Application Scenarios
Application Scenarios
Improve the Development Capability
Enterprise ModelArts development capabilities are insufficient and require all-round development guidance services from professional teams.
.svg)
.svg){kind=icon}
Improve Modeling Capabilities
Enterprises are not capable of developing model algorithms. Professional teams are required to provide expert guidance services for developing model algorithms.
.svg)
Improve the Optimization Capability
Professional teams are required to provide model optimization, acceleration, and performance improvement services.
Service Highlights
Service Highlights
Extensive Industry Experience
Based on a large number of industry delivery experience, the team has experience and capabilities in various complex service scenarios, and can better provide scenario-specific modeling analysis and design services for customers.
Efficient R&D Force
Experienced and efficient R&D teams help customers quickly analyze and design scenario-specific modeling solutions.
Advanced Service Concept
Adhere to the customer-centric service concept and focus on solving customers' business pain points.
Service Content
Service Content
AI Platform Scenario-specific Model Development Service
Provides modeling consulting services based on customer service scenarios to help customers build scenario-specific models, plan and design, customize cloud brain development services, and orchestration services.
_1.svg)
AI Platform Scenario-specific Model Tuning Service
Provides model training, fine tuning, distillation, optimization, acceleration, and performance improvement services for existing models.
_1.svg)
_1.svg){kind=icon}
AI Cloud Service Migration and Adaptation Service
Provides migration adaptation services based on the AI platform and AI cloud service solution to help customers complete AI cloud service migration adaptation.
AI Platform Implementation Service
Provides computing power base platform design and implementation services, including high-performance network and storage.
AI Platform Development Support Service
Provides SDK and API usage guide, model guide, and environment setup guide.
Service Pricing Service Pricing
-
AI Platform Scenario-specific Model Development Service
-
AI Platform Scenario-specific Model Tuning Service
-
AI Cloud Service Migration and Adaptation Service
-
AI Platform Implementation Service
-
AI Platform Development Support Service
- AI Platform Scenario-specific Model Development Service
- AI Platform Scenario-specific Model Deployment Service
- Application and AI Agent Orchestration Service
Basic: Customers need to design AI solutions based on actual problem scenarios and need Huawei's help to establish AI solutions (excluding UseCase). Model indicators are verified and guaranteed in historical service data of customers.
Standard: Customers need to design AI solutions based on their actual problem scenarios. Huawei helps build AI solutions and provide them as use cases. Model indicators are verified and ensured in historical service data of customers.
Professional: Customers need to design AI solutions based on their actual problem scenarios and need Huawei to help build AI solutions (excluding use cases). Model indicators can be verified and ensured in real-time service data of customers.
Platinum: The customer needs to complete algorithm training based on actual data to obtain algorithm model results and ensure the precision of the results. Huawei needs to provide scenario-based use cases to verify and ensure model indicators in real-time service data of the customer.
Basic: The customer needs to complete E2E training for the delivered solution and deploy models as RESTAPIs to connect to the online system. Based on the technical solution, the customer needs to develop inference scripts and create cloud inference images and complete deployment, guiding the customer to interconnect algorithm APIs with service systems. The QPS of the deployment service is 1, and the number of resource devices is 1.
Standard: Customers need to perform E2E training on AI assets on Huawei cloud and deploy models as RESTAPIs to connect to the online system. Based on the technical solution, customers need to develop inference scripts and cloud inference images and complete the deployment, guiding customers to interconnect algorithm APIs with service systems. The QPS of the deployment service ranges from 2 to 5, and the number of resource devices ranges from 2 to 5.
Professional: Customers need to perform E2E training on AI assets on Huawei cloud and deploy models as RESTAPIs to connect to online systems. Based on the technical solution, customers need to develop inference scripts and cloud inference images and complete the deployment, guiding customers to interconnect algorithm APIs with service systems. The QPS of the deployment service ranges from 6 to 50, and the number of resource devices ranges from 6 to 10.
Platinum: Customers need to complete E2E training for delivered solutions or AI assets on HUAWEI CLOUD, deploy models as a REST API to connect to customers' online systems, deploy models in edge systems, and finally perform service logic acceptance. The QPS of the deployment service is greater than 50, and the number of resource devices is greater than 10.
Standard: Supports quick integration of DeepSeek APIs and functional modules to meet medium and low complexity requirements. (e.g. database connection, networked search, knowledge base, etc.).
Professional: For complex service systems that require in-depth customization, customers can integrate DeepSeek APIs and functional modules. (e.g. database connection, networked search, knowledge base, etc.), supports the construction of multi-agent matrices, and completes complex tasks through functions such as task planning and reflection execution.
- AI Platform Scenario-specific Model Tuning Service
- Model Distillation Service
- Incremental Model Training Service
- Model Fine-Tuning Service
Basic: Customers need to optimize the indicators or precision of the delivered solution. we provides the algorithm optimization guide and verifies the optimization in offline data.
Standard: Customers need to optimize scenario-specific indicators or precision for AI assets on HUAWEI CLOUD. Huawei provides algorithm optimization guidance and verifies the optimization in actual services.
Professional: If the customer needs to optimize the scenario migration KPIs or precision for the delivered solution, we provides the algorithm optimization guide.
Platinum: Customers need to optimize scenario migration indicators or precision for delivered AI assets or solutions. We provide algorithm optimization guidance, verify the optimization in actual services, and provide new use cases.
Standard: It is applicable to the expected distillation dataset with thinking chain capability and strong logic capability.
Professional: This scenario is applicable to customers who expect to obtain domain models with strong thinking chain capabilities and logic capabilities.
Platinum: Applicable to customers who expect to obtain domain models with strong thinking chain capabilities and logic capabilities.
Standard: Applicable to customers who expect to obtain large industry models without thinking chain capabilities and with average logic capabilities.
Professional: Applicable to customers who want to obtain large industry models with strong logic capabilities but no thinking chain capabilities.
Platinum: Applicable to customers who expect to obtain large industry models with strong thinking chain capabilities and logic capabilities.
Standard: This mode is applicable to the scenario where customers have a small amount of industry instruction knowledge data and can quickly build a scenario model based on a scenario.
Professional: This mode is applicable to the scenario where customers have rich industry instruction knowledge data and can quickly build a scenario model based on a scenario.
Platinum: This feature is applicable to the scenario where customers have rich industry instruction knowledge data and quickly build a scenario model based on a scenario to guide customers to make better decisions in complex tasks.
- AI Cloud Service Migration and Adaptation Service
- AI Cloud Service AI Model Migration Service - per Model
- AI Cloud Service Application Migration Service - per Application
Basic: This document is applicable to the scenario where the customer adapts the algorithm model migration and adaptation in the open source adaptation list provided by Huawei. Based on the guide, the demo solution is demonstrated and technical support is provided during the implementation of the customer's service.
Standard: This document is applicable to the scenario where Huawei provides the migration adaptation service. For the algorithm model migration adaptation in the Huawei open source adaptation list, implement the solution based on the guide document, and complete precision alignment, performance optimization, model deployment based on the dataset and open source weight file specified by the customer. A specified dataset and open source weight file are specified.
Professional: Applicable to the scenario where Huawei provides the migration adaptation service. For the algorithm model migration adaptation that is not included in the Huawei open source adaptation list, perform precision alignment, performance optimization, and model deployment based on the dataset and open source weight file specified by the customer when all operators meet the requirements. Limits a specified dataset and open source weight.
Applies to multi-model scenarios. After the basic migration service is provided for customers, other models need to be added for adaptation. Provides support services for migrating data to the AI cloud service platform, including migration analysis, adaptation, verification, and model deployment.
This feature is applicable to scenarios where customers have application migration requirements.
Basic: Designs the implementation of the computing backbone platform, including high-performance networks and storage.
Designs the implementation of the AI platform solution, including resource scheduling, model training solution, and model inference solution.
Application:ModelArts cluster supports 64 PUs.
Standard: Designs the implementation of the computing backbone platform, including high-performance networks and storage.
Designs the implementation of the AI platform solution, including resource scheduling, model training solution, and model inference solution.
Application:ModelArts cluster supports 64 to 256 PUs.
Professional: Designs the implementation of the computing backbone platform, including high-performance networks and storage.
Designs the implementation of the AI platform solution, including resource scheduling, model training solution, and model inference solution.
Application:ModelArts cluster supports 64 to 1024 PUs.
Platinum: Design and implementation based on the computing power foundation platform, including high-performance network, storage, resource scheduling, model training solution implementation, and model inference solution implementation.
Application:ModelArts cluster over 1024 PUs.
Basic: Lightweight AI exploration: Customers want to quickly experience basic functions of ModelArts without complex development or integration.
Standard: Individual/Small-scale project verification: used by individual developers or students to complete course projects and small-scale experiments.
Professional: Customization requirements for small- and medium-sized enterprises: AI capabilities need to be embedded in existing service systems (such as customer service and report analysis).
Platinum: Preliminarily interconnecting with third-party models: Integrate open-source models (such as Hugging Face) to extend functions.
Flagship: Optimization in complex service scenarios: high-precision model training (such as industrial quality inspection) and multi-model pipeline orchestration.
