Coding & Development

Pligence is an AI-driven cybersecurity and privacy company that provides comprehensive solutions for mobile security, threat intelligence, and governance, risk, and compliance (GRC) management. Their offerings include Privacy Defender, a mobile privacy and security suite for consumers, Pligence Connect for enterprise mobility and security with MDM capabilities, and a Threat Intelligence and Assessment Platform for identifying and disrupting cyber threats. Additionally, Pligence offers a GRC Lens SAAS platform for risk management and compliance. The company leverages machine learning to proactively identify and neutralize evolving cyber threats, ensuring the security and resilience of digital environments for mobile users, IoTs, enterprises, financial institutions, government organizations, and service providers.

NeuralTalk is a Python+numpy project designed for developing Multimodal Recurrent Neural Networks capable of describing images with sentences. This open-source tool, though now deprecated in favor of NeuralTalk2, remains valuable for educational purposes in image captioning and natural language processing research. It implements models like those proposed by Vinyals et al. (Google CNN + LSTM) and Karpathy and Fei-Fei (Stanford CNN + RNN), allowing users to train models on datasets such as Flickr8K, Flickr30K, and MSCOCO. The project supports both training and prediction stages, with utilities for visualizing results and evaluating performance using BLEU scores. Users can also adapt the system for their own datasets, requiring feature extraction using tools like VGG network from Caffe.

nn4nlp-code is a comprehensive GitHub repository offering code examples specifically designed for the 2017 edition of CMU CS 11-747 Neural Networks for NLP course. Developed by Graham Neubig, Daniel Clothiaux, Zhengzhong Liu, and Xuezhe Ma, this resource provides practical, hands-on implementations of various neural network models pertinent to natural language processing. It serves as an invaluable learning tool for students and researchers looking to understand and apply NLP concepts through code. The repository is open-source, making it accessible for educational purposes, experimentation, and further development in the field of AI and NLP.

Monk_Object_Detection is a comprehensive, low-code repository designed to simplify object detection pipelines for computer vision engineers and AI developers. It acts as a unified wrapper over major deep learning frameworks like PyTorch, TensorFlow, and GluonCV, making it easier to install and use state-of-the-art algorithms. The toolkit addresses common challenges such as installing different deep learning pipelines, setting up algorithms for custom data, and tuning hyperparameters. It supports various applications including object detection, image segmentation, face detection, pose estimation, and activity recognition, with easy deployment options and hands-on tutorials.

Research is a GitHub repository by PaddlePaddle dedicated to novel deep learning research works. It features a comprehensive collection of top conference papers and competition-winning models, covering key areas such as Computer Vision (CV), Natural Language Processing (NLP), Knowledge Graph (KG), and Spatial-Temporal Data-Mining (STDM). The repository offers detailed descriptions, paper links, and implementations for various tasks within these domains, making it a valuable resource for researchers and developers working with PaddlePaddle. It is open-source and freely accessible, encouraging collaboration and advancement in deep learning.

Simple_Reinforcement_Learning is an open-source toolkit designed for the development and testing of reinforcement learning algorithms. It provides a structured environment for implementing various RL techniques, including stateless problems, Markov Decision Processes, dynamic programming, temporal difference algorithms, DynaQ, DQN, policy gradient, Actor-Critic, PPO, DDPG, SAC, imitation learning, offline learning, MPC, MBPO, goal-oriented reinforcement learning, and multi-agent systems. The toolkit is built to run on Python 3.9, PyTorch 1.12.1, and Gym 0.26.2, making it compatible with widely used machine learning libraries and environments. It serves as a valuable resource for researchers and engineers looking to explore and experiment with different reinforcement learning paradigms.

SUSI.AI server backend is an open-source Artificial Intelligence server designed to power personal assistants. It facilitates both chat and voice interactions, allowing users to perform a wide range of actions such as music playback, managing to-do lists, setting alarms, streaming podcasts, and accessing real-time information like weather and traffic. The server's core intelligence and personality drive the SUSI.AI platform, with additional functionalities extensible via external APIs. It is highly customizable, allowing users to deploy it on various platforms like Heroku, Google Cloud, AWS, Azure, and Digital Ocean, and even develop custom skills for tailored AI experiences. The project emphasizes community contributions for bug fixes and feature enhancements.

Labophase is a platform designed to offer users access to a diverse range of AI tools, encompassing capabilities like prompt-based image generation and advanced text models. The platform supports both free and premium user tiers, ensuring accessibility for a broad audience. Labophase is committed to continuous improvement, regularly integrating and updating models such as Claude Opus and GPT-4o. Its primary goal is to deliver significant value through a user-friendly interface combined with extensive AI functionalities, making advanced AI accessible for various creative and analytical tasks.

Awesome-AI-GPTs is an open-source directory designed to be a comprehensive resource for all things related to OpenAI GPTs. It features a curated collection of various GPTs, useful prompts, and plugins, along with articles and guides on how to use them effectively. The project also covers topics like GPTs security, prompt protection, and custom plugin installation, making it a valuable hub for both beginners and experienced users looking to explore and leverage the power of AI. It aims to empower users by providing access to AI resources and interesting applications, fostering a community for collaboration and knowledge sharing.

Auto-GPT-ZH is the Chinese version of Auto-GPT, an open-source platform designed for building, deploying, and managing continuous AI agents capable of automating complex workflows. It offers a comprehensive solution for AI enthusiasts and developers looking to leverage autonomous AI. The platform provides a full visual interface, a drag-and-drop agent builder, and cloud deployment options, making it accessible for both technical and non-technical users. While the classic version is no longer maintained, the platform version and the Forge toolkit for developers offer robust capabilities for creating custom AI agents, integrating tools and APIs, managing context memory, and performing web browsing and file operations. It supports various applications, including automated content generation for videos and social media.

Baichuan-13B is a 13-billion parameter open-source large language model developed by Baichuan Intelligent Technology. Building upon Baichuan-7B, it expands its parameter count and has been trained on 1.4 trillion tokens of high-quality data, surpassing LLaMA-13B in training data volume. The model supports both Chinese and English, utilizes ALiBi positional encoding, and has a context window length of 4096. It is available in both a pre-trained base version (Baichuan-13B-Base) and an aligned chat version (Baichuan-13B-Chat) with strong conversational capabilities. For efficient deployment, Baichuan-13B also provides int8 and int4 quantized versions, significantly reducing hardware requirements without substantial performance loss, making it deployable on consumer-grade GPUs like Nvidia 3090. It is free for academic research and available for free commercial use upon application.

ChatGLM-Finetuning is a comprehensive open-source project designed for fine-tuning ChatGLM-6B, ChatGLM2-6B, and ChatGLM3-6B models. It offers a variety of fine-tuning methods, including Freeze, Lora, P-tuning, and full parameter fine-tuning, allowing users to select the most suitable approach for their specific needs. The tool supports both single-card and multi-card training environments, making it adaptable to different hardware setups. It is particularly useful for tasks such as information extraction, generation, and classification. The project emphasizes maintaining model performance without severe catastrophic forgetting, even after fine-tuning. It also provides detailed instructions and code examples for implementing each fine-tuning method, along with memory usage benchmarks for different configurations, aiding developers in optimizing their training processes.

Bert-Chinese-Text-Classification-Pytorch is an open-source project designed for Chinese text classification, leveraging powerful pre-trained language models like Bert and ERNIE. Implemented in PyTorch, this tool offers an out-of-the-box solution for developers and researchers working with Chinese language data. It includes pre-trained models and a dataset of 200,000 Chinese news titles across 10 categories, making it ready for immediate use. The project also explores the integration of Bert with other neural network architectures such as CNN, RNN, RCNN, and DPCNN for comparative analysis of classification performance. It provides clear instructions for setting up the environment, using custom datasets, and running training and testing scripts.

claude-skills is an open-source GitHub repository that provides a collection of skills designed to extend the capabilities of Anthropic's Claude AI model within its code interpreter environment. The repository serves as a resource for developers and users interested in understanding and implementing custom functionalities for Claude. It includes archived files that were previously part of Claude's `/mnt/skills` folder, offering a historical perspective and practical examples of how skills are structured. Users can leverage this repository to learn about skill creation, adapt existing examples, or contribute to the development of new tools for Claude. The project also references an official Anthropic repository for current skill development.

BlueLM is an open-source large language model developed by vivo AI Lab, available in 7B base and chat versions, including models supporting 32K long-text contexts. Trained on a massive 2.6 trillion token corpus covering Chinese, English, and some Japanese/Korean data, BlueLM-7B-Chat demonstrates competitive performance on benchmarks like C-Eval and CMMLU. The project provides resources for inference deployment, including command-line and web demos, and supports OpenAI-compatible API for integration. It also offers tools for model fine-tuning and quantized versions (4-bit) for reduced GPU memory usage, making it accessible for various research and development scenarios.

Awesome-ChatGPT-Prompts-CN is an open-source GitHub repository offering a comprehensive guide and collection of prompts for ChatGPT, primarily in Chinese. It aims to help users effectively interact with and leverage the capabilities of ChatGPT for various tasks. The repository includes examples for different roles, such as acting as a Linux terminal, English translator, interviewer, JavaScript console, Excel sheet, and more. It also provides guidance on registration and usage, addressing common issues like country restrictions. The project encourages community contributions and offers resources for further learning and development with OpenAI and ChatGPT.

Awesome-Chinese-LLM is a comprehensive open-source repository dedicated to Chinese large language models (LLMs). The collection prioritizes models that are smaller in scale, suitable for private deployment, and have lower training costs, making them accessible to a wider range of users. It encompasses a variety of resources, including foundational base models like ChatGLM, LLaMA, Baichuan, and Qwen, as well as models fine-tuned for vertical domains such as healthcare, law, finance, and education. Beyond models, the repository also provides valuable datasets for pre-training, SFT, and preference alignment, along with tutorials covering LLM basics, prompt engineering, application development, and practical implementation. This makes it an invaluable resource for researchers, developers, and practitioners working with Chinese LLMs.

Fairlearn is a Python package designed to empower developers of artificial intelligence (AI) systems to assess and mitigate unfairness in their machine learning models. It offers a comprehensive suite of tools, including metrics for identifying groups negatively impacted by a model and algorithms for mitigating unfairness across various AI tasks and fairness definitions. The package focuses on addressing allocation harms (e.g., in hiring or lending) and quality-of-service harms, following a group fairness approach. Fairlearn enables users to compare multiple models based on fairness and accuracy metrics, providing a robust framework for responsible AI development. It is open-source and includes Jupyter notebooks for practical usage examples.

DeepMatcher is a Python package designed for entity and text matching tasks using deep learning. It offers built-in neural networks and essential utilities, enabling users to train and apply advanced deep learning models for entity matching with less than 10 lines of code. The package supports data processing for training, validation, and test CSV data, model definition with customizable neural network architectures, and model training and application. Its modular design allows for easy customization of subcomponents, making it flexible for various matching tasks beyond traditional entity matching, such as question answering. DeepMatcher is ideal for researchers and developers looking to leverage deep learning for data integration and record linkage.

EduChat is an open-source educational chat model developed by ICALK at East China Normal University, designed to support personalized learning and holistic development. It integrates diverse educational data with methods like instruction fine-tuning and value alignment to offer rich functionalities such as automatic question generation, homework grading, emotional support, and course tutoring. The project has evolved through several versions, culminating in EduChat-R1, which focuses on "Thinking before teaching" to provide intelligent educational solutions. It also includes specialized products like MindCare@EduChat for psychological assessment, Shell@EduChat for value alignment, and AiBoard@EduChat as an AI teaching assistant, catering to the needs of teachers, students, and parents.

deep-learning-resources is an open-source GitHub repository that curates a comprehensive collection of deep learning materials. It is designed to guide learners from foundational concepts to advanced topics, with content continuously updated. The repository includes interactive playgrounds for hands-on experience, a curated list of online courses from leading institutions like Stanford and MIT, practical tools such as Colaboratory and TensorBoard, and a selection of high-quality articles and classic papers. It serves as a valuable hub for anyone looking to start or deepen their understanding of deep learning, providing structured learning paths and practical applications.

Firefly is a comprehensive open-source project designed for training large language models, offering support for pre-training, instruction fine-tuning, and DPO (Direct Preference Optimization). It is compatible with numerous mainstream models such as Qwen2.5, Qwen2, Yi1.5, Phi-3, Llama3, Gemma, MiniCPM, and many others. The platform facilitates full parameter training, as well as efficient training methods like LoRA and QLoRA, making it accessible even with limited computational resources. Firefly also integrates with Unsloth for accelerated training and reduced VRAM usage, and provides curated instruction fine-tuning datasets. It offers pre-trained Firefly series models and has demonstrated effectiveness on the Open LLM Leaderboard.

electerm is a versatile, open-source terminal client designed for developers and system administrators, supporting a wide array of connection types including SSH, SFTP, FTP, Telnet, serial port, RDP, VNC, and Spice. Available across Linux, macOS, and Windows, it offers features like global hotkeys, multi-language support, and the ability to directly edit small remote files. A key differentiator is its AI assistant integration, supporting DeepSeek, OpenAI, and other AI APIs, to provide command suggestions, assist with script writing, and explain terminal content. It also includes a Model Context Protocol (MCP) widget for AI assistants and external tools, enhancing productivity for technical users.

food-101-keras is an open-source deep learning project hosted on GitHub, designed for food classification using Keras and Tensorflow. It leverages Convolutional Neural Networks (CNNs) to identify 101 different food classes from the Food-101 dataset. The project demonstrates how to fine-tune a pre-trained Google InceptionV3 model, achieving high accuracy in food recognition. It includes detailed steps for data loading, preprocessing, image augmentation, model training, and evaluation. The repository also provides insights into handling large datasets and exporting models for mobile applications, making it a valuable resource for machine learning practitioners and researchers interested in computer vision and food recognition.