SlideShare uma empresa Scribd logo
1 de 50
Baixar para ler offline
Vertex AI
Unified ML Platform for the entire
AI workflow on Google Cloud
DevFest Season, November 2021
Márton Kodok
Google Developer Expert at REEA.net
● Among the Top3 romanians on Stackoverflow 195k reputation
● Google Developer Expert on Cloud technologies
● Crafting Web/Mobile backends at REEA.net
● BigQuery + Redis database engine expert
Slideshare: martonkodok
Articles: martonkodok.medium.com
Twitter: @martonkodok
StackOverflow: pentium10
GitHub: pentium10
Vertex AI - Unified ML Platform @martonkodok
About me
1. What is Vertex AI
2. Gather, Import & label datasets
3. Build, train and deploy ML solutions
4. Manage your models with confidence
5. Using Pipelines throughout your ML workflow
6. Adapting to changes of data
7. Conclusions
Agenda
Vertex AI - Unified ML Platform @martonkodok
“VertexAI is a managed ML platform for practitioners
to accelerate experiments and deploy AI models.
Vertex AI - Unified ML Platform @martonkodok
Where does VertexAI fit in?
Application Servers
Vertex AI
Desktop client
Mobile client
End-to-end platform for
ML model development
and deployment
Backend
Vertex AI - Unified ML Platform @martonkodok
Application Logic
VertexAI is a unified MLOps platform
Vertex AI - Unified ML Platform @martonkodok
Operational
Model
Programming
Model
No Infra Management Managed Security Pay only for usage
Model-as-a-service
oriented
Streamlined model
development
Open SDKs,
integrates with ML frameworks
What’s included in VertexAI?
Vertex AI - Unified ML Platform @martonkodok
Data Labeling
AutoML
DL Environment (DL VM + DL Container)
Prediction
Feature Store Training
Experiments
Data Readiness
Feature
Engineering
Training/
HP-Tuning
Model
Monitoring
Model serving
Understanding/
Tuning
Edge
Model
Management
Notebooks
Pipelines (Orchestration)
Explainable AI
Hybrid AI
Model
Monitoring
Metadata
Vision Translation Tables
Language
Video
AI Accelerators
Models
Datasets
Custom Models
Containers
Python
Endpoints
Vizier Optimization
VertexAI supports...
Vertex AI - Unified ML Platform @martonkodok
UI based model development
# Define job
job = aiplatform.AutoMLTabularTrainingJob(
display_name='price-predict-training',
optimization_prediction_type='regression'
)
# Run job
model = job.run(
dataset=ds,
target_column='median_house_value',
model_display_name='house-value-prediction',
)
Code-based model development
Using Vertex AI throughout your ML workflow
Vertex AI - Unified ML Platform @martonkodok
Gather data Train model
Scalably
deploy
model
Evaluate,
monitor,
retrain
@martonkodok
Gather,
Import & label
datasets at
scale
Part #1
VertexAI: Gather, Import & label datasets at scale
Vertex AI - Unified ML Platform @martonkodok
Data Labeling
AutoML
DL Environment (DL VM + DL Container)
Prediction
Feature Store Training
Experiments
Data Readiness
Feature
Engineering
Training/
HP-Tuning
Model
Monitoring
Model serving
Understanding/
Tuning
Edge
Model
Management
Notebooks
Pipelines (Orchestration)
Explainable AI
Hybrid AI
Model
Monitoring
Metadata
Vision Translation Tables
Language
Video
AI Accelerators
Models
Datasets
Custom Models
Containers
Python
Endpoints
Vizier Optimization
“ Datasets
Vertex AI - Unified ML Platform @martonkodok
Datasets
Vertex AI - Unified ML Platform @martonkodok
Vertex AI datasets (fully managed)
• Fully serverless
• Region based
• Free to store
• In tandem with AutoML managed models
Custom Datasets
Cloud Storage, BigQuery or on Internet
Accessing managed dataset from your app:
- JSONL (default)
- CSV or BigQuery stream
“ VertexAI Managed Datasets + Objectives
(AutoML*)
Vertex AI - Unified ML Platform @martonkodok
* legacy name, previous generation naming from AI Platform
- Regression/classification
- Forecasting
- Single-label classification
- Multi-label classification
- Text entity extraction
- Text sentiment analysis
- Video action recognition
- Video classifications
for entire video, shots, frames
- Video object tracking
Vertex AI: Managed datasets + objectives
Vertex AI - Unified ML Platform @martonkodok
Image Tabular Text Video
- Single-label classification
- Multi-label classification
- Image object detection
- Image segmentation
Vertex AI: Managed dataset + objectives
Vertex AI - Unified ML Platform @martonkodok
Image
Vertex AI: Managed dataset + objectives
Vertex AI - Unified ML Platform @martonkodok
Tabular
Vertex AI: Managed dataset + objectives
Vertex AI - Unified ML Platform @martonkodok
Text
Vertex AI: Managed dataset + objectives
Vertex AI - Unified ML Platform @martonkodok
Video
Data labeling + Feature Store
Vertex AI - Unified ML Platform @martonkodok
Data labeling (fully managed)
• Create data labeling, annotation tasks
• Use human labelers
• Use Google’s labeler workforce
• Use your own workforce
Feature Store (fully managed)
• Centralized repository for organizing, storing,
and serving ML features
• Organization can efficiently share, discover,
re-use features
• Data Model: Entity Type -> Feature
• Ingest data from BigQuery or Cloud Storage
Pro: point-in-time lookup from time series
@martonkodok
Build,
train & deploy
models at
scale
Part #2
2. Train models
Vertex AI - Unified ML Platform @martonkodok
Gather data Train model
Scalably
deploy
model
Evaluate,
monitor,
retrain
VertexAI: Build, train & deploy models at scale
Vertex AI - Unified ML Platform @martonkodok
Data Labeling
AutoML
DL Environment (DL VM + DL Container)
Prediction
Feature Store Training
Experiments
Data Readiness
Feature
Engineering
Training/
HP-Tuning
Model
Monitoring
Model serving
Understanding/
Tuning
Edge
Model
Management
Notebooks
Pipelines (Orchestration)
Explainable AI
Hybrid AI
Model
Monitoring
Metadata
Vision Translation Tables
Language
Video
AI Accelerators
Models
Datasets
Custom Models
Containers
Python
Endpoints
Vizier Optimization
1. AutoML out-of-the box training integration
No-code solution. You must target one of the AutoML’s predefined objectives.
2. Custom Training - run your own training applications in the cloud
Train with one of the Google’s pre-builtcontainers or useyourown.
3. Hyperparameter tuning jobs - searchesforbestcombination of hyperparameter values by optimizing
values across a series of trials. Available for custom training. Your training app must adhere to
accepting Vertex AI parameters. You need to report metrics to Vertex AI.
Training
https://cloud.google.com/vertex-ai/docs/training/using-hyperparameter-tuning @martonkodok
Pre-built containers for custom training
https://cloud.google.com/vertex-ai/docs/training/pre-built-containers @martonkodok
Tensorflow
ML Framework version 1.15, 2.1-2.4
use with Cuda 11.x GPU
scikit-learn
ML Framework version 0.23
No GPUs
PyTorch
ML Framework version 1.4 - 1.7
use with Cuda 11.x GPU
XGBoost
ML Framework version 1.1
No GPUs
3. Deploying models
Vertex AI - Unified ML Platform @martonkodok
Gather data Train model
Scalably
deploy
model
Evaluate,
monitor,
retrain
“ You can deploy models on VertexAI
and get a HTTPs Endpointto serve predictions
rapidly and reliably.
Vertex AI - Unified ML Platform @martonkodok
1. Deploy a model and get aREST endpointto serve predictions realtime or batched
2. You can use models whetherornotthemodelwastrained on Vertex AI.
3. Specify a prediction traffic split in your endpoint.
4. VPC Private Network option for custom-trained models/tabular models
Vertex AI: Endpoints
Vertex AI - Unified ML Platform @martonkodok
Vertex AI Endpoints Backend Prediction
deploy REST
Manage your
models with
confidence
Part #3
VertexAI: Manage your models with confidence
Vertex AI - Unified ML Platform @martonkodok
Data Labeling
AutoML
DL Environment (DL VM + DL Container)
Prediction
Feature Store Training
Experiments
Data Readiness
Feature
Engineering
Training/
HP-Tuning
Model
Monitoring
Model serving
Understanding/
Tuning
Edge
Model
Management
Notebooks
Pipelines (Orchestration)
Explainable AI
Hybrid AI
Model
Monitoring
Metadata
Vision Translation Tables
Language
Video
AI Accelerators
Models
Datasets
Custom Models
Containers
Python
Endpoints
Vizier
Optimization
VertexAI: Manage your models with confidence
Vertex AI - Unified ML Platform @martonkodok
Data Labeling
AutoML
DL Environment (DL VM + DL Container)
Prediction
Feature Store Training
Experiments
Data Readiness
Feature
Engineering
Training/
HP-Tuning
Model
Monitoring
Model serving
Understanding/
Tuning
Edge
Model
Management
Notebooks
Pipelines (Orchestration)
Explainable AI
Hybrid AI
Model
Monitoring
Metadata
Vision Translation Tables
Language
Video
AI Accelerators
Models
Datasets
Custom Models
Containers
Python
Endpoints
Vizier Optimization
Explainable AI
Vertex AI - Unified ML Platform @martonkodok
Explainable AI
• Interpret predictions made by ML models
• Receive a score explaining how much each
factor contributed to the model predictions
• What-If Tool lets you investigate model
behavior at a glance
AI Explanations samples
Github: GoogleCloudPlatform/ai-platform-samples
- Training, deploying, and explaining a tabular data model
- Training, deploying, and explaining an image model
Limitations: doesn’t work well on low-contrast, X-rays, one shade,
panoramas, very tall, very wide images.
Explainable AI: What-if Tool (pair-code.github.io/what-if-tool)
Vertex AI - Unified ML Platform @martonkodok
What-if Tool
• Model probing, from within any workflow
• test performance in hypothetical situations
• analyze the importance of different data
features
• visualize model behavior across multiple
models and subsets of data
Tutorials, demos onpair-code.github.io/what-if-tool
• Available on many platforms (TensorBoard, Jupyter,
Colaboratory, Vertex AI)
• Supports what-if Analyses (explore counterfactuals, fairness
measures, partial dependence plots)
• Visualizes Model Performances (threshold simulation, up to 2
model comparison, dataset summary statistics)
Part #4
Using Pipelines
throughout your
ML workflow
VertexAI: Pipelines - Orchestrate your model
Vertex AI - Unified ML Platform @martonkodok
Data Labeling
AutoML
DL Environment (DL VM + DL Container)
Prediction
Feature Store Training
Experiments
Data Readiness
Feature
Engineering
Training/
HP-Tuning
Model
Monitoring
Model serving
Understanding/
Tuning
Edge
Model
Management
Notebooks
Pipelines (Orchestration)
Explainable AI
Hybrid AI
Model
Monitoring
Metadata
Vision Translation Tables
Language
Video
AI Accelerators
Models
Datasets
Custom Models
Containers
Python
Endpoints
Vizier Optimization
“ Why are MLpipelines useful?
Vertex AI: Pipelines for your MLOps workflows @martonkodok
1. Orchestrate ML workflow steps as a process.
We no longer handle all data gathering, model training, tuning, evaluation, deployment as a monolith.
2. Adopt MLOps for production models. We need a repeatable, verifiable, and automatic process for
making any change to a production model.
3. Develop steps independently -as you scale out, enables you to share your ML workflow with others on
your team, so they can run it, and contribute code. Enablesyoutotracktheinputandoutputfromeach
stepinareproducibleway.
Why are ML pipelines useful?
@martonkodok
“Continuousdelivery and automationpipelines
for machinelearning systems.
Vertex AI: Pipelines for your MLOps workflows @martonkodok
What is MLOps?
MLOps level 0: Manual process
MLOps level 1: ML pipeline automation
MLOps level 2: CI/CD pipeline automation
Levelsofautomation defines maturity of theMLprocess
@martonkodok
MLOps level 0: Manual process - Process for building and deploying ML models is entirely manual.
Infrequent release iterations. No CI, No CD. Disconnection between ML and operations.
MLOps level 1: ML pipeline automation - Continuous training of the model by automating the ML pipeline;
achieve continuous delivery of model prediction service. New pipelines mostly based on new data.
MLOps level 2: CI/CD pipeline automation -iteratively try out new ML algorithms and new modeling where
the experiment steps are orchestrated. The output of this stage is the source code of the ML pipeline steps
that are then pushed to a source repository. Build source. Run test. Output is pipeline.
Levelsofautomation defines maturity of theMLprocess
@martonkodok
MLOps level 2: CI/CDpipelineautomation
@martonkodok
Levelsofautomation defines maturity of theMLprocess
Vertex AI: Pipelines
Vertex AI - Unified ML Platform
Source: Piero Esposito
https://github.com/piEsposito/vertex-ai-tutorials
Pipeline SDK: Condition
Vertex AI: Pipelines for your MLOps workflows
automl_tabular_classification_beans.ipynb
github.com/GoogleCloudPlatform/vertex-ai-samples
1. Use of the Google Cloud Pipeline Components, which support easy access to Vertex AI services
2. Custom Components - function that compiles to a task ‘factory’ function that can be used by pipelines
3. No more Kubeflow Pipelines that must be deployed on a Kubernetes Cluster.
4. Sharing component specifications - the YAML format allows the component to be put under version
control and shared with others, or be used by other pipelines by calling the load_from_url function.
5. Leveraging Pipeline step caching to develop and debug
6. Vertex AI Metadata service + Artifacts Lineage tracking - inverse of pipeline DAG
Developer friendly components
@martonkodok
Part #5
Adapting to
changes of data
Automatic CI / CD Perspective with GCP Services
Vertex AI: Pipelines for your MLOps workflows @martonkodok
Eventarc
• Detect changes on data
• React to events from Cloud services
• Handle events on Cloud Workflows,
Cloud Functions, Cloud Run
• Reuse pipeline spec.json from GCS
• Trigger Vertex AI pipeline
• Detect changes in codebase
• Build pipeline
• Pipeline spec.json to Cloud Storage
• Image to Cloud Registry
• Trigger Vertex AI pipeline
Cloud Build
Cloud Scheduler
• Poll for changes of any data
• Launch based on schedule
• In tandem with Cloud Workflows
• Trigger Vertex AI pipeline
Conclusion
Vertex AI - Unified ML Platform @martonkodok
1. Build with the groundbreaking ML tools that power Google
2. Approachable from the non-ML developer perspective (AutoML, managed models, training)
3. Ease the life of a data scientist/ML (has feature store, managed datasets, endpoints, notebooks)
4. End-to-end integration for data and AI with build pipelines that outperform and solve complex ML tasks
5. GitOps-style continuous delivery with Cloud Build
6. Explainable AI and TensorBoard to visualize and track ML experiments
Vertex AI: Enhanced developer experience
Vertex AI: Pipelines for your MLOps workflows @martonkodok
Thank you.
Slides available on:
slideshare.net/martonkodok
Reea.net - Integrated web solutions driven by creativity
to deliver projects.

Mais conteúdo relacionado

Mais procurados

MLFlow: Platform for Complete Machine Learning Lifecycle
MLFlow: Platform for Complete Machine Learning Lifecycle MLFlow: Platform for Complete Machine Learning Lifecycle
MLFlow: Platform for Complete Machine Learning Lifecycle
Databricks
 
MLflow: Infrastructure for a Complete Machine Learning Life Cycle with Mani ...
 MLflow: Infrastructure for a Complete Machine Learning Life Cycle with Mani ... MLflow: Infrastructure for a Complete Machine Learning Life Cycle with Mani ...
MLflow: Infrastructure for a Complete Machine Learning Life Cycle with Mani ...
Databricks
 

Mais procurados (20)

Microsoft Build 2023 Updates – Copilot Stack and Azure OpenAI Service (Machin...
Microsoft Build 2023 Updates – Copilot Stack and Azure OpenAI Service (Machin...Microsoft Build 2023 Updates – Copilot Stack and Azure OpenAI Service (Machin...
Microsoft Build 2023 Updates – Copilot Stack and Azure OpenAI Service (Machin...
 
MLFlow: Platform for Complete Machine Learning Lifecycle
MLFlow: Platform for Complete Machine Learning Lifecycle MLFlow: Platform for Complete Machine Learning Lifecycle
MLFlow: Platform for Complete Machine Learning Lifecycle
 
MLflow: Infrastructure for a Complete Machine Learning Life Cycle with Mani ...
 MLflow: Infrastructure for a Complete Machine Learning Life Cycle with Mani ... MLflow: Infrastructure for a Complete Machine Learning Life Cycle with Mani ...
MLflow: Infrastructure for a Complete Machine Learning Life Cycle with Mani ...
 
Google Cloud Platform Tutorial | GCP Fundamentals | Edureka
Google Cloud Platform Tutorial | GCP Fundamentals | EdurekaGoogle Cloud Platform Tutorial | GCP Fundamentals | Edureka
Google Cloud Platform Tutorial | GCP Fundamentals | Edureka
 
LanGCHAIN Framework
LanGCHAIN FrameworkLanGCHAIN Framework
LanGCHAIN Framework
 
Global Azure Bootcamp Pune 2023 - Lead the AI era with Microsoft Azure.pdf
Global Azure Bootcamp Pune 2023 -  Lead the AI era with Microsoft Azure.pdfGlobal Azure Bootcamp Pune 2023 -  Lead the AI era with Microsoft Azure.pdf
Global Azure Bootcamp Pune 2023 - Lead the AI era with Microsoft Azure.pdf
 
Building End-to-End Delta Pipelines on GCP
Building End-to-End Delta Pipelines on GCPBuilding End-to-End Delta Pipelines on GCP
Building End-to-End Delta Pipelines on GCP
 
From Data Science to MLOps
From Data Science to MLOpsFrom Data Science to MLOps
From Data Science to MLOps
 
MLOps Virtual Event: Automating ML at Scale
MLOps Virtual Event: Automating ML at ScaleMLOps Virtual Event: Automating ML at Scale
MLOps Virtual Event: Automating ML at Scale
 
Introduction to MLflow
Introduction to MLflowIntroduction to MLflow
Introduction to MLflow
 
Build an LLM-powered application using LangChain.pdf
Build an LLM-powered application using LangChain.pdfBuild an LLM-powered application using LangChain.pdf
Build an LLM-powered application using LangChain.pdf
 
DevOps for Databricks
DevOps for DatabricksDevOps for Databricks
DevOps for Databricks
 
Azure Synapse Analytics Overview (r1)
Azure Synapse Analytics Overview (r1)Azure Synapse Analytics Overview (r1)
Azure Synapse Analytics Overview (r1)
 
Introduction to Azure Databricks
Introduction to Azure DatabricksIntroduction to Azure Databricks
Introduction to Azure Databricks
 
adb.pdf
adb.pdfadb.pdf
adb.pdf
 
Databricks Overview for MLOps
Databricks Overview for MLOpsDatabricks Overview for MLOps
Databricks Overview for MLOps
 
Generative-AI-in-enterprise-20230615.pdf
Generative-AI-in-enterprise-20230615.pdfGenerative-AI-in-enterprise-20230615.pdf
Generative-AI-in-enterprise-20230615.pdf
 
AzureOpenAI.pptx
AzureOpenAI.pptxAzureOpenAI.pptx
AzureOpenAI.pptx
 
Pythonsevilla2019 - Introduction to MLFlow
Pythonsevilla2019 - Introduction to MLFlowPythonsevilla2019 - Introduction to MLFlow
Pythonsevilla2019 - Introduction to MLFlow
 
Apply MLOps at Scale
Apply MLOps at ScaleApply MLOps at Scale
Apply MLOps at Scale
 

Semelhante a Vertex AI - Unified ML Platform for the entire AI workflow on Google Cloud

Infrastructure Agnostic Machine Learning Workload Deployment
Infrastructure Agnostic Machine Learning Workload DeploymentInfrastructure Agnostic Machine Learning Workload Deployment
Infrastructure Agnostic Machine Learning Workload Deployment
Databricks
 
Borys Rybak “How to make your data smart with Artificial Intelligence and Mac...
Borys Rybak “How to make your data smart with Artificial Intelligence and Mac...Borys Rybak “How to make your data smart with Artificial Intelligence and Mac...
Borys Rybak “How to make your data smart with Artificial Intelligence and Mac...
Lviv Startup Club
 

Semelhante a Vertex AI - Unified ML Platform for the entire AI workflow on Google Cloud (20)

DevBCN Vertex AI - Pipelines for your MLOps workflows
DevBCN Vertex AI - Pipelines for your MLOps workflowsDevBCN Vertex AI - Pipelines for your MLOps workflows
DevBCN Vertex AI - Pipelines for your MLOps workflows
 
MLOps Virtual Event | Building Machine Learning Platforms for the Full Lifecycle
MLOps Virtual Event | Building Machine Learning Platforms for the Full LifecycleMLOps Virtual Event | Building Machine Learning Platforms for the Full Lifecycle
MLOps Virtual Event | Building Machine Learning Platforms for the Full Lifecycle
 
Infrastructure Agnostic Machine Learning Workload Deployment
Infrastructure Agnostic Machine Learning Workload DeploymentInfrastructure Agnostic Machine Learning Workload Deployment
Infrastructure Agnostic Machine Learning Workload Deployment
 
GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...
GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...
GDG Cloud Southlake #16: Priyanka Vergadia: Scalable Data Analytics in Google...
 
Serverless machine learning architectures at Helixa
Serverless machine learning architectures at HelixaServerless machine learning architectures at Helixa
Serverless machine learning architectures at Helixa
 
201906 02 Introduction to AutoML with ML.NET 1.0
201906 02 Introduction to AutoML with ML.NET 1.0201906 02 Introduction to AutoML with ML.NET 1.0
201906 02 Introduction to AutoML with ML.NET 1.0
 
201908 Overview of Automated ML
201908 Overview of Automated ML201908 Overview of Automated ML
201908 Overview of Automated ML
 
Build 2019 Recap
Build 2019 RecapBuild 2019 Recap
Build 2019 Recap
 
.NET Fundamentals and Business Application Development
.NET Fundamentals and Business Application Development.NET Fundamentals and Business Application Development
.NET Fundamentals and Business Application Development
 
Introducing MlFlow: An Open Source Platform for the Machine Learning Lifecycl...
Introducing MlFlow: An Open Source Platform for the Machine Learning Lifecycl...Introducing MlFlow: An Open Source Platform for the Machine Learning Lifecycl...
Introducing MlFlow: An Open Source Platform for the Machine Learning Lifecycl...
 
AML_service.pptx
AML_service.pptxAML_service.pptx
AML_service.pptx
 
NextGenML
NextGenML NextGenML
NextGenML
 
DICE & Cloudify – Quality Big Data Made Easy
DICE & Cloudify – Quality Big Data Made EasyDICE & Cloudify – Quality Big Data Made Easy
DICE & Cloudify – Quality Big Data Made Easy
 
Borys Rybak “How to make your data smart with Artificial Intelligence and Mac...
Borys Rybak “How to make your data smart with Artificial Intelligence and Mac...Borys Rybak “How to make your data smart with Artificial Intelligence and Mac...
Borys Rybak “How to make your data smart with Artificial Intelligence and Mac...
 
ODSC East 2020 Accelerate ML Lifecycle with Kubernetes and Containerized Da...
ODSC East 2020   Accelerate ML Lifecycle with Kubernetes and Containerized Da...ODSC East 2020   Accelerate ML Lifecycle with Kubernetes and Containerized Da...
ODSC East 2020 Accelerate ML Lifecycle with Kubernetes and Containerized Da...
 
Global AI Bootcamp Madrid - Azure Databricks
Global AI Bootcamp Madrid - Azure DatabricksGlobal AI Bootcamp Madrid - Azure Databricks
Global AI Bootcamp Madrid - Azure Databricks
 
Overview of VS2010 and .NET 4.0
Overview of VS2010 and .NET 4.0Overview of VS2010 and .NET 4.0
Overview of VS2010 and .NET 4.0
 
Machine Learning at the Edge
Machine Learning at the EdgeMachine Learning at the Edge
Machine Learning at the Edge
 
201906 04 Overview of Automated ML June 2019
201906 04 Overview of Automated ML June 2019201906 04 Overview of Automated ML June 2019
201906 04 Overview of Automated ML June 2019
 
"The Suitcase" Project Cloud QTR meeting presentation @ Disney/ABC
"The Suitcase"  Project Cloud QTR meeting presentation @ Disney/ABC"The Suitcase"  Project Cloud QTR meeting presentation @ Disney/ABC
"The Suitcase" Project Cloud QTR meeting presentation @ Disney/ABC
 

Mais de Márton Kodok

Mais de Márton Kodok (20)

Gen Apps on Google Cloud PaLM2 and Codey APIs in Action
Gen Apps on Google Cloud PaLM2 and Codey APIs in ActionGen Apps on Google Cloud PaLM2 and Codey APIs in Action
Gen Apps on Google Cloud PaLM2 and Codey APIs in Action
 
Discover BigQuery ML, build your own CREATE MODEL statement
Discover BigQuery ML, build your own CREATE MODEL statementDiscover BigQuery ML, build your own CREATE MODEL statement
Discover BigQuery ML, build your own CREATE MODEL statement
 
Cloud Run - the rise of serverless and containerization
Cloud Run - the rise of serverless and containerizationCloud Run - the rise of serverless and containerization
Cloud Run - the rise of serverless and containerization
 
BigQuery best practices and recommendations to reduce costs with BI Engine, S...
BigQuery best practices and recommendations to reduce costs with BI Engine, S...BigQuery best practices and recommendations to reduce costs with BI Engine, S...
BigQuery best practices and recommendations to reduce costs with BI Engine, S...
 
Cloud Workflows What's new in serverless orchestration and automation
Cloud Workflows What's new in serverless orchestration and automationCloud Workflows What's new in serverless orchestration and automation
Cloud Workflows What's new in serverless orchestration and automation
 
Serverless orchestration and automation with Cloud Workflows
Serverless orchestration and automation with Cloud WorkflowsServerless orchestration and automation with Cloud Workflows
Serverless orchestration and automation with Cloud Workflows
 
Serverless orchestration and automation with Cloud Workflows
Serverless orchestration and automation with Cloud WorkflowsServerless orchestration and automation with Cloud Workflows
Serverless orchestration and automation with Cloud Workflows
 
Serverless orchestration and automation with Cloud Workflows
Serverless orchestration and automation with Cloud WorkflowsServerless orchestration and automation with Cloud Workflows
Serverless orchestration and automation with Cloud Workflows
 
BigdataConference Europe - BigQuery ML
BigdataConference Europe - BigQuery MLBigdataConference Europe - BigQuery ML
BigdataConference Europe - BigQuery ML
 
DevFest Romania 2020 Keynote: Bringing the Cloud to you.
DevFest Romania 2020 Keynote: Bringing the Cloud to you.DevFest Romania 2020 Keynote: Bringing the Cloud to you.
DevFest Romania 2020 Keynote: Bringing the Cloud to you.
 
Applying BigQuery ML on e-commerce data analytics
Applying BigQuery ML on e-commerce data analyticsApplying BigQuery ML on e-commerce data analytics
Applying BigQuery ML on e-commerce data analytics
 
Supercharge your data analytics with BigQuery
Supercharge your data analytics with BigQuerySupercharge your data analytics with BigQuery
Supercharge your data analytics with BigQuery
 
Vibe Koli 2019 - Utazás az egyetem padjaitól a Google Developer Expertig
Vibe Koli 2019 - Utazás az egyetem padjaitól a Google Developer ExpertigVibe Koli 2019 - Utazás az egyetem padjaitól a Google Developer Expertig
Vibe Koli 2019 - Utazás az egyetem padjaitól a Google Developer Expertig
 
BigQuery ML - Machine learning at scale using SQL
BigQuery ML - Machine learning at scale using SQLBigQuery ML - Machine learning at scale using SQL
BigQuery ML - Machine learning at scale using SQL
 
Google Cloud Platform Solutions for DevOps Engineers
Google Cloud Platform Solutions  for DevOps EngineersGoogle Cloud Platform Solutions  for DevOps Engineers
Google Cloud Platform Solutions for DevOps Engineers
 
GDG DevFest Romania - Architecting for the Google Cloud Platform
GDG DevFest Romania - Architecting for the Google Cloud PlatformGDG DevFest Romania - Architecting for the Google Cloud Platform
GDG DevFest Romania - Architecting for the Google Cloud Platform
 
Next18 Extended Targu Mures - Bringing the Cloud to you
Next18 Extended Targu Mures - Bringing the Cloud to youNext18 Extended Targu Mures - Bringing the Cloud to you
Next18 Extended Targu Mures - Bringing the Cloud to you
 
6. DISZ - Webalkalmazások skálázhatósága a Google Cloud Platformon
6. DISZ - Webalkalmazások skálázhatósága  a Google Cloud Platformon6. DISZ - Webalkalmazások skálázhatósága  a Google Cloud Platformon
6. DISZ - Webalkalmazások skálázhatósága a Google Cloud Platformon
 
GCP - A felhőalapú architektúrák és szolgáltatások
GCP - A felhőalapú architektúrák és szolgáltatásokGCP - A felhőalapú architektúrák és szolgáltatások
GCP - A felhőalapú architektúrák és szolgáltatások
 
GDG Heraklion - Architecting for the Google Cloud Platform
GDG Heraklion - Architecting for the Google Cloud PlatformGDG Heraklion - Architecting for the Google Cloud Platform
GDG Heraklion - Architecting for the Google Cloud Platform
 

Último

%+27788225528 love spells in Huntington Beach Psychic Readings, Attraction sp...
%+27788225528 love spells in Huntington Beach Psychic Readings, Attraction sp...%+27788225528 love spells in Huntington Beach Psychic Readings, Attraction sp...
%+27788225528 love spells in Huntington Beach Psychic Readings, Attraction sp...
masabamasaba
 
+971565801893>>SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHAB...
+971565801893>>SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHAB...+971565801893>>SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHAB...
+971565801893>>SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHAB...
Health
 
%+27788225528 love spells in Knoxville Psychic Readings, Attraction spells,Br...
%+27788225528 love spells in Knoxville Psychic Readings, Attraction spells,Br...%+27788225528 love spells in Knoxville Psychic Readings, Attraction spells,Br...
%+27788225528 love spells in Knoxville Psychic Readings, Attraction spells,Br...
masabamasaba
 
Abortion Pills In Pretoria ](+27832195400*)[ 🏥 Women's Abortion Clinic In Pre...
Abortion Pills In Pretoria ](+27832195400*)[ 🏥 Women's Abortion Clinic In Pre...Abortion Pills In Pretoria ](+27832195400*)[ 🏥 Women's Abortion Clinic In Pre...
Abortion Pills In Pretoria ](+27832195400*)[ 🏥 Women's Abortion Clinic In Pre...
Medical / Health Care (+971588192166) Mifepristone and Misoprostol tablets 200mg
 
Love witchcraft +27768521739 Binding love spell in Sandy Springs, GA |psychic...
Love witchcraft +27768521739 Binding love spell in Sandy Springs, GA |psychic...Love witchcraft +27768521739 Binding love spell in Sandy Springs, GA |psychic...
Love witchcraft +27768521739 Binding love spell in Sandy Springs, GA |psychic...
chiefasafspells
 
%+27788225528 love spells in Atlanta Psychic Readings, Attraction spells,Brin...
%+27788225528 love spells in Atlanta Psychic Readings, Attraction spells,Brin...%+27788225528 love spells in Atlanta Psychic Readings, Attraction spells,Brin...
%+27788225528 love spells in Atlanta Psychic Readings, Attraction spells,Brin...
masabamasaba
 

Último (20)

%+27788225528 love spells in Huntington Beach Psychic Readings, Attraction sp...
%+27788225528 love spells in Huntington Beach Psychic Readings, Attraction sp...%+27788225528 love spells in Huntington Beach Psychic Readings, Attraction sp...
%+27788225528 love spells in Huntington Beach Psychic Readings, Attraction sp...
 
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...
 
WSO2Con204 - Hard Rock Presentation - Keynote
WSO2Con204 - Hard Rock Presentation - KeynoteWSO2Con204 - Hard Rock Presentation - Keynote
WSO2Con204 - Hard Rock Presentation - Keynote
 
+971565801893>>SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHAB...
+971565801893>>SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHAB...+971565801893>>SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHAB...
+971565801893>>SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHAB...
 
WSO2CON 2024 - Cloud Native Middleware: Domain-Driven Design, Cell-Based Arch...
WSO2CON 2024 - Cloud Native Middleware: Domain-Driven Design, Cell-Based Arch...WSO2CON 2024 - Cloud Native Middleware: Domain-Driven Design, Cell-Based Arch...
WSO2CON 2024 - Cloud Native Middleware: Domain-Driven Design, Cell-Based Arch...
 
%+27788225528 love spells in Knoxville Psychic Readings, Attraction spells,Br...
%+27788225528 love spells in Knoxville Psychic Readings, Attraction spells,Br...%+27788225528 love spells in Knoxville Psychic Readings, Attraction spells,Br...
%+27788225528 love spells in Knoxville Psychic Readings, Attraction spells,Br...
 
Abortion Pills In Pretoria ](+27832195400*)[ 🏥 Women's Abortion Clinic In Pre...
Abortion Pills In Pretoria ](+27832195400*)[ 🏥 Women's Abortion Clinic In Pre...Abortion Pills In Pretoria ](+27832195400*)[ 🏥 Women's Abortion Clinic In Pre...
Abortion Pills In Pretoria ](+27832195400*)[ 🏥 Women's Abortion Clinic In Pre...
 
Crypto Cloud Review - How To Earn Up To $500 Per DAY Of Bitcoin 100% On AutoP...
Crypto Cloud Review - How To Earn Up To $500 Per DAY Of Bitcoin 100% On AutoP...Crypto Cloud Review - How To Earn Up To $500 Per DAY Of Bitcoin 100% On AutoP...
Crypto Cloud Review - How To Earn Up To $500 Per DAY Of Bitcoin 100% On AutoP...
 
%in Benoni+277-882-255-28 abortion pills for sale in Benoni
%in Benoni+277-882-255-28 abortion pills for sale in Benoni%in Benoni+277-882-255-28 abortion pills for sale in Benoni
%in Benoni+277-882-255-28 abortion pills for sale in Benoni
 
WSO2CON 2024 - Does Open Source Still Matter?
WSO2CON 2024 - Does Open Source Still Matter?WSO2CON 2024 - Does Open Source Still Matter?
WSO2CON 2024 - Does Open Source Still Matter?
 
Architecture decision records - How not to get lost in the past
Architecture decision records - How not to get lost in the pastArchitecture decision records - How not to get lost in the past
Architecture decision records - How not to get lost in the past
 
WSO2CON 2024 - WSO2's Digital Transformation Journey with Choreo: A Platforml...
WSO2CON 2024 - WSO2's Digital Transformation Journey with Choreo: A Platforml...WSO2CON 2024 - WSO2's Digital Transformation Journey with Choreo: A Platforml...
WSO2CON 2024 - WSO2's Digital Transformation Journey with Choreo: A Platforml...
 
MarTech Trend 2024 Book : Marketing Technology Trends (2024 Edition) How Data...
MarTech Trend 2024 Book : Marketing Technology Trends (2024 Edition) How Data...MarTech Trend 2024 Book : Marketing Technology Trends (2024 Edition) How Data...
MarTech Trend 2024 Book : Marketing Technology Trends (2024 Edition) How Data...
 
Love witchcraft +27768521739 Binding love spell in Sandy Springs, GA |psychic...
Love witchcraft +27768521739 Binding love spell in Sandy Springs, GA |psychic...Love witchcraft +27768521739 Binding love spell in Sandy Springs, GA |psychic...
Love witchcraft +27768521739 Binding love spell in Sandy Springs, GA |psychic...
 
%in kempton park+277-882-255-28 abortion pills for sale in kempton park
%in kempton park+277-882-255-28 abortion pills for sale in kempton park %in kempton park+277-882-255-28 abortion pills for sale in kempton park
%in kempton park+277-882-255-28 abortion pills for sale in kempton park
 
%in Rustenburg+277-882-255-28 abortion pills for sale in Rustenburg
%in Rustenburg+277-882-255-28 abortion pills for sale in Rustenburg%in Rustenburg+277-882-255-28 abortion pills for sale in Rustenburg
%in Rustenburg+277-882-255-28 abortion pills for sale in Rustenburg
 
%+27788225528 love spells in Atlanta Psychic Readings, Attraction spells,Brin...
%+27788225528 love spells in Atlanta Psychic Readings, Attraction spells,Brin...%+27788225528 love spells in Atlanta Psychic Readings, Attraction spells,Brin...
%+27788225528 love spells in Atlanta Psychic Readings, Attraction spells,Brin...
 
WSO2CON 2024 - API Management Usage at La Poste and Its Impact on Business an...
WSO2CON 2024 - API Management Usage at La Poste and Its Impact on Business an...WSO2CON 2024 - API Management Usage at La Poste and Its Impact on Business an...
WSO2CON 2024 - API Management Usage at La Poste and Its Impact on Business an...
 
WSO2CON 2024 Slides - Open Source to SaaS
WSO2CON 2024 Slides - Open Source to SaaSWSO2CON 2024 Slides - Open Source to SaaS
WSO2CON 2024 Slides - Open Source to SaaS
 
%in kaalfontein+277-882-255-28 abortion pills for sale in kaalfontein
%in kaalfontein+277-882-255-28 abortion pills for sale in kaalfontein%in kaalfontein+277-882-255-28 abortion pills for sale in kaalfontein
%in kaalfontein+277-882-255-28 abortion pills for sale in kaalfontein
 

Vertex AI - Unified ML Platform for the entire AI workflow on Google Cloud

  • 1. Vertex AI Unified ML Platform for the entire AI workflow on Google Cloud DevFest Season, November 2021 Márton Kodok Google Developer Expert at REEA.net
  • 2. ● Among the Top3 romanians on Stackoverflow 195k reputation ● Google Developer Expert on Cloud technologies ● Crafting Web/Mobile backends at REEA.net ● BigQuery + Redis database engine expert Slideshare: martonkodok Articles: martonkodok.medium.com Twitter: @martonkodok StackOverflow: pentium10 GitHub: pentium10 Vertex AI - Unified ML Platform @martonkodok About me
  • 3. 1. What is Vertex AI 2. Gather, Import & label datasets 3. Build, train and deploy ML solutions 4. Manage your models with confidence 5. Using Pipelines throughout your ML workflow 6. Adapting to changes of data 7. Conclusions Agenda Vertex AI - Unified ML Platform @martonkodok
  • 4. “VertexAI is a managed ML platform for practitioners to accelerate experiments and deploy AI models. Vertex AI - Unified ML Platform @martonkodok
  • 5. Where does VertexAI fit in? Application Servers Vertex AI Desktop client Mobile client End-to-end platform for ML model development and deployment Backend Vertex AI - Unified ML Platform @martonkodok Application Logic
  • 6. VertexAI is a unified MLOps platform Vertex AI - Unified ML Platform @martonkodok Operational Model Programming Model No Infra Management Managed Security Pay only for usage Model-as-a-service oriented Streamlined model development Open SDKs, integrates with ML frameworks
  • 7. What’s included in VertexAI? Vertex AI - Unified ML Platform @martonkodok Data Labeling AutoML DL Environment (DL VM + DL Container) Prediction Feature Store Training Experiments Data Readiness Feature Engineering Training/ HP-Tuning Model Monitoring Model serving Understanding/ Tuning Edge Model Management Notebooks Pipelines (Orchestration) Explainable AI Hybrid AI Model Monitoring Metadata Vision Translation Tables Language Video AI Accelerators Models Datasets Custom Models Containers Python Endpoints Vizier Optimization
  • 8. VertexAI supports... Vertex AI - Unified ML Platform @martonkodok UI based model development # Define job job = aiplatform.AutoMLTabularTrainingJob( display_name='price-predict-training', optimization_prediction_type='regression' ) # Run job model = job.run( dataset=ds, target_column='median_house_value', model_display_name='house-value-prediction', ) Code-based model development
  • 9. Using Vertex AI throughout your ML workflow Vertex AI - Unified ML Platform @martonkodok Gather data Train model Scalably deploy model Evaluate, monitor, retrain
  • 11. VertexAI: Gather, Import & label datasets at scale Vertex AI - Unified ML Platform @martonkodok Data Labeling AutoML DL Environment (DL VM + DL Container) Prediction Feature Store Training Experiments Data Readiness Feature Engineering Training/ HP-Tuning Model Monitoring Model serving Understanding/ Tuning Edge Model Management Notebooks Pipelines (Orchestration) Explainable AI Hybrid AI Model Monitoring Metadata Vision Translation Tables Language Video AI Accelerators Models Datasets Custom Models Containers Python Endpoints Vizier Optimization
  • 12. “ Datasets Vertex AI - Unified ML Platform @martonkodok
  • 13. Datasets Vertex AI - Unified ML Platform @martonkodok Vertex AI datasets (fully managed) • Fully serverless • Region based • Free to store • In tandem with AutoML managed models Custom Datasets Cloud Storage, BigQuery or on Internet Accessing managed dataset from your app: - JSONL (default) - CSV or BigQuery stream
  • 14. “ VertexAI Managed Datasets + Objectives (AutoML*) Vertex AI - Unified ML Platform @martonkodok * legacy name, previous generation naming from AI Platform
  • 15. - Regression/classification - Forecasting - Single-label classification - Multi-label classification - Text entity extraction - Text sentiment analysis - Video action recognition - Video classifications for entire video, shots, frames - Video object tracking Vertex AI: Managed datasets + objectives Vertex AI - Unified ML Platform @martonkodok Image Tabular Text Video - Single-label classification - Multi-label classification - Image object detection - Image segmentation
  • 16. Vertex AI: Managed dataset + objectives Vertex AI - Unified ML Platform @martonkodok Image
  • 17. Vertex AI: Managed dataset + objectives Vertex AI - Unified ML Platform @martonkodok Tabular
  • 18. Vertex AI: Managed dataset + objectives Vertex AI - Unified ML Platform @martonkodok Text
  • 19. Vertex AI: Managed dataset + objectives Vertex AI - Unified ML Platform @martonkodok Video
  • 20. Data labeling + Feature Store Vertex AI - Unified ML Platform @martonkodok Data labeling (fully managed) • Create data labeling, annotation tasks • Use human labelers • Use Google’s labeler workforce • Use your own workforce Feature Store (fully managed) • Centralized repository for organizing, storing, and serving ML features • Organization can efficiently share, discover, re-use features • Data Model: Entity Type -> Feature • Ingest data from BigQuery or Cloud Storage Pro: point-in-time lookup from time series
  • 22. 2. Train models Vertex AI - Unified ML Platform @martonkodok Gather data Train model Scalably deploy model Evaluate, monitor, retrain
  • 23. VertexAI: Build, train & deploy models at scale Vertex AI - Unified ML Platform @martonkodok Data Labeling AutoML DL Environment (DL VM + DL Container) Prediction Feature Store Training Experiments Data Readiness Feature Engineering Training/ HP-Tuning Model Monitoring Model serving Understanding/ Tuning Edge Model Management Notebooks Pipelines (Orchestration) Explainable AI Hybrid AI Model Monitoring Metadata Vision Translation Tables Language Video AI Accelerators Models Datasets Custom Models Containers Python Endpoints Vizier Optimization
  • 24. 1. AutoML out-of-the box training integration No-code solution. You must target one of the AutoML’s predefined objectives. 2. Custom Training - run your own training applications in the cloud Train with one of the Google’s pre-builtcontainers or useyourown. 3. Hyperparameter tuning jobs - searchesforbestcombination of hyperparameter values by optimizing values across a series of trials. Available for custom training. Your training app must adhere to accepting Vertex AI parameters. You need to report metrics to Vertex AI. Training https://cloud.google.com/vertex-ai/docs/training/using-hyperparameter-tuning @martonkodok
  • 25. Pre-built containers for custom training https://cloud.google.com/vertex-ai/docs/training/pre-built-containers @martonkodok Tensorflow ML Framework version 1.15, 2.1-2.4 use with Cuda 11.x GPU scikit-learn ML Framework version 0.23 No GPUs PyTorch ML Framework version 1.4 - 1.7 use with Cuda 11.x GPU XGBoost ML Framework version 1.1 No GPUs
  • 26. 3. Deploying models Vertex AI - Unified ML Platform @martonkodok Gather data Train model Scalably deploy model Evaluate, monitor, retrain
  • 27. “ You can deploy models on VertexAI and get a HTTPs Endpointto serve predictions rapidly and reliably. Vertex AI - Unified ML Platform @martonkodok
  • 28. 1. Deploy a model and get aREST endpointto serve predictions realtime or batched 2. You can use models whetherornotthemodelwastrained on Vertex AI. 3. Specify a prediction traffic split in your endpoint. 4. VPC Private Network option for custom-trained models/tabular models Vertex AI: Endpoints Vertex AI - Unified ML Platform @martonkodok Vertex AI Endpoints Backend Prediction deploy REST
  • 30. VertexAI: Manage your models with confidence Vertex AI - Unified ML Platform @martonkodok Data Labeling AutoML DL Environment (DL VM + DL Container) Prediction Feature Store Training Experiments Data Readiness Feature Engineering Training/ HP-Tuning Model Monitoring Model serving Understanding/ Tuning Edge Model Management Notebooks Pipelines (Orchestration) Explainable AI Hybrid AI Model Monitoring Metadata Vision Translation Tables Language Video AI Accelerators Models Datasets Custom Models Containers Python Endpoints Vizier Optimization
  • 31. VertexAI: Manage your models with confidence Vertex AI - Unified ML Platform @martonkodok Data Labeling AutoML DL Environment (DL VM + DL Container) Prediction Feature Store Training Experiments Data Readiness Feature Engineering Training/ HP-Tuning Model Monitoring Model serving Understanding/ Tuning Edge Model Management Notebooks Pipelines (Orchestration) Explainable AI Hybrid AI Model Monitoring Metadata Vision Translation Tables Language Video AI Accelerators Models Datasets Custom Models Containers Python Endpoints Vizier Optimization
  • 32. Explainable AI Vertex AI - Unified ML Platform @martonkodok Explainable AI • Interpret predictions made by ML models • Receive a score explaining how much each factor contributed to the model predictions • What-If Tool lets you investigate model behavior at a glance AI Explanations samples Github: GoogleCloudPlatform/ai-platform-samples - Training, deploying, and explaining a tabular data model - Training, deploying, and explaining an image model Limitations: doesn’t work well on low-contrast, X-rays, one shade, panoramas, very tall, very wide images.
  • 33. Explainable AI: What-if Tool (pair-code.github.io/what-if-tool) Vertex AI - Unified ML Platform @martonkodok What-if Tool • Model probing, from within any workflow • test performance in hypothetical situations • analyze the importance of different data features • visualize model behavior across multiple models and subsets of data Tutorials, demos onpair-code.github.io/what-if-tool • Available on many platforms (TensorBoard, Jupyter, Colaboratory, Vertex AI) • Supports what-if Analyses (explore counterfactuals, fairness measures, partial dependence plots) • Visualizes Model Performances (threshold simulation, up to 2 model comparison, dataset summary statistics)
  • 35. VertexAI: Pipelines - Orchestrate your model Vertex AI - Unified ML Platform @martonkodok Data Labeling AutoML DL Environment (DL VM + DL Container) Prediction Feature Store Training Experiments Data Readiness Feature Engineering Training/ HP-Tuning Model Monitoring Model serving Understanding/ Tuning Edge Model Management Notebooks Pipelines (Orchestration) Explainable AI Hybrid AI Model Monitoring Metadata Vision Translation Tables Language Video AI Accelerators Models Datasets Custom Models Containers Python Endpoints Vizier Optimization
  • 36. “ Why are MLpipelines useful? Vertex AI: Pipelines for your MLOps workflows @martonkodok
  • 37. 1. Orchestrate ML workflow steps as a process. We no longer handle all data gathering, model training, tuning, evaluation, deployment as a monolith. 2. Adopt MLOps for production models. We need a repeatable, verifiable, and automatic process for making any change to a production model. 3. Develop steps independently -as you scale out, enables you to share your ML workflow with others on your team, so they can run it, and contribute code. Enablesyoutotracktheinputandoutputfromeach stepinareproducibleway. Why are ML pipelines useful? @martonkodok
  • 38. “Continuousdelivery and automationpipelines for machinelearning systems. Vertex AI: Pipelines for your MLOps workflows @martonkodok What is MLOps?
  • 39. MLOps level 0: Manual process MLOps level 1: ML pipeline automation MLOps level 2: CI/CD pipeline automation Levelsofautomation defines maturity of theMLprocess @martonkodok
  • 40. MLOps level 0: Manual process - Process for building and deploying ML models is entirely manual. Infrequent release iterations. No CI, No CD. Disconnection between ML and operations. MLOps level 1: ML pipeline automation - Continuous training of the model by automating the ML pipeline; achieve continuous delivery of model prediction service. New pipelines mostly based on new data. MLOps level 2: CI/CD pipeline automation -iteratively try out new ML algorithms and new modeling where the experiment steps are orchestrated. The output of this stage is the source code of the ML pipeline steps that are then pushed to a source repository. Build source. Run test. Output is pipeline. Levelsofautomation defines maturity of theMLprocess @martonkodok
  • 41. MLOps level 2: CI/CDpipelineautomation @martonkodok
  • 43. Vertex AI: Pipelines Vertex AI - Unified ML Platform Source: Piero Esposito https://github.com/piEsposito/vertex-ai-tutorials
  • 44. Pipeline SDK: Condition Vertex AI: Pipelines for your MLOps workflows automl_tabular_classification_beans.ipynb github.com/GoogleCloudPlatform/vertex-ai-samples
  • 45. 1. Use of the Google Cloud Pipeline Components, which support easy access to Vertex AI services 2. Custom Components - function that compiles to a task ‘factory’ function that can be used by pipelines 3. No more Kubeflow Pipelines that must be deployed on a Kubernetes Cluster. 4. Sharing component specifications - the YAML format allows the component to be put under version control and shared with others, or be used by other pipelines by calling the load_from_url function. 5. Leveraging Pipeline step caching to develop and debug 6. Vertex AI Metadata service + Artifacts Lineage tracking - inverse of pipeline DAG Developer friendly components @martonkodok
  • 47. Automatic CI / CD Perspective with GCP Services Vertex AI: Pipelines for your MLOps workflows @martonkodok Eventarc • Detect changes on data • React to events from Cloud services • Handle events on Cloud Workflows, Cloud Functions, Cloud Run • Reuse pipeline spec.json from GCS • Trigger Vertex AI pipeline • Detect changes in codebase • Build pipeline • Pipeline spec.json to Cloud Storage • Image to Cloud Registry • Trigger Vertex AI pipeline Cloud Build Cloud Scheduler • Poll for changes of any data • Launch based on schedule • In tandem with Cloud Workflows • Trigger Vertex AI pipeline
  • 48. Conclusion Vertex AI - Unified ML Platform @martonkodok
  • 49. 1. Build with the groundbreaking ML tools that power Google 2. Approachable from the non-ML developer perspective (AutoML, managed models, training) 3. Ease the life of a data scientist/ML (has feature store, managed datasets, endpoints, notebooks) 4. End-to-end integration for data and AI with build pipelines that outperform and solve complex ML tasks 5. GitOps-style continuous delivery with Cloud Build 6. Explainable AI and TensorBoard to visualize and track ML experiments Vertex AI: Enhanced developer experience Vertex AI: Pipelines for your MLOps workflows @martonkodok
  • 50. Thank you. Slides available on: slideshare.net/martonkodok Reea.net - Integrated web solutions driven by creativity to deliver projects.