No Experience Needed. ModelArts Makes It Easy.
ModelArts ExeML automates model design, parameter optimization; and model training, compression, and deployment based on labeled data. You can build AI models and apply them to actual services without any coding and model development experience.

ModelArts ExeML

ModelArts ExeML automates model design, parameter optimization, and model training, compression, and deployment based on labeled data. You can build AI models and apply them to actual services without any coding and model development experience.

Create an AI Model in Three Steps, No Coding or AI Experience Needed

Compared with conventional AI model training and deployment, ModelArts ExeML greatly simplifies AI development and reduces costs by more than 90%.

4+ Application Scenarios

ModelArts supports image classification, object detection, predictive analytics, and sound classification.
Image classification
Object detection
Predictive analytics
Sound classification

AI Development Lifecycle

Efficient and Accurate AI Development
Data Preparation

7 Labeling Scenes for 3 Scenarios

7 labeling scenes for images, sounds, and texts to meet data requirements of popular models
Use Now
Sound segmentation
Sound classification
Image classification
Text comparison
Text translation
Text classification
Named entity recognition
Object detection

Reduced Labor Costs

Auto labeling based on pre-labeling models of active learning and scenarios of high value (such as MDC) reduces labor costs by 50% to 80%.

Visualized Version Management

Visualized management of data versions facilitates model development. Datasets can be imported and exported conveniently.
Use Now
Model Creation

Out-of-the-Box Online Development Environment

Out-of-the-Box Online Development Environment

ModelArts integrates Jupyter Notebook to provide online interactive development and debugging tools. You can create a development environment to develop, debug, and train models. The environment installs common machine learning engines and software libraries for you to use immediately.
Use Now

Accelerated Model Development That Is Kept Simple

Accelerated Model Development That Is Kept Simple

1. ModelArts provides a large number of built-in machine learning models trained on large datasets with optimized precision and speed. You can use the models for transfer learning, simplifying model development.

2. ModelArts integrates the Huawei-developed accelerated framework, MoXing, built on open source frameworks such as TensorFlow, MXNet, PyTorch, and Keras. You can use MoXing to simplify model development and optimization.

Only a small amount of code is required to transition from standalone training to distributed training.

The distributed model training based on MoXing can be performed on hundreds of GPUs, greatly reducing the time required for model training.

ModelArts automatically performs optimization analysis and searches for appropriate model hyper-parameters. This saves the time that would normally have to be spent on manual parameter adjustment and on multiple rounds of model training. It can cut days or even weeks off the time required for model training and can significantly improve model quality.

Use Now
Model Deployment

Model Deployment

Models trained on ModelArts can be quickly published as real-time inference services that can handle high throughput with low latency.

ModelArts supports batch inference tasks. It can process a large amount of data, with efficient distributed computing.

ModelArts can optimize models to better support edge devices. Working with Huawei HiLens, ModelArts can easily deploy models on cameras or other edge devices.

Buy Now

Interconnecting Local IDEs with ModelArts

In addition to the management console, ModelArts also provides Python SDKs. You can use these SDKs to access ModelArts through any local IDE, and then create and train models and deploy services. This makes AI development more user-friendly.

AI Marketplace

Developer Ecosystem Built Based on ModelArts

The AI marketplace is a developer ecosystem built based on ModelArts. It allows you to share AI models, data, and competition cases, as well as perform API transactions. The AI marketplace is a secure and open sharing and transaction environment for scientific research institutes, AI application developers, solution integrators, enterprises, and individual developers. The AI marketplace connects all of the participants in the AI development ecosystem and accelerates the development and implementation of AI products.

Coming Soon
Featured Stories
HUAWEI CLOUD helps Miaopai, the short-video platform hailed as China's Vine, increase the average click-through rate for the short video service by 20.01% and increase the average number of views per person by 10.42%.

Year-End Sales Carnival, Free Coupons Worth $1,000

Claim Now