Platform For AI:ModelVersion

The model version, which is unique for the model. If you leave this parameter empty, 0.1.0 is the first version by default. Then, the minor version number incremented by 1 is used as the second version 0.2.0.

The version consists of a major version number, a minor version number, and a patch version number. The version numbers are separated with periods (.). The major and minor version numbers are digits, and the patch version number starts with a digit followed by an underscore (_) and a letter. such as 1.1.0 or 2.3.4_beta.

Regular expression: "^\\d+\\.\\d+\\.\\d+(_\\w+)?$"

The time when the model was created, in UTC. The time follows the ISO 8601 standard.

The time when the model was last modified, in UTC. The time follows the ISO 8601 standard.

The user ID.

The ID of the Alibaba Cloud account.

The URI of the model version, which is the location where the model is stored. The value can be the HTTP(S) address of the model, such as https://myweb.com/mymodel.tar.gz. If the model is stored in an Object Storage Service (OSS) bucket, the value is in the oss://<bucket>.<endpoint>/object format. The endpoint can be queried in the OSS console, such as oss://mybucket.oss-cn-beijing.aliyuncs.com/mypath/.

The model version description.

The labels.

The model format.

• OfflineModel
• SavedModel
• Keras H5
• Frozen Pb
• Caffe Prototxt
• TorchScript
• XGBoost
• PMML
• AlinkModel
• ONNX

The model framework.

• Pytorch
• XGBoost
• Keras
• Caffe
• Alink
• Xflow
• TensorFlow

The extended field. The value is a JSON string.

The model metrics.

The training configuration, used for fine-tuning and incremental training.

The inference configurations applied to the downstream, such as the configuration of the processor or container of Elastic Algorithm Service (EAS). Example: { "processor": "tensorflow_gpu_1.12" }

The evaluation configuration.

The compression configuration.

The source type of the model. Default value: Custom.

• Custom
• PAIFlow
• TrainingService

The source ID.

• If the source type is Custom, this field is not limited.
• If the source type is PAIFlow or TrainingService, the format is:

region=<region_id>,workspaceId=<workspace_id>,kind=<kind>,id=<id>

region is the ID of the Alibaba Cloud region. workspacceId is the ID of the workspace. kind is the type. Valid values: PipelineRun (PAIFlow pipeline) and ServiceJob (training service). id is the unique identifier.

The approval status. Valid values:

• Pending
• Approved
• Rejected

The additional information.