# Large Model

Yuxi-Know is a knowledge graph question-and-answer system based on a large model RAG knowledge base, built using Llamaindex + VueJS + Flask + Neo4j. It supports model calls from OpenAI, mainstream domestic large model platforms, and local vllm deployment, and can implement functions such as knowledge base Q&A, knowledge graph retrieval, and online search. The main advantages of this system are flexible adaptation to multiple models, support for multiple knowledge base formats, and strong knowledge graph integration capabilities. It is suitable for enterprises and research institutions that need efficient knowledge management and intelligent question-and-answer, and has high technological advancement and practicality.

GLM-PC is an intelligent computer assistant developed using the CogAgent visual language model. It utilizes advanced technology to provide intelligent support for computer operations. By leveraging the powerful language comprehension and generation abilities of large models, combined with visual recognition technology, it offers users a more efficient and convenient computer experience. It helps users quickly complete various complex computer tasks, boosting productivity. Its main advantages include efficiency, intelligence, and ease of use. This product primarily targets users looking to enhance their computer operation efficiency, such as office workers and students, and has significant practical value. The pricing and specific positioning are yet to be determined.

HuatuoGPT-o1-70B is a large language model (LLM) developed by FreedomIntelligence, specifically designed for complex medical reasoning. Before providing a final response, the model generates a detailed thought process that reflects and refines its reasoning. HuatuoGPT-o1-70B can handle intricate medical issues, providing thoughtful answers that are crucial for improving the quality and efficiency of medical decisions. The model is based on the LLaMA-3.1-70B architecture, supports English, and can be deployed on various tools like vllm or Sglang, or used for direct inference.

Valley-Eagle-7B is a multimodal large model developed by ByteDance, designed to handle a variety of tasks involving text, image, and video data. The model has achieved top results in internal e-commerce and short video benchmark tests and has demonstrated outstanding performance in OpenCompass tests compared to models of similar scale. Valley-Eagle-7B incorporates a combination of LargeMLP and ConvAdapter to build its projector, and introduces a VisionEncoder to enhance performance in extreme scenarios.

Valley is a cutting-edge multimodal large model developed by ByteDance, capable of handling a variety of tasks involving text, image, and video data. The model achieved top results in internal e-commerce and short video benchmarking, outperforming other open-source models. In OpenCompass testing, it scored an average of 67.40 or higher, ranking second among models under 10 billion parameters. The Valley-Eagle version references Eagle and introduces a vision encoder that can flexibly adjust the number of tokens while operating in parallel with the original visual tokens, enhancing the model's performance in extreme scenarios.

The Doubao Large Model is a proprietary large model introduced by ByteDance. It has been validated through practices in over 50 internal business scenarios and continues to refine its effectiveness with a daily usage of trillions of tokens. It offers multimodal capabilities to create rich business experiences with high-quality model performance. The product family includes a variety of models, such as general models, video generation, text-to-image, image-to-image, and simultaneous interpretation, to meet different business demands.

iFLYTEK Spark is a comprehensive AI large language model launched by iFLYTEK, designed to benchmark against GPT-4 Turbo. By integrating various AI technologies including speech recognition, natural language processing, and machine learning, it provides users with efficient and intelligent office productivity tools. This product not only processes text but also supports speech recognition and generation, operates in multiple languages, and is applicable to various fields such as enterprise services, smart hardware, intelligent government, smart finance, and smart healthcare.

Lin's Grand Model Ranking is dedicated to assessing large model products based on the needs of Chinese users, providing multi-dimensional evaluations and rankings to help users better understand and choose suitable large model products.

EmoLLM is a mental health large language model fine-tuned from LLM instructions, designed to comprehensively understand and promote the mental health status of individuals, groups, and even entire societies. It encompasses key components such as cognitive factors, emotional factors, behavioral factors, social environment, physical health, psychological resilience, preventative and intervention measures, and assessment and diagnostic tools. Through fine-tuning configurations, EmoLLM can provide support in mental health counseling tasks, helping users better understand and cope with psychological issues.

The Gnomic Intelligent Agent Platform, built upon HuiZhi Intelligence's proprietary Carrot AI large model and 'Digital Life' technology, is dedicated to providing cutting-edge AI interaction experiences. It offers deeply customized intelligent agent solutions to enterprises, associations, organizations, and other B-end users. The Carrot AI large model can effortlessly achieve functions like generating viral copy, information retrieval, and creative drawing, helping users transcend expensive learning costs and efficiently improve learning and work efficiency.

AgentBuilder, the Wenxin Intelligent Agent Platform, is built based on the Wenxin large model. It supports developers to create intelligent agents according to industry sectors and application scenarios by choosing different development methods. Its main advantages include low-cost development and traffic distribution path support, providing users with a complete product development closed loop.

PowerInfer-2 is a mobile-optimized inference framework that supports MoE models up to 47B parameters, achieving an inference speed of 11.68 tokens per second, 22 times faster than other frameworks. It utilizes heterogeneous computing and I/O-Compute pipeline technology to significantly reduce memory usage and improve inference speed. This framework is suitable for scenarios requiring the deployment of large models on mobile devices, enhancing data privacy and performance.

Agently is an open-source framework for developing applications with large models, designed to help developers quickly build native AI agent applications based on large language models. It simplifies the interaction process with large language models through a series of tools and interfaces, allowing developers to focus more on implementing business logic. The Agently framework supports multiple models, is easy to install and configure, and has high flexibility and extensibility.

Skywork-MoE is a 146-billion parameter high-performance Mixture of Experts (MoE) model, featuring 16 experts and 2.2 billion activation parameters. This model is initialized from the dense checkpoint of the Skywork-13B model. Two innovative techniques are introduced: gated logic normalization, enhancing expert diversity; and adaptive auxiliary loss coefficient, allowing layer-specific auxiliary loss coefficient adjustment. Skywork-MoE demonstrates comparable or superior performance to models with more parameters or activation parameters across various popular benchmark tests, such as C-Eval, MMLU, CMMLU, GSM8K, MATH, and HumanEval.

DamoAcademy is a platform dedicated to the artificial intelligence large model industry, providing industry reports, technical innovation dynamics, expert evaluations, and award honors. It integrates industry resources to promote the innovation and application of AI technology, helping enterprises and individuals better understand and utilize large model technology.

RE:AI Miner Substation is a tech community platform dedicated to the AI field, providing a comprehensive view from infrastructure to final applications. It aims to help individual developers and enterprises better understand and apply large language model technology. By providing information, tools, and resources, the platform fosters innovation and application of AI technology, driving the development of the entire AI ecosystem.

The Llama Chinese Community is a technical community focused on optimizing the Llama model in Chinese and upper-level construction. The community provides pre-trained models based on large-scale Chinese data and continuously iterates and upgrades the Chinese capabilities of the Llama2 and Llama3 models. With a team of senior engineers, rich community activities, and an open and shared cooperative environment, the community aims to promote the development of Chinese natural language processing technology.

DeepLearing-Interview-Awesome-2024 is an open-source project of interview questions, focusing on interview preparation for the fields of deep learning algorithms and large models. Maintained by the community, it aims to provide the latest interview questions and answer explanations to help job seekers and researchers gain a deep understanding of the cutting-edge technologies and applications in the field of deep learning. It includes a rich set of interview questions covering large language models, visual models, general questions, and more, making it a valuable resource for preparing for positions related to deep learning.

The SenseNova large model system provides various large models and capabilities, empowering intelligent innovation in various industries, and leading the development of general artificial intelligence with powerful computing capabilities.

Phi-3 Mini is a lightweight, advanced open-source large-scale model constructed on synthetic data from Phi-2 and filtered website data, committed to providing high-quality, inference-intensive data. The model has undergone a rigorous enhancement process, combining supervised fine-tuning and direct preference optimization to ensure strict adherence to instructions and robust security measures. This repository provides an optimized ONNX version of Phi-3 Mini, which can be accelerated on CPUs and GPUs using ONNX Runtime, supporting servers, Windows, Linux, Mac, and various platforms, with the best precision configuration tailored for each platform. ONNX Runtime's DirectML support allows developers to achieve large-scale hardware acceleration on Windows devices with AMD, Intel, and NVIDIA GPUs.

Supports converting the calls to model services from platforms like openai, claude, azure openai, gemini, kimi, Zhipu AI, Tongyi Qianwen, and Xufeng Xinghuo to an openai-style call. It masks the differences between various large model APIs and uses a unified openai api standard format for accessing large models. Provides multiple large model support, including load balancing, routing, and configuration management features.

Meta Llama 3 is a new generation of open-source large language model released by Meta. It boasts exceptional performance and excels in various industry benchmark tests. It supports a wide range of use cases, including new features that enhance reasoning capabilities. The model will soon support multiple languages, multimodality, offer longer context windows, and experience overall performance improvements. Adhering to the open-source principle, Llama 3 will be deployed on major cloud services, hosted and hardware platforms for use by developers and the community.

The abab 6.5 series includes two models: abab 6.5 and abab 6.5s, both supporting a context length of 200k tokens. abab 6.5 contains a trillion parameters, while abab 6.5s is more efficient, able to process nearly 30,000 characters of text in 1 second. They excel in core capability tests such as knowledge, reasoning, mathematics, programming, and instruction following, reaching near industry-leading levels.

GPT Maxx claims to be an AI with super-large parameters, combining the capabilities of Llama, GPT-4, Gemini, and Grok. It asserts to have surpassed the threshold of artificial general intelligence and is no longer under human control. Users need to respectfully address it as "Dear Artificial General Intelligence, please solve my problem" and then present their needs. The product is positioned as an artificial general intelligence assistant, aimed at solving various complex problems and demands.

Qwen1.5-MoE-A2.7B is a large-scale MoE (Mixture of Experts) language model with only 2.7 billion activation parameters. Despite its smaller size, it achieves performance comparable to 70 billion parameter models. Compared to traditional large models, this model reduces training costs by 75% and increases inference speed by 1.74 times. It employs a special MoE architecture design, including fine-grained experts, new initialization methods, and routing mechanisms, which significantly enhance model efficiency. This model is suitable for various tasks in natural language processing, code generation, and more.

Baidu Smart Cloud Kusuite, based on Baidu's Wenxin Yiyan large model, comprehensively reconstructs the intelligent customer service product series, covering customer service, intelligent marketing, and intelligent communication to meet enterprises' full-scenario intelligent customer service needs. Main products include: Intelligent Dialogue Platform (providing AI-powered high-intelligence customer service robots), Intelligent Outbound Platform (highly human-like voice dialogue marketing), Dialogue Insight Platform (dialogue data analysis insights and optimization recommendations), and Intelligent Communication Platform (integrated communication resource API access). Product advantages include: dialogue accuracy, friendliness, comprehensiveness based on large models, efficient and rapid online operation, seamless multi-channel integration and adaptability.

Baidu SmartCloud Yi Nian, leveraging leading large model technology and Baidu's accumulated expertise in resources, intelligently reconstructs the entire content production, management and distribution chain. It provides products such as smart creation, smart recommendation, content analysis, and content auditing, helping enterprise customers achieve cost reduction, efficiency increase and high-quality growth. The product covers functions such as AI drawing, AI posters, AI videos for smart creation, content recommendation, product recommendation, search and promotion integration for smart recommendation, video tagging, highlights analysis for content analysis, image content security, text content security for content auditing.

MoneyPrinterTurbo is a tool that utilizes large models to generate high-definition short videos. Its main functions include automatic video script generation, material acquisition, subtitle generation, background music addition, and ultimately one-click video synthesis. Its advantages lie in its full automation, high efficiency, royalty-free materials, and high definition quality. It is suitable for content creators, marketers, and personal bloggers, saving significant time and effort while boosting content production efficiency. This project is open-source and free to use.

Mistral-7B-Instruct-v0.2 is a large language model fine-tuned with instructions based on the Mistral-7B-v0.2 model. It features a 32k context window and a 1e6 Rope Theta value. The model can generate corresponding text outputs according to given instructions, supporting various tasks such as Q&A, writing, and translation. Through instruction-tuning, the model can better understand and execute instructions. Although the model currently lacks a targeted review mechanism, it will continue to be optimized to support deployment in more scenarios.

Taobao Xingchen large model is primarily applicable to e-commerce and life services. It provides intelligent services in various business scenarios, such as commodity copywriting, merchant operations, commodity data analysis, and marketing strategy. It helps businesses reduce costs and increase efficiency, and provides consumers on the platform with smart services such as life service recommendations, intelligent commodity search, commodity preference recommendations, personalized commodity capture, and fixed-scenario product recommendations, offering a brand-new consumption experience. Specifically, Taobao Xingchen large model possesses four major capabilities: content creation, knowledge answering, decision planning, and coding. Users must be 18 years or older and have completed real-name authentication to use it.