edf86a8ae7
ModelArts GA UMN 06052022 from R&D R&D provided a new version of the ModelArts User Manual in May 2022. Depends-On: #11 Reviewed-by: Artem Goncharov <Artem.goncharov@gmail.com>
4.1 KiB
Where Are Models Generated by ExeML Stored? What Other Operations Are Supported?
Unified Model Management
For an ExeML project, after the model training is complete, the generated model is automatically displayed on the Model Management > Models page. The model name is automatically generated by the system. Its prefix is the same as the name of the ExeML project for easy identification.
Models generated by ExeML cannot be downloaded.
What Other Operations Are Supported for Models Generated by ExeML?
- Deploying models as real-time and batch services
On the ExeML page, models can only be deployed as real-time services. You can deploy models as batch services on the Model Management > Models page.
It should be noted that resources with other specifications, such as dedicated resource pools and GPU resources, can be used when you create a model deployment task on the Model Management > Models page. On the ExeML project page, only ExeML (CPU) can be used to deploy models.
- Creating a version
When creating a new version, you can select a meta model only from a ModelArts training job, OBS, model template, or custom image. You cannot create a version from the original ExeML project.
- Deleting a model or its version