AI large model base computing power center and mounting device

Abstract

The invention provides an AI large model base computing power center and installation device, and the device comprises a dynamic environment adjustment and control module which is used for collecting and adjusting environment data in real time; the power management module is used for providing electric power and reasonably distributing the electric power to each device; the training and pushing module is used for integrating hardware equipment, supporting large model training and reasoning at the same time, and realizing two-way data transmission with the intelligent middle platform module; the intelligent middle platform module is used for providing reuse capabilities of data management, AI algorithms, business components and the like, supporting rapid development and collaboration of the AI education module, and guaranteeing safety and data management of the AI education module at the same time; and the AI education module is used for editing digital teaching materials, assisting teachers in teaching innovation, generating personalized learning paths based on the data and use data provided by the intelligent middle table module and evaluating use behaviors of students and teachers, fusion of an AI large model and the computing power center is realized, and the multifunctional integrated computing power center is obtained.

Description

Technical Field

[0001] This invention belongs to the field of AI large-scale model technology, and specifically relates to an AI large-scale model base computing center and installation device. Background Technology

[0002] The AI ​​Large Model Base Computing Center is an integrated support platform combining computing power, algorithms, and data, specifically designed for AI large models with hundreds of billions to trillions of parameters. Unlike traditional general-purpose computing centers, its core positioning is to solve four major pain points of large models: long training cycles, high inference latency, difficulty in resource coordination, and high energy consumption costs. It provides "high computing power, low latency, high reliability, and green" base-level support for the research, iteration, and industrial application of large models. It is a core infrastructure for the large-scale application of AI technology in the digital economy era. As the parameter scale of AI large models breaks through from tens of billions to trillions (such as GPT-4, Wenxin Yiyan 4.0, etc.), the demand for computing power is growing exponentially.

[0003] As large-scale models develop and their applications in the education field gradually increase, there is an urgent need for an integrated and high-efficiency AI large-scale model base computing center. Summary of the Invention

[0004] Purpose of the invention: To address the problems existing in the above-mentioned background technology, the present invention provides an AI large model base computing power center, including: a dynamic environment adjustment and control module, used to collect and adjust environmental data in real time; The power management module is used to provide power and distribute it appropriately to various devices; The training and inference module is used to integrate hardware devices, support large model training and inference, and enable bidirectional data transmission with the intelligent middleware module. The intelligent middleware module provides reusable capabilities for data management, AI algorithms, and business components, supporting the rapid development and collaboration of the AI ​​education module while ensuring the security and data governance of the AI ​​education module. The AI ​​education module is used to edit digital teaching materials, assist teachers in teaching innovation, generate personalized learning paths based on data provided by the intelligent platform module and usage data, and evaluate the usage behavior of students and teachers.

[0005] The dynamic environment control module monitors and regulates the environment of the entire computing center to ensure the working environment of each module and avoid affecting work efficiency. By centrally installing the training and push module, intelligent middleware module, and AI education module, and realizing a variety of functions beneficial to the education field based on the AI ​​big model, the efficiency of the computing center is improved.

[0006] Preferably, the power management module includes a mains power supply, a power distribution unit, and a UPS power supply. The power distribution unit is connected to the UPS power supply and the mains power supply. The power distribution unit distributes the power from the UPS power supply and the mains power supply according to the power consumption, so that the entire computing center can achieve energy saving.

[0007] Preferably, the dynamic environment control module includes an intelligent environmental monitoring system and a precision air conditioner. The intelligent environmental monitoring system is used to collect environmental data from the computing center in real time and control the temperature of the precision air conditioning. Precision air conditioning is used to provide a constant temperature and humidity operating environment for the computing center, ensuring that the operating environment of the computing center is at its optimal level.

[0008] Preferably, the training and push module includes an integrated training and push machine, a switch, and a pneumatic CDU. The training and inference integrated machine is used for large-scale model training and inference. The switch is used to ensure signal transmission between the training and promotion all-in-one machine, the intelligent middleware module, and the AI ​​education module; The liquid cooling CDU provides liquid cooling for the training and push integrated machine, enabling the use of large AI models in the computing center and improving the efficiency of the computing center.

[0009] Preferably, the AI ​​education module includes a digital textbook development system, a digital teacher innovation system, an AI learning system, an AI evaluation system, an AI assistant management system, and an AI training assistance system. The digital textbook development system is used for the digital creation of textbooks and can generate digital textbooks adapted to different teaching scenarios; The Digital Teacher Innovation System is designed to assist teachers in innovative teaching methods, such as intelligent lesson preparation, resource integration, and interactive courseware creation. An AI learning system is used to receive student data and generate personalized learning paths based on that data. AI evaluation system is used to receive multimodal data analysis to assess teachers' teaching effectiveness or students' learning outcomes; AI-assisted management system for intelligent class scheduling, attendance statistics, asset management, and logistics dispatch; The AI-assisted training system provides simulated training and personalized training programs for skills training scenarios for teachers or students, and offers real-time feedback. This makes the entire computing center a multi-functional integrated system for both teachers and students, improving the efficiency of the entire process and enabling the computing center to achieve multi-functional integration.

[0010] Preferably, the intelligent middleware module is an intelligent middleware system, which ensures the data source and integrates technologies such as cloud computing, big data, and AI.

[0011] Preferably, the multimodal data in the AI ​​evaluation system includes data from classroom interactions, assignments, and exams; skills data collected from students can also be collected from teachers.

[0012] Preferably, the intelligent environmental monitoring system includes a dynamic environmental monitoring system and a dynamic environmental management system. The dynamic environment monitoring system is used to collect environmental data of the computing center in real time. The dynamic environment management system receives the data collected by the dynamic environment monitoring system and controls the precision air conditioning according to the data to manage the environment of the computing center, ensure the stability of the working environment of the computing center, and avoid overheating that causes premature aging of the equipment in the computing center.

[0013] Preferably, an AI large-scale model base computing center installation device includes: Multiple cabinets connected in sequence are used to install dynamic environment control module, power management module, training and promotion module, intelligent middleware module and AI education module; The display screens are installed on every two cabinets to show the working status of the dynamic environment control module, power management module, training and push module, intelligent middleware module, and AI education module, ensuring that the working status and content of the entire computing center can be seen intuitively.

[0014] Compared with existing technical solutions, the technical solution adopted in this invention has the following advantages: 1. The dynamic environment adjustment and control module monitors and detects the environment of the entire computing center, and adjusts the environment of the computing center to ensure the working environment of each module of the computing center and avoid affecting work efficiency; 2. By integrating the training and promotion module, the intelligent middleware module, and the AI ​​education module together, and based on the AI ​​big model, a variety of functions beneficial to the education field are realized, thereby improving the efficiency of the computing center; 3. Based on a large AI model, it integrates functions from multiple educational fields to help improve teaching efficiency. Attached Figure Description

[0015] Figure 1 A plan view of a computing center and installation device for a large AI model base; Figure 2 This is a three-dimensional structural diagram of a computing center and installation device for a large AI model base.

[0016] Figure label: 1. Cabinet; 2. Display screen. Detailed Implementation Example

[0017] This invention provides an AI large-scale model base computing center for teaching scenarios, such as... Figures 1-2As shown, it includes: a dynamic environment adjustment and control module, used to collect and adjust environmental data in real time; The power management module is used to provide power and distribute it appropriately to various devices; The training and inference module is used to integrate hardware devices, support large model training and inference, and enable bidirectional data transmission with the intelligent middleware module. The intelligent middleware module provides reusable capabilities for data management, AI algorithms, and business components, supporting the rapid development and collaboration of the AI ​​education module while ensuring the security and data governance of the AI ​​education module. The AI ​​education module is used to edit digital teaching materials, assist teachers in teaching innovation, generate personalized learning paths based on data provided by the intelligent platform module and usage data, and evaluate the usage behavior of students and teachers.

[0018] The dynamic environment control module maintains a suitable temperature throughout the computing center, slowing down its aging process and reducing the possibility of internal hardware damage. The power management module manages and distributes power throughout the entire computing center. Two-way data transmission between the training and push modules and the intelligent platform module enables pre-training and fine-tuning of large AI models, facilitating their deployment and use. It also stores and transmits relevant data from the AI ​​education module, which comprises multiple systems and implements various functions within the education field. This makes the entire computing center multifunctional, integrated, and energy-efficient. The power management module includes mains power, power distribution unit and UPS power. Under normal circumstances, the computing center is powered by mains power, but when the mains power is interrupted, it seamlessly switches to UPS power to ensure the normal operation of the computing center and avoid data loss due to sudden power outages.

[0019] The training and inference module includes a training and inference all-in-one machine, a switch, and a liquid cooling CDU. The training and inference all-in-one machine is a dedicated device for integrating hardware such as GPUs and CPUs, and supports large-scale model training and inference. Large-scale model training includes model pre-training and fine-tuning, while large-scale model inference involves deploying large models according to the services to be provided. The switch is responsible for network connection and data transmission of devices within the computing center, ensuring high-speed communication between the training and inference all-in-one machine and the AI ​​education module. The liquid cooling CDU provides liquid cooling for the training and inference all-in-one machine, removing heat from the device through liquid circulation.

[0020] The dynamic environment control module includes an intelligent environmental monitoring system and a precision air conditioner. The intelligent environmental monitoring system is used to collect environmental data of the computing center in real time and control the temperature of the precision air conditioner. In this embodiment, existing products are used, which are well known to those skilled in the art. Precision air conditioning is used to provide a constant temperature and humidity operating environment for the computing center, ensuring that the operating environment of the computing center is at its optimal level.

[0021] The intelligent environmental monitoring system includes a dynamic environmental monitoring system and a dynamic environmental management system. The dynamic environment monitoring system is used to collect environmental data of the computing center in real time. The dynamic environment management system receives the data collected by the dynamic environment monitoring system and controls the precision air conditioning according to the data to manage the environment of the computing center, ensure the stability of the working environment of the computing center, and avoid overheating that causes premature aging of the equipment in the computing center.

[0022] The AI ​​education module includes a digital textbook development system, a digital teacher innovation system, an AI learning system, an AI evaluation system, an AI assistant management system, and an AI training assistance system. The digital textbook development system is used for the digital creation of textbooks and can generate digital textbooks adapted to different teaching scenarios. Based on the needs of different scenarios, it processes the textbooks through an intelligent middleware system and training and promotion modules to form digital textbooks that meet the requirements.

[0023] The Digital Teacher Innovation System, based on data and large-scale models from the intelligent platform system and training and promotion modules, assists teachers in innovative teaching methods, such as intelligent lesson preparation, resource integration, and interactive courseware creation. The AI ​​learning system is used to receive student data, generate personalized learning paths based on the student data, recommend suitable learning resources, automatically answer questions such as breaking down problem-solving steps, and provide real-time feedback on learning progress to improve learning efficiency and interest. An AI-powered evaluation system is used to receive multimodal data for analysis of teachers' teaching effectiveness or students' learning outcomes. Specifically, it includes the following functions: For teachers: Quantify classroom behaviors, such as the percentage of teacher-student interaction, identify teaching weaknesses and provide improvement suggestions; For students: Assess their knowledge mastery and generate personalized ability reports; AI-assisted management system is used for intelligent class scheduling, attendance statistics, asset management, and logistics scheduling. It is mainly used to assist schools or management institutions in improving the efficiency of the arrangement and management of teachers and public property. The AI-assisted training system provides simulated training and personalized training programs for skills training scenarios for teachers or students, and offers real-time feedback. This makes the entire computing center a multi-functional integrated system for both teachers and students, improving the efficiency of the entire process and enabling the computing center to achieve multi-functional integration. Example

[0024] This invention also provides an AI large model base computing center installation device, comprising: Multiple cabinets 1, in this embodiment, include one long cabinet and six short cabinets. The seven cabinets 1 are welded together in sequence and are used to install the dynamic environment regulation and control module, power management module, training and push module, intelligent middleware module and AI education module. The surface of each of the seven cabinets 1 is provided with a heat dissipation mesh to ensure the heat dissipation effect of the internal modules.

[0025] Display screen 2 is installed on every two low cabinets 1. It is used to display the working status of the dynamic environment adjustment and control module, power management module, training and push module, intelligent middleware module and AI education module, so as to ensure that the working status and content of the entire computing center can be seen intuitively and to ensure the visibility of the entire computing power working in the center.

[0026] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An AI large-scale model base computing center, characterized in that, include: The dynamic environment regulation and control module is used to collect and regulate environmental data in real time. The power management module is used to provide power and distribute it appropriately to various devices; The training and inference module is used to integrate hardware devices, support large model training and inference, and enable bidirectional data transmission with the intelligent middleware module. The intelligent middleware module provides reusable capabilities for data management, AI algorithms, and business components, supporting the rapid development and collaboration of the AI ​​education module while ensuring the security and data governance of the AI ​​education module. The AI ​​education module is used to edit digital teaching materials, assist teachers in teaching innovation, generate personalized learning paths based on data provided by the intelligent platform module and usage data, and evaluate the usage behavior of students and teachers.

2. The AI ​​large-scale model base computing center according to claim 1, characterized in that: The power management module includes mains power, a power distribution unit, and a UPS power supply. The power distribution unit connects the UPS power supply and the mains power supply, and distributes power between the UPS power supply and the mains power supply according to power consumption.

3. The AI ​​large-scale model base computing center according to claim 2, characterized in that: The dynamic environment control module includes an intelligent environmental monitoring system and a precision air conditioner. The intelligent environmental monitoring system is used to collect environmental data from the computing center in real time and control the temperature of the precision air conditioning. Precision air conditioning is used to provide a constant temperature and humidity operating environment for computing centers.

4. The AI ​​large-scale model base computing center according to claim 3, characterized in that: The training and push module includes a training and push all-in-one machine, a switch, and a pneumatic-hydraulic CDU. The training and inference integrated machine is used for large-scale model training and inference. The switch is used to ensure signal transmission between the training and promotion all-in-one machine, the intelligent middleware module, and the AI ​​education module; The air-liquid CDU provides liquid cooling for the training and propulsion integrated machine.

5. The AI ​​large-scale model base computing center according to claim 4, characterized in that: The AI ​​education module includes a digital textbook development system, a digital teacher innovation system, an AI learning system, an AI evaluation system, an AI assistant management system, and an AI training assistance system. The digital textbook development system is used for the digital creation of textbooks and can generate digital textbooks adapted to different teaching scenarios; The Digital Teacher Innovation System is designed to assist teachers in innovative teaching methods, such as intelligent lesson preparation, resource integration, and interactive courseware creation. An AI learning system is used to receive student data and generate personalized learning paths based on that data. AI evaluation system is used to receive multimodal data analysis to assess teachers' teaching effectiveness or students' learning outcomes; AI-assisted management system for intelligent class scheduling, attendance statistics, asset management, and logistics dispatch; AI-assisted training systems are used to provide simulated training and personalized training programs for skills training scenarios for teachers or students, and to provide real-time feedback.

6. The AI ​​large-scale model base computing center according to claim 5, characterized in that: The intelligent middleware module is an intelligent middleware system.

7. The AI ​​large-scale model base computing center according to claim 6, characterized in that: The multimodal data in the AI ​​assessment system includes data from classroom interactions, assignments, and exams.

8. The AI ​​large-scale model base computing center according to claim 3, characterized in that: The intelligent environmental monitoring system includes a dynamic environmental monitoring system and a dynamic environmental management system. The dynamic environment monitoring system is used to collect environmental data of the computing center in real time. The dynamic environment management system receives the data collected by the dynamic environment monitoring system and controls the precision air conditioning according to the data to manage the environment of the computing center.

9. An installation device for an AI large-scale model base computing center, characterized in that, include: Multiple cabinets connected in sequence are used to install the dynamic environment regulation and control module, power management module, training and promotion module, intelligent middleware module and AI education module as described in any one of claims 1-8; The display screens are installed on every two cabinets to display the working status of the dynamic environment regulation and control module, power management module, training and promotion module, intelligent middleware module and AI education module as described in any one of claims 1-8.

Images