a wide range of systems. - **Documentation:** Comprehensive documentation and tutorials available, including a detailed 50-page modern Big Data Algorithms PDF. **Tech Stack:** HyperLearn is written entirely in PyTorch, Numba, NumPy, Pandas, Scipy, and LAPACK, utilizing C++, C, Python, Cython, and Assembly. It mirrors Scikit-Learn's functionality, offering a seamless integration experience. **Why It Matters:** HyperLearn is essential for developers and researchers looking to enhance their machine learning workflows by significantly improving algorithm performance and efficiency, making it a valuable addition to any data science toolkit. **GitHub Stars:** 2,527 **Repository:** [unslothai/hyperlearn](https://github.com/unslothai/hyperlearn) **Primary Language:** Jupyter Notebook **GitHub Description:** 2-2000x faster ML algos, 50% less memory usage, works on all hardware - new and old. **GitHub Description (Excerpt):** <p align

### Overview: This project is an official implementation of the paper "Are Transformers Effective for Time Series Forecasting?" presented at AAAI 2023. It provides a comprehensive set of linear models for time series forecasting, including Linear, NLinear, and DLinear, along with a benchmark for long-term time series forecasting. ### Key Features: - **Benchmark for Long-Term Time Series Forecasting**: Provides a comprehensive benchmark for long-term time series forecasting. - **Support for Univariate and Multivariate Forecasting**: Handles both univariate and multivariate time series forecasting tasks. - **Visualization of Weights**: Allows visualization of model weights for better understanding. - **Support for Different Look-Back Window Sizes**: Scripts available for different look-back window sizes. - **Linear, NLinear, and DLinear Models**: Includes three linear models with different architectures and functionalities. - **Support for Autoformer, Informer, Pyraformer, and FEDformer Transformers**: Provides re-implementations of the results from the paper using these transformers. ### Why It Matters: This project is significant because it offers a robust and efficient alternative to traditional transformer-based models for time series forecasting, providing clear insights into why existing LTSF-Transformers perform poorly on the LTSF problem. The inclusion of linear models alongside transformers demonstrates the effectiveness of linear models in certain scenarios, making it a valuable resource for researchers and practitioners in the field of time series forecasting.

Overview: 雅意大模型是基于大规模中英文多领域指令数据训练的 LLaMA 2 & BLOOM 系列模型，由中科闻歌算法团队研发。它在百万级高质量领域数据上进行指令微调，覆盖媒体宣传、舆情分析、公共安全、金融风控、城市治理等五大领域，上百种自然语言指令任务。通过不断优化，雅意大模型提升了中文基础能力和领域分析能力，增加了多轮对话和部分插件功能，并通过用户反馈持续提升性能和安全性。 Key features: - 基于 LLaMA 2 & BLOOM 系列模型 - 多领域指令任务覆盖：媒体宣传、舆情分析、公共安全、金融风控、城市治理 - 优化的中文基础能力和领域分析能力 - 多轮对话和部分插件功能 - 用户反馈持续优化性能和安全性 Tech stack: Python Why it matters: 雅意大模型的开源为促进中文预训练大模型的社区发展，贡献了自己的一份力量，通过开源，与每一位合作伙伴共建雅意大模型生态。

Overview: Bolt is a fast algorithm for compressing vectors of real-valued data and performing mathematical operations on compressed representations, potentially reducing space and computation time by up to 10-200x. Key features: - Efficient vector compression - Fast matrix and vector operations - Compressible and lossy compression - Theoretical error bounds Tech stack: C++ Why it matters: Bolt offers significant space and computational savings for large collections of dense vectors, making it valuable for applications requiring efficient data compression and parallel processing. Repository: dblalock/bolt GitHub stars: 2523 GitHub description: 10x faster matrix and vector operations README excerpt: <p align="center"> <img src="https://github.com/dblalock/bolt/blob/master/assets/bolt.jpg?raw=true" alt="Bolt" width="611px" height="221px"/> <!-- <img src="https://github.com/dblalock/bolt/blob/master/assets/bolt.jpg?raw=true" alt="Bolt" width="685px" height="248px"/> --> </p> Bolt is an algorithm for compressing vectors of real-valued data and running mathematical operations directly on the compressed representations. If you have a large collection of mostly-dense vectors and can tolerate lossy compression, Bolt can probably save you 10-200x space and compute time. Bolt also has [theoretical guarantees](https://github.com/dblalock/bolt/blob/master

### Overview: This repository is a Semantic Segmentation Suite designed to simplify the implementation, training, and testing of various Semantic Segmentation models using TensorFlow. It includes training and testing modes, data augmentation, support for multiple state-of-the-art models, flexibility with any dataset, and comprehensive evaluation metrics. ### Key Features: - **Training and Testing Modes**: Facilitates easy setup for both training and testing of Semantic Segmentation models. - **Data Augmentation**: Enhances model robustness through various data augmentation techniques. - **State-of-the-Art Models**: Includes popular and advanced models like SegNet, ResNet, and others. - **Plug-and-Play**: Allows seamless integration with different models without modification. - **Versatile Data Support**: Compatible with any dataset, ensuring broad applicability. - **Evaluation Metrics**: Provides precision, recall, F1 score, average accuracy, per-class accuracy, and mean IoU for comprehensive performance assessment. ### Tech Stack: TensorFlow ### Why It Matters: This repository is essential for researchers and developers working on Semantic Segmentation projects, offering a comprehensive solution for training, testing, and evaluating various models efficiently. Its flexibility and ease of use make it a valuable tool for anyone looking to improve their Semantic Segmentation capabilities.