Vertex AI Models and Predictions
Vertex AI Models provides a platform for managing ML models. With Vertex AI Models, you can develop and manage ML models in many ways:
- Create model: Users can choose to create a new model and be redirected to the training Pipelines screen.
- Upload model: Users can upload a model that’s been trained elsewhere for use within their Vertex AI project.
- Deploy model: Users can deploy a selected model to an endpoint, making it available through a REST API.
- Export model: Users can export a trained model to a GCS bucket, where it can be stored or used in another project.
After the models are trained, they can be exported or deployed publicly or privately to predict production cases. When you deploy a model to an endpoint resource for online predictions, or when you request batch predictions, you can always customize the VM types that the prediction service uses, and you can also configure prediction nodes to use GPUs. We will...
