检测到您已登录华为云国际站账号，为了您更好的体验，建议您访问国际站服务网站 https://www.huaweicloud.com/intl/zh-cn
不再显示此消息
检测到您已登录华为云国际站账号，为了您更好的体验，建议您访问国际站服务网站 https://www.huaweicloud.com/intl/zh-cn
不再显示此消息
Importantly, MoXing Framework is decoupled from specific AI engines and can be seamlessly integrated with all major AI engines (including TensorFlow, MXNet, PyTorch, and MindSpore) supported by ModelArts.
An ExeML project supports multiple rounds of training, and each round generates an AI application version. For example, the first training version is 0.0.1, and the next version is 0.0.2. The trained models can be managed by training version.
Algorithm Source Frequently-used Select an AI engine and its version and specify Code Directory and Boot File. The framework selected for the AI engine must be the same as the one you select when compiling training code.
An ExeML project supports multiple rounds of training, and each round generates an AI application version. For example, the first training version is 0.0.1, and the next version is 0.0.2. The trained models can be managed by training version.
Figure 1 Selecting an AI engine After the file is created, the JupyterLab page is displayed by default. Figure 2 Encoding page Calling mox.file. Enter the following code to implement the following simple functions: Introduce MoXing Framework.
An ExeML project supports multiple rounds of training, and each round generates an AI application version. For example, the first training version is 0.0.1, and the next version is 0.0.2. The trained models can be managed by training version.
An ExeML project supports multiple rounds of training, and each round generates an AI application version. For example, the first training version is 0.0.1, and the next version is 0.0.2. The trained models can be managed by training version.
Managing Team Labeling Tasks For datasets with team labeling enabled, you can create team labeling tasks and assign the labeling tasks to different teams so that team members can complete the labeling tasks together. During data labeling, members can initiate acceptance, continue
Specifications for Importing the Manifest File The manifest file defines the mapping between labeling objects and content. The Manifest file import mode means that the manifest file is used for dataset import. The manifest file can be imported from OBS. When importing a manifest file
Data Selection Overview of Data Selection Operators ModelArts provides the following data selection operators: The SimDeduplication operator can implement image deduplication based on the similarity threshold you set. Image deduplication is a common method for image data processing
Data Validation MetaValidation Operator Overview ModelArts data validation is implemented by the MetaValidation operator. ModelArts supports the following image formats: JPG, JPEG, BMP, and PNG. The object detection scenario supports the XML labeling format but does not support the
Data Cleansing PCC Operator Overview ModelArts data cleansing is implemented by the PCC operator. The dataset used for image classification or object detection may contain images that do not belong to the required categories. These images need to be removed to avoid interference to
Data Processing Introduction to Data Processing Creating a Data Processing Task Managing and Viewing Data Processing Tasks Built-in Operators Parent topic: Data Management (Old Version to Be Terminated)
Autocompletion for ma-cli Commands CLI autocomplete enables you to get a list of supported ma-cli commands by typing a command prefix and pressing Tab on your terminal. Autocomplete for ma-cli commands needs to be enabled in Terminal. After running the ma-cli auto-completion command
Configuring Workflow Parameters Description A workflow parameter is a placeholder object that can be configured when the workflow runs. The following data types are supported: int, str, bool, float, Enum, dict, and list. You can display fields (such as algorithm hyperparameters) in
Creating Workflow Phases Creating a Dataset Phase Creating a Dataset Labeling Phase Creating a Dataset Import Phase Creating a Dataset Release Phase Creating a Training Job Phase Creating a Model Registration Phase Creating a Service Deployment Phase Parent topic: Workflow Development
Creating a Multi-Branch Workflow Multi-Branch Workflow Creating a Condition Phase to Control Branch Execution Configuring Phase Parameters to Control Branch Execution Configuring Multi-Branch Phase Data Parent topic: Workflow Development Command Reference
Advanced Workflow Capabilities Using Big Data Capabilities (MRS) in a Workflow Specifying Certain Phases to Run in a Workflow Parent topic: Workflow Development Command Reference
Configuring SFS Turbo and OBS Interworking SFS Turbo HPC file systems can access objects stored in OBS buckets seamlessly. You can specify an SFS Turbo interworking directory and associate it with an OBS bucket. Log in to the SFS console. In the left navigation pane, choose SFS Turbo
GalleryModel: defines a model subscribed from AI Gallery. This object is used for model registration. Placeholder data objects, which are specified when a workflow is running DatasetPlaceholder: defines datasets to be specified when a workflow is running.
