How does Vertex AI help with the MLOps workflow, part 1?

1. How does Vertex AI help with the MLOps workflow, part 1?

Now let’s explore how Vertex AI can help with MLOps processes. Vertex AI gives you total flexibility in the tools you use to complete your ML journeys. Vertex AI was designed to automate ML processes to prevent ML practitioners from starting from the beginning. You can scale your ML models faster and more efficiently with the managed infrastructure provided by Vertex AI. You can also set up ML environments quickly, automate orchestration, manage large clusters, and set up low latency applications. Vertex AI provides a range of AI and ML tools for users of different skill sets alongside built-in MLOps capabilities, which lets enterprises improve insights, make predictions, and automate core business processes across their organization. With Vertex AI, you can: Manage and govern your ML models. Leverage Google Cloud’s managed services to simplify your MLOps processes. Reveal the explanations behind your models and predictions. Monitor your data and models’ performance, and Track and compare multiple experiment runs and analyze main model metrics. Let’s look at Vertex AI’s key components that let you perform those capabilities. Let’s start with managing and governing capabilities. You can manage and govern your ML models with Vertex AI Feature Store, Model Registry, ML Metadata, and model evaluation. Manage features: You can use Vertex AI Feature Store to create and manage features. Vertex AI Feature Store lets you: Share and reuse ML features across use cases. Serve ML features at scale with low latency. And alleviate training serving skew. Manage models: You can use the Vertex AI Model Registry to manage your ML models and use Vertex ML Metadata to track and analyze the metadata produced by your ML workflows. Vertex AI Model Registry is a central repository where you can manage the lifecycle of your ML models. Vertex AI Model Registry lets you: Register, organize, track, and version your trained and deployed ML models. Govern the model launch process. And maintain model documentation and reporting. Vertex ML Metadata lets you record the metadata and artifacts produced by your ML system. Therefore, it lets you: Automatically track inputs and outputs of all components. Query the metadata to help analyze, debug, and audit the performance of your ML system or the artifacts that it produces. And visualize, analyze, and compare detailed ML lineage. Evaluate models: You can run model evaluations in Vertex AI in several ways: First, you can create evaluations through Vertex AI Model Registry in the Google Cloud console. Second, you can use the model evaluation feature from Vertex AI as a pipeline component with Vertex AI Pipelines. Let’s move on to orchestrating ML workflow’s capabilities. You can simplify ML operations by using Vertex AI Pipelines to automate, monitor, and govern your ML systems. Vertex AI Pipelines lets you orchestrate your ML workflow in a serverless manner thanks to Google Cloud’s managed services suite such as BigQuery, Vertex Training, or Dataflow. ML pipelines are portable and scalable ML workflows that are based on containers. Therefore, ML practitioners can iterate faster, get more of their work into production, and work more independently. ML pipelines are composed of a set of input parameters and a list of steps. Each step is an instance of a pipeline component. ML pipelines let ML practitioners spend their time building ML solutions instead of building the infrastructure needed to get those solutions into production. You can use ML pipelines to: Apply MLOps strategies to automate and monitor repeatable processes. Experiment by running an ML workflow with different sets of hyperparameters, for example, different number of training steps or iterations. Reuse a pipeline's workflow to train a new model. You can use Vertex AI Pipelines to run pipelines that were built using the Kubeflow Pipelines SDK or TensorFlow Extended. To learn more about choosing between the Kubeflow Pipelines SDK and TFX, check the reading at the end of section. After you’re satisfied about your ML workflow, the next step is to understand your model’s behavior. You’ll explore that process in the next video.

2. Let's practice!

Create Your Free Account

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.