Model Training

无问芯穹是由Infinigence推出的企业级AI大模型服务平台，专注于AI 2.0时代的解决方案，提供从算力、模型到应用的一站式服务，支持多种主流模型和国产芯片的高效部署。

AgentGPT是一款开源项目，允许用户在浏览器中轻松创建、配置和部署自主AI代理，无需编程技能，适用于个人助理、内容创作和软件开发等多种场景。

Coze is an all-in-one AI development platform launched by ByteDance, enabling users to create AI robots without coding in 30 seconds. It integrates over 60 plugins covering areas like news reading and travel planning, making it suitable for users without programming experience to quickly create and deploy AI applications.

Obviously AI is a no-code data science platform designed to help businesses and individuals rapidly build and deploy machine learning models without writing code. Its core features include rapid model building, one-click deployment, automated model monitoring, and real-time API integration, significantly lowering the barrier to AI adoption.

ColossalChat, developed by Colossal-AI, is a free open-source chatbot based on the LLaMA pre-trained model, utilizing a complete RLHF mechanism to offer robust bilingual conversation capabilities in English and Chinese, ideal for developers and AI enthusiasts.

Scale AI is a San Francisco-based technology company specializing in providing high-quality data annotation and data management services for machine learning and artificial intelligence applications. Its core products include Scale Data Engine and Scale GenAI Platform, designed to help enterprises optimize AI model performance and accelerate the development and deployment of AI applications.

SenseChat, developed by SenseTime, is an AI chat assistant based on the self-developed 'SenseNova' model, offering multi-turn conversations, logical reasoning, language correction, content creation, and more, catering to the needs of Gen Z and internet users for intelligence, convenience, and efficiency.

OpenBMB, jointly initiated by Tsinghua University and the Beijing Academy of Artificial Intelligence, is dedicated to building a large-scale pre-trained language model library and related tools, lowering the threshold for large model usage, accelerating their training, fine-tuning, and inference processes, and promoting the standardization, popularization, and practical application of large models.

Jan is a free and open-source AI chat tool that supports running over 20 large models, including LLaMa, Mistral, and Phi-2, locally on Windows, Mac, and Linux systems. It offers a user-friendly chat interface and a built-in API server, catering to developers and AI enthusiasts.

Meitu's self-developed AI visual model, MiracleVision, integrates profound aesthetic understanding with advanced AI technology, offering text-to-image and image-to-image functionalities. It is widely applied in e-commerce, advertising, gaming, animation, and film industries, significantly enhancing creative efficiency.

Label Studio, developed by Human Signal (formerly Heartex), is a free open-source data labeling tool supporting various data types, offering flexible configurations and machine learning assistance, enabling developers to efficiently fine-tune large language models, prepare training data, or validate AI models.

JAX是由Google开发的Python库，旨在提供高效的数值计算和机器学习功能，支持自动微分、即时编译和自动向量化等特性。

MLX is an open-source framework developed by Apple's machine learning research team, designed specifically for Apple Silicon chips. It offers a NumPy-like API, supports automatic differentiation and lazy evaluation, aiming to assist developers in efficiently developing, training, and deploying machine learning models on Apple's M-series chips.

Tencent Hunyuan Model is a self-developed large-scale language model by Tencent, featuring excellent Chinese comprehension and creation capabilities, logical reasoning, and reliable task execution, widely applied in document creation, meeting assistance, advertisement generation, and more.

Jinsha Aggregation is a platform dedicated to providing professional API interface services for AI practitioners, covering various functions such as AI Q&A, AI painting, and AI video, aiming to enhance developers' work efficiency and product quality.

Cohere is a Canadian technology company founded in 2019, specializing in providing large language models for enterprises. Its platform offers services such as multilingual embeddings, semantic search, text generation, and classification, enabling businesses to rapidly deploy conversational AI, generative search engines, and text summarization applications. Cohere emphasizes data security and privacy protection, supporting various deployment options to meet diverse enterprise needs.

Hugging Face has released BLOOM, a 176 billion parameter multilingual large language model supporting 46 natural languages and 13 programming languages, aiming to advance open research and applications in artificial intelligence.

Meta has released the new Segment Anything Model 2 (SAM 2), a unified open-source AI model capable of real-time object segmentation in images and videos. SAM 2 marks a significant breakthrough in image and video segmentation, supporting real-time processing and excelling in various visual tasks.

Mobvoi's 'Sequence Monkey' is a large language model with multimodal generation capabilities, covering six dimensions: knowledge, dialogue, mathematics, logic, reasoning, and planning. It supports various tasks such as text, image, 3D content, voice generation, and recognition, aiming to provide users with a smooth and convenient interactive experience.

AutoGPT is an open-source AI agent designed to automate digital workflows, assisting users in efficiently completing tasks and focusing on innovative and impactful work.

DeepSpeed is an open-source deep learning optimization library by Microsoft, designed to reduce computational resources and memory usage, offering efficient distributed model training, especially suitable for large-scale language model training.

BigModel是智谱AI专为开发者设计的大模型开发平台，提供模型能力、开发资源、知识库与应用体验中心等工具，帮助开发者充分应用GLM大模型的技术能力，打造多种应用。近百万开发者选择调用大模型，注册即送2500万Tokens资源包。

Lobe is a no-code development platform provided by Microsoft, designed to help users easily create and train machine learning models. With a simplified three-step process, users can train and deploy models without writing any code.

Huixing Cloud (formerly Houdeyun) is a professional AI computing power platform offering high-performance, convenient, and secure GPU resources to accelerate AI research and application processes.

Keras is an open-source deep learning API written in Python, designed to simplify the construction and training of neural networks. Its user-friendly design, modular structure, and multi-backend support enable developers to efficiently develop and deploy deep learning models.

PaddlePaddle is an open-source deep learning platform developed by Baidu, integrating core frameworks, model libraries, development kits, and tool components. It is widely used in industries such as manufacturing, agriculture, and enterprise services, serving over 10.7 million developers and 235,000 enterprises.

CuspAI is an AI-driven material search engine launched by the University of Cambridge. Utilizing generative AI technology, it rapidly generates and evaluates potential molecular structures based on user-defined material properties, significantly accelerating the discovery of new materials. In June 2024, CuspAI completed a $30 million seed funding round, focusing on developing new materials for carbon capture and storage to support global carbon neutrality goals.

Modelers Community, co-developed by Tianyi Cloud and Huawei, offers diverse services such as model hosting, dataset management, and online experiences, empowering AI developers for efficient innovation.

Qdrant is an open-source vector database and similarity search engine developed in Rust, supporting efficient storage and retrieval of multi-modal data, widely used in personalized recommendations, image recognition, and real-time data analysis.

Apache MXNet is an open-source deep learning framework supporting multiple programming languages and devices, offering flexible programming models and efficient distributed training capabilities.