products
Product matrix
The products of AI Lab consist of intelligent terminals, platform courses and training kits. The intelligent terminals, including teaching interaction terminals, teaching recording and broadcasting terminals and other teaching - supporting terminals, are used to build the basic functions of smart classrooms. The platform courses are composed of AI general knowledge courses, experimental and practical projects, together with supporting algorithms and datasets. The training kits include physical simulation kits and virtual simulation kits.
Physical Simulation Kit: AI Comprehensive Training Box
Functional Features
Cover mainstream AI teaching directions
The AI training box is designed to meet the teaching needs of artificial intelligence and related majors. It covers multiple core areas such as Python programming, machine learning, deep learning, digital image processing, computer vision, speech recognition, embedded systems, and intelligent robots. It can support multi - level teaching from basic programming to comprehensive applications.
Out-of-the-box integrated training
The training box features an integrated experimental - box design. With a solid overall structure and an aluminum - alloy frame for reinforcement, it is suitable for high - frequency teaching and long - term use. The device integrates a 17 - inch high - definition display, a full - size keyboard, a mouse, and necessary experimental teaching aids. Users can directly carry out teaching and training without additional peripheral devices.
Supports multimodal combined training
The AI training box integrates the computing power of edge - computing terminals, supports the deployment of mainstream deep - learning frameworks such as PyTorch and TensorFlow, and integrates various technical components including the Linux system, machine vision, speech recognition, embedded sensors, and robotic arm control.
Physical Simulation Kit: AI+Intelligent Manufacturing Integrated Training Sand Table
Functional Features
Deep integration of AI and manufacturing
This desktop - type intelligent manufacturing production line, as an important carrier for the integrated training of AI + majors, introduces artificial intelligence technology into real manufacturing scenarios, supporting the application of AI capabilities such as machine vision, intelligent recognition, data analysis, and robot control in the manufacturing process.
Real production line presented on the desktop
This solution restores real - world application scenarios at a lower cost. It is suitable for carrying out multiple parallel training sessions in a teaching environment, providing a sustainable practical foundation for the integrated teaching of AI and manufacturing.
Support collaborative teaching for multiple majors
Through modular design and a unified control platform, it supports combined experiments of multiple courses and technical modules. It guides students to understand the collaborative application of AI technology in complex systems through cross - professional collaboration, enhancing their comprehensive engineering practice ability.
Virtual Simulation Kit: Naked - eye 3D Holographic All-in-One Machine
Functional Features
Naked - eye 3D holographic display technology
No need to wear any device. It has an adaptive viewing angle and a self - developed ultra - low - latency 3D display algorithm, enabling you to enjoy realistic and natural 3D content immediately.
27-inch 4K 60Hz Ultra HD display
Equipped with a 12th generation Intel Core i7 processor and an NVIDIA GeForce RTX 3060 graphics card. The resolution is 3840 x 2160 in 2D mode and 1920 x 2160 in 3D mode, with a refresh rate of ≥60Hz.
Note: A 16-inch version is also available.
Rich selection of accessories, app store, and content ecosystem
A 6DOF spatial interaction pen, designed specifically for glasses-free 3D environments. It features a proprietary app store offering a wealth of software resources. It also boasts a leading VR education ecosystem in China, with content covering K12 and numerous vocational and technical fields.
Virtual Simulation Kit: Naked - eye 3D Laptop
Functional Features
High-performance processor and graphics card configuration
Equipped with Intel HX series processors and RTX 5060/5070 graphics cards, it can smoothly run complex 3D model rendering and engineering simulations, meeting the demanding computational performance requirements of high-end teaching and practical training.
Naked - eye 3D visual presentation
Using display technology that doesn't require 3D glasses, it can directly present high-fidelity stereoscopic images, making three-dimensional models of engineering structures, biological anatomy, and other subjects more intuitive and vivid.
Integrated design for professional training
Optimized specifically for advanced model analysis, engineering simulation, and medical 3D learning scenarios, it provides complete and efficient technical support for complex courses and practical training.
Virtual simulation kit: Multi - person 3D interactive intelligent large - screen
Functional Features
Ultra-high-definition 3D display, with vivid and lifelike images
It features an 86-inch 4K ultra-high-definition display, supporting adaptive 2D/3D switching. Combined with GPR 3D display technology, users can experience ultra-high-definition stereoscopic images by wearing lightweight 3D glasses, delivering realistic and immersive visuals.
AI gesture recognition for intuitive, efficient interaction
Equipped with a self-developed AI intelligent recognition algorithm, it can accurately capture hand gestures. Interaction can be completed without external devices, significantly improving operational efficiency during teaching and presentations, making human-computer interaction more natural and convenient.
Flagship hardware config, powering AI teaching scenarios
It features a built-in Rockchip RK3588 octa-core CPU, a third-generation Valhalla architecture GPU, and a 6 TOPS AI accelerator. Combined with 8GB of RAM, 128GB of expandable storage, and a rich set of interfaces, it can smoothly run 3D course resources developed with Unity/UE, fully supporting digital simulation training and other applications.
Platform courses: AI Teaching and Training Platform
Functional Features
Interactive programming environment
The platform is built on the Jupyter open-source framework and provides a complete programming interface accessible directly in your browser. Users do not need to configure a local environment and can directly create various programming resources such as Notebooks, consoles, and terminals, meeting the immersive practical needs ranging from basic syntax to project development.。
Hierarchical management and teaching collaboration
The platform provides a comprehensive user, role, and class management system. It allows for the flexible creation of administrative or teaching classes, adapting to different teaching organizational structures and enabling scalable academic management needs such as course assignment and progress tracking.
Multilingual support and real-time feedback
It deeply integrates Jupyter's interactive programming features and supports multiple language kernels such as Python, C++, and Java. Combined with project management and keyboard shortcuts, it significantly improves code writing and debugging efficiency.
Platform courses: Large Language Model Training and Inference Training Platform
Functional Features
One-stop training for the whole process
The platform connects data annotation, model fine-tuning, and deployment evaluation into a coherent closed loop. Users can complete the entire process, from raw data preparation to model performance validation, without switching between multiple independent tools.
Low - threshold and highly customizable
The platform strikes a good balance between ease of use and professional functionality. On the one hand, it provides a clear and intuitive interface, allowing beginners to get started quickly. On the other hand, it offers advanced users granular control capabilities, including Lora configuration and hyperparameter tuning.
Combat - oriented evaluation and motivation
The platform incorporates a multi-dimensional model evaluation system (using metrics such as ROUGE and BLEU) and a class ranking function, gamifying and making the learning process more practical.
Platform courses: Online Learning Platform
Functional Features
Systematized cutting-edge course content
The platform offers a comprehensive course content system, covering a full range of cutting-edge topics from large language model fundamentals, data annotation, and training (including full-parameter training, fine-tuning, and distillation) to prompt engineering, RAG, and intelligent agent application development.
Clear and intuitive learning progress management
The platform provides a real-time dashboard of your personal learning progress. In the "My Learning" center, you can view your learning progress and recent study time for each course at any time.
Convenient course selection and organizational management
The platform supports filtering by academic level and course category, as well as various sorting options, to help you quickly find the courses you need. The platform also provides comprehensive multi-role permission and class management functions.
Platform courses: Virtual Simulation Teaching Platform
Functional Features
Massive resources and convenient lesson preparation
The platform provides teachers and students with over 4,000 school-based model resources and a corresponding VR editor. Teachers can easily combine their personal resources or self-created courseware into engaging lessons, just like dragging and dropping building blocks.
Immersive interaction and flexible teaching
In the classroom, teachers can intelligently control and manage student devices, conducting virtual demonstrations and experimental assessments. Students, in turn, perform practical exercises in a highly realistic virtual environment, with the system providing immediate feedback on their actions.
Data-driven and precise evaluation
The platform has powerful BI analysis capabilities. It can conduct data statistics from multiple dimensions such as resource utilization, teaching behavior, and experimental scores. The system automatically generates class reports and teacher profiles to help teachers accurately evaluate teaching effectiveness.
Teaching support: Large Language Model AI Inference All-in-One Machine
Functional Features
Ready to use out-of-the-box, quick to deploy
The host computer comes pre-installed with Ubuntu, mainstream open-source models, and all necessary software dependencies, enabling rapid deployment within 10 minutes of powering on. With its powerful 330 TFLOPS computing power and 48GB of large video memory, it can easily utilize large language models such as DeepSeek and Qwen with a single click.
Rich in functions, comprehensively intelligent
With a throughput of up to 901 tokens/s and support for 50 concurrent users, it can simultaneously handle multiple complex tasks, meeting the real-time intelligent interaction and analysis needs of various scenarios, including education and enterprise applications.
Safe and reliable, with complete services
All data processing is completed locally, ensuring data privacy and security, and without any additional cloud service costs.