Kruize-HPO-meetup-Aug2022.pdf

•

0 likes•3 views

S

Choosing performance tuables is not that simple

Choosing performance
tunables is not that simple
Chandrakala Subramanyam @chandra25ms
Kusuma Chalasani @kusumach

2
Agenda
➢ Challenges / Motivation
➢ Introduction to Bayesian Optimization / HPO
➢ HPO Use cases
➢ HPO Demo
➢ Experiment results

SRE
Optimize App Stack - Challenges
3

4
Choices Are Never This simple!

5
Building an ML Model?
Fix the Hyperparameters first!
⚫ Model architecture
⚫ Learning rate
⚫ Number of epochs
⚫ Number of branches in a decision tree
⚫ Number of clusters in a clustering algorithm

6
Fixing Hyperparameters - Options
⚫ Manual Search
⚫ Random Search
⚫ Grid Search
⚫ Automated Hyperparameter Tuning - Bayesian Optimization
⚫ ...

Ref: https://distill.pub/2020/bayesian-optimization/
Bayesian Optimization
7

HPO Architecture
8

9
HPO - Usecases
Native
Containers
Kubernetes
Kruize HPO
OS settings tuning
Resource Usage
Microservices App
tuning

HPO Demo
10

Be careful
what you
wish for !
11

Kruize HPO vs Default Config
[Obj Fn = maximize(1.25*Throughput / 1.5*Response_time / 0.25*Max_response_time),
Fixed Resources]
Default
14.21 ms
2.39 ms
12

Kruize HPO vs Default Config
[Obj Fn = maximize(1.25*Throughput / 1.5*Response_time / 0.25*Max_Response_Time),
Fixed Resources]
83% better response time 1.3% better throughput
13

Interested in HPO?
⚫ HPO - https://github.com/kruize/hpo
⚫ Kruize Demos - https://github.com/kruize/kruize-demos
Call for collaboration !
Kruize Slack
kruize@redhat.com
14

Questions

More Related Content

Similar to Kruize-HPO-meetup-Aug2022.pdf

On a dartboard, the triple 20 is the square that gives the most points for a single dart throw, which is why most players always try to hit the triple 20. Of course, developers and engineers also try to get the most out of their applications and infrastructure with the least amount of effort. As with darts, precision and efficiency are key. In our webinar ”Triple 20 IT - How to reduce costs on target while increasing speed and quality” you can learn how to efficiently reduce cloud costs with the help of some counterintuitive approaches and the latest advances in machine-learning. By attending this webinar, you’ll learn how to: - Reduce costs and also deliver all features in the backlog on time, - Stay agile despite cost-cutting pressure, - Follow counterintuitive approaches instead of wasting time on unhelpful “reduce infrastructure and licensing costs” tips, - Automate Kubernetes resource efficiency as you scale to day 2 operations, reducing cloud costs and improving application performance without the trade-offs.

Triple 20 IT – How to reduce costs on target while increasing speed and quality

Triple 20 IT – How to reduce costs on target while increasing speed and quality

Triple 20 IT – How to reduce costs on target while increasing speed and quality

Optimizing Java

Optimizing Java

Optimizing Java

When is a project ready for Software Automation_NEW

When is a project ready for Software Automation_NEW

When is a project ready for Software Automation_NEW

Mike Christesen

PCM_PDI_BestPractices.pdf

PCM_PDI_BestPractices.pdf

PCM_PDI_BestPractices.pdf

scale_perf_best_practices

scale_perf_best_practices

scale_perf_best_practices

Synergy 2015 Session Slides: SYN408 XenDesktop 7.6 Architecture - Dealing Wit...

Synergy 2015 Session Slides: SYN408 XenDesktop 7.6 Architecture - Dealing Wit...

Synergy 2015 Session Slides: SYN408 XenDesktop 7.6 Architecture - Dealing Wit...

Hadoop-DS: Which SQL-on-Hadoop Rules the Herd

Hadoop-DS: Which SQL-on-Hadoop Rules the Herd

Hadoop-DS: Which SQL-on-Hadoop Rules the Herd

Building on the "Best Practices for Hyperparameter Tuning with MLflow" talk, we will present advanced topics in HPO for deep learning, including early stopping, multi-metric optimization, and robust optimization. We will then discuss implementations using open source tools. Finally, we will discuss how we can leverage MLflow with these tools and techniques to analyze the performance of our models.

Advanced Hyperparameter Optimization for Deep Learning with MLflow

Advanced Hyperparameter Optimization for Deep Learning with MLflow

Advanced Hyperparameter Optimization for Deep Learning with MLflow

Cutting edge hyperparameter tuning made simple with ray tune

Cutting edge hyperparameter tuning made simple with ray tune

Cutting edge hyperparameter tuning made simple with ray tune

For the full video of this presentation, please visit: https://www.edge-ai-vision.com/2021/09/a-practical-guide-to-implementing-ml-on-embedded-devices-a-presentation-from-the-chamberlain-group/ Nathan Kopp, Principal Software Architect for Video Systems at the Chamberlain Group, presents the “Practical Guide to Implementing ML on Embedded Devices” tutorial at the May 2021 Embedded Vision Summit. Deploying machine learning onto edge devices requires many choices and trade-offs. Fortunately, processor designers are adding inference-enhancing instructions and architectures to even the lowest cost MCUs, tools developers are constantly discovering optimizations that extract a little more performance out of existing hardware, and ML researchers are refactoring the math to achieve better accuracy using faster operations and fewer parameters. In this presentation, Kopp takes a high-level look at what is involved in running a DNN model on existing edge devices, exploring some of the evolving tools and methods that are finally making this dream a reality. He also takes a quick look at a practical example of running a CNN object detector on low-compute hardware.

“A Practical Guide to Implementing ML on Embedded Devices,” a Presentation fr...

“A Practical Guide to Implementing ML on Embedded Devices,” a Presentation fr...

“A Practical Guide to Implementing ML on Embedded Devices,” a Presentation fr...

Edge AI and Vision Alliance

Optimisation and prediction are two goals that are often confused. This video explains the differences and when each approach should be taken. This presentation has been produced by the Tesseract Academy (http://tesseract.academy), a company that educates decision makers in deep technical topics such as data science, analytics, machine learning, AI and blockchain. For the video version please visit: https://youtu.be/7oJRaxrIqNw

Optimisation vs prediction

Optimisation vs prediction

Optimisation vs prediction

Dr. Stylianos Kampakis

Q2.12: Benchmarking Techniques

Q2.12: Benchmarking Techniques

Q2.12: Benchmarking Techniques

Software Testing: Application And Script Independent Automation Framework: Th...

Software Testing: Application And Script Independent Automation Framework: Th...

Software Testing: Application And Script Independent Automation Framework: Th...

Humana strives to help the communities we serve and our individual members achieve their best health – no small task in the past year! We had the opportunity to rethink our existing operations and reimagine what a collaborative ML platform for hundreds of data scientists might look like. The primary goal of our ML Platform, named FlorenceAI, is to automate and accelerate the delivery lifecycle of data science solutions at scale. In this presentation, we will walk through an end-to-end example of how to build a model at scale on FlorenceAI and deploy it to production. Tools highlighted include Azure Databricks, MLFlow, AppInsights, and Azure Data Factory. We will employ slides, notebooks and code snippets covering problem framing and design, initial feature selection, model design and experimentation, and a framework of centralized production code to streamline implementation. Hundreds of data scientists now use our feature store that has tens of thousands of features refreshed in daily and monthly cadences across several years of historical data. We already have dozens of models in production and also daily provide fresh insights for our Enterprise Clinical Operating Model. Each day, billions of rows of data are generated to give us timely information. We already have examples of teams operating orders of magnitude faster and at a scale not within reach using fixed on-premise resources. Given rapid adoption from a dozen pilot users to over 100 MAU in the first 5 months, we will also share some anecodotes about key early wins created by the platform. We want FlorenceAI to enable Humana’s data scientists to focus their efforts where they add the most value so we can continue to deliver high-quality solutions that remain fresh, relevant and fair in an ever changing world.

FlorenceAI: Reinventing Data Science at Humana

FlorenceAI: Reinventing Data Science at Humana

FlorenceAI: Reinventing Data Science at Humana

Many industries use multiple similar Work Center / resources to manufacture a product. In industries like Pharmaceutical, FMCG, chemical industry, there are many Material which can be produced on different work center which have the same Capacity. For example, one product can be pack on 9 different packing line. So the main question arises, should I create 9 routing/Master Recipe and 9 production versions? OR Is there any other effective way of dealing with this situation? In this document, we will follow the best approach, i.e. Resource Selection through classification. We will need to create Characteristic, Class, need to update Work Center, Master Recipe in order to achieve this solution.

Multiple Work Center/Resource Selection and Classification in Master Recipe f...

Multiple Work Center/Resource Selection and Classification in Master Recipe f...

Multiple Work Center/Resource Selection and Classification in Master Recipe f...

Performance and optimization CakeFest 2014

Performance and optimization CakeFest 2014

Performance and optimization CakeFest 2014

Query Execution Time and Query Optimization.

Query Execution Time and Query Optimization.

Query Execution Time and Query Optimization.

Radhe Krishna Rajan

You know what your query does, but do you know how it does it? Do you know what type of resources your query uses? This session covered these questions and more as we walked through reading execution plans. We saw how SQL breaks down the execution of your query and what each step tells us about the overall query. These slides provide the additional resources that go into the depth we couldn't get into in the session.

Before you optimize: Understanding Execution Plans

Before you optimize: Understanding Execution Plans

Before you optimize: Understanding Execution Plans

Bayesian Global Optimization

Bayesian Global Optimization

Bayesian Global Optimization

Amazon Web Services

Benchmarking PyCon AU 2011 v0

Benchmarking PyCon AU 2011 v0

Benchmarking PyCon AU 2011 v0

Tennessee Leeuwenburg

Similar to Kruize-HPO-meetup-Aug2022.pdf (20)

Triple 20 IT – How to reduce costs on target while increasing speed and quality

Triple 20 IT – How to reduce costs on target while increasing speed and quality

Triple 20 IT – How to reduce costs on target while increasing speed and quality

Optimizing Java

Optimizing Java

Optimizing Java

When is a project ready for Software Automation_NEW

When is a project ready for Software Automation_NEW

When is a project ready for Software Automation_NEW

PCM_PDI_BestPractices.pdf

PCM_PDI_BestPractices.pdf

PCM_PDI_BestPractices.pdf

scale_perf_best_practices

scale_perf_best_practices

scale_perf_best_practices

Synergy 2015 Session Slides: SYN408 XenDesktop 7.6 Architecture - Dealing Wit...

Synergy 2015 Session Slides: SYN408 XenDesktop 7.6 Architecture - Dealing Wit...

Synergy 2015 Session Slides: SYN408 XenDesktop 7.6 Architecture - Dealing Wit...

Hadoop-DS: Which SQL-on-Hadoop Rules the Herd

Hadoop-DS: Which SQL-on-Hadoop Rules the Herd

Hadoop-DS: Which SQL-on-Hadoop Rules the Herd

Advanced Hyperparameter Optimization for Deep Learning with MLflow

Advanced Hyperparameter Optimization for Deep Learning with MLflow

Advanced Hyperparameter Optimization for Deep Learning with MLflow

Cutting edge hyperparameter tuning made simple with ray tune

Cutting edge hyperparameter tuning made simple with ray tune

Cutting edge hyperparameter tuning made simple with ray tune

“A Practical Guide to Implementing ML on Embedded Devices,” a Presentation fr...

“A Practical Guide to Implementing ML on Embedded Devices,” a Presentation fr...

“A Practical Guide to Implementing ML on Embedded Devices,” a Presentation fr...

Optimisation vs prediction

Optimisation vs prediction

Optimisation vs prediction

Q2.12: Benchmarking Techniques

Q2.12: Benchmarking Techniques

Q2.12: Benchmarking Techniques

Software Testing: Application And Script Independent Automation Framework: Th...

Software Testing: Application And Script Independent Automation Framework: Th...

Software Testing: Application And Script Independent Automation Framework: Th...

FlorenceAI: Reinventing Data Science at Humana

FlorenceAI: Reinventing Data Science at Humana

FlorenceAI: Reinventing Data Science at Humana

Multiple Work Center/Resource Selection and Classification in Master Recipe f...

Multiple Work Center/Resource Selection and Classification in Master Recipe f...

Multiple Work Center/Resource Selection and Classification in Master Recipe f...

Performance and optimization CakeFest 2014

Performance and optimization CakeFest 2014

Performance and optimization CakeFest 2014

Query Execution Time and Query Optimization.

Query Execution Time and Query Optimization.

Query Execution Time and Query Optimization.

Before you optimize: Understanding Execution Plans

Before you optimize: Understanding Execution Plans

Before you optimize: Understanding Execution Plans

Bayesian Global Optimization

Bayesian Global Optimization

Bayesian Global Optimization

Benchmarking PyCon AU 2011 v0

Benchmarking PyCon AU 2011 v0

Benchmarking PyCon AU 2011 v0

Recently uploaded

Talk given at NewCrafts Paris Conference 2024 Abstract: Have you ever observed quality problems with what your team is delivering? Have you ever felt your team is learning about the actual status quo late, too late to include these insights in your next steps? Have you ever wondered how to influence change when you’re “just” a specialist on the team, maybe even felt powerless? I’ve mostly worked with people who never experienced effective collaboration across roles and specialties. How to get such a team to move toward testing throughout the delivery workflow from idea to production, on all kinds of levels, for all parts of the product, by the whole team? How to consider quality from the start and have it continuously inform our decision-making? Over the years, I’ve built a repertoire of tactics to help teams transform toward holistic testing and quality. They range from meeting people where they are to creating guidelines, pairing and many others. It’s a toolbox of approaches to try out and find the sweet spot where people start moving toward better directions together. These tactics are fallible, not all of them work in all contexts, and yet they have helped my teams get to a better place. Join this talk to get inspired to take things into your hands. Learn what you can try yourself in your context to help your team make the first steps on your journey towards holistic testing and quality. Key learnings: - Learn how you can influence team culture, no matter your role or experience level - Add team transformation tactics to your toolbox of approaches to try - Understand the many ways we can move towards holistic testing and quality

Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...

Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...

Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...

Nearly a decade ago, a small group of menders (and one maker) found community with each other at a conference and Legacy Code Rocks was born. Since then, the Legacy Code Rocks community has grown to over 1000 people, 150 podcast episodes, a weekly meetup and of course, MenderCon. In this talk, Andrea Goulet, the maker who kindled the spark, will take us back to those early days and reflect on how the bedrock of the Legacy Code Rocks community has always been creating a sense of belonging. Compassion and empathy are at the core of being a mender. Mending is caring deeply, seeing potential, and being willing to work through the frustration of making things better one small step at a time. This commitment to helping others feel seen instead of shame is why the Legacy Code Rocks community is so special — and it’s exactly why having the heart of the mender is what our world needs now more than ever.

Community is Just as Important as Code by Andrea Goulet

Community is Just as Important as Code by Andrea Goulet

Community is Just as Important as Code by Andrea Goulet

Anypoint Code Builder - Munich MuleSoft Meetup - 16th May 2024

Anypoint Code Builder - Munich MuleSoft Meetup - 16th May 2024

Anypoint Code Builder - Munich MuleSoft Meetup - 16th May 2024

MulesoftMunichMeetup

Workshop: Enabling GenAI Breakthroughs with Knowledge Graphs - GraphSummit Milan

Workshop: Enabling GenAI Breakthroughs with Knowledge Graphs - GraphSummit Milan

Workshop: Enabling GenAI Breakthroughs with Knowledge Graphs - GraphSummit Milan

A Deep Dive into Secure Product Development Frameworks.pdf

A Deep Dive into Secure Product Development Frameworks.pdf

A Deep Dive into Secure Product Development Frameworks.pdf

Dive into the innovative world of Rapidoform and discover how our dynamic platform is reshaping the landscape of form building. In this presentation, we unveil the power of Rapidoform's intuitive interface, interactive question types, and seamless design customization options. Explore real-world use cases and success stories, showcasing how organizations are leveraging Rapidoform to captivate audiences and gather valuable insights. From customer feedback surveys to event registrations, Rapidoform offers a versatile toolkit to elevate engagement and streamline data collection. Join us as we embark on a journey to reimagine forms and unlock new possibilities with Rapidoform.

Rapidoform for Modern Form Building and Insights

Rapidoform for Modern Form Building and Insights

Rapidoform for Modern Form Building and Insights

Azure Native Qumulo scales elastically for common High Performance Compute (HPC) workloads based on application requirements for: Financial Services, Automotive, Genomics / Life Sciences, Media and Entertainment, Energy, Oil and Gas, and more. Performance can be increased (and elastically decreased) much higher than the examples shown here. These slides offer a glimpse into ANQ's HPC capabilities, although at a smaller scale. We invite YOU to do your own testing (with a free ANQ trial) and work with us to test your HPC workloads in Azure.

AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdf

AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdf

AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdf

Workshop 1. Architecting Innovative Graph Applications Join this hands-on workshop for beginners led by Neo4j experts guiding you to systematically uncover contextual intelligence. Using a real-life dataset we will build step-by-step a graph solution; from building the graph data model to running queries and data visualization. The approach will be applicable across multiple use cases and industries.

Workshop - Architecting Innovative Graph Applications- GraphSummit Milan

Workshop - Architecting Innovative Graph Applications- GraphSummit Milan

Workshop - Architecting Innovative Graph Applications- GraphSummit Milan

Abortion Pill Prices Jozini ](+27832195400*)[ 🏥 Women's Abortion Clinic in Jo...

Abortion Pill Prices Jozini ](+27832195400*)[ 🏥 Women's Abortion Clinic in Jo...

Abortion Pill Prices Jozini ](+27832195400*)[ 🏥 Women's Abortion Clinic in Jo...

Medical / Health Care (+971588192166) Mifepristone and Misoprostol tablets 200mg

Abortion Clinic In Stanger ](+27832195400*)[ 🏥 Safe Abortion Pills In Stanger...

Abortion Clinic In Stanger ](+27832195400*)[ 🏥 Safe Abortion Pills In Stanger...

Abortion Clinic In Stanger ](+27832195400*)[ 🏥 Safe Abortion Pills In Stanger...

Medical / Health Care (+971588192166) Mifepristone and Misoprostol tablets 200mg

Alluxio Monthly Webinar May. 14, 2024 For more Alluxio Events: https://www.alluxio.io/events/ Speaker: - ChanChan Mao (Developer Advocate, Alluxio) - Bin Fan (VP of Technology, Alluxio) Running AI/ML workloads in different clouds present unique challenges. The key to a manageable multi-cloud architecture is the ability to seamlessly access data across environments with high performance and low cost. This webinar is designed for data platform engineers, data infra engineers, data engineers, and ML engineers who work with multiple data sources in hybrid or multi-cloud environments. Chanchan and Bin will guide the audience through using Alluxio to greatly simplify data access and make model training and serving more efficient in these environments. You will learn: - How to access data in multi-region, hybrid, and multi-cloud like accessing a local file system - How to run PyTorch to read datasets and write checkpoints to remote storage with Alluxio as the distributed data access layer - Real-world examples and insights from tech giants like Uber, AliPay and more

Alluxio Monthly Webinar | Simplify Data Access for AI in Multi-Cloud

Alluxio Monthly Webinar | Simplify Data Access for AI in Multi-Cloud

Alluxio Monthly Webinar | Simplify Data Access for AI in Multi-Cloud

These are the slides of my OpenSIPS Summit 2024 presentation about automating your test calls. It dives into why automated call testing is crucial, how to integrate it into your CI/CD pipeline and how to extend testing of single calls into load testing and testing of other protocols. The presentation also provides details how open source components such as sipp, asterisk and rtpengine are used to implement agents to generate test calls for SIP, Fax and WebRTC at scale.

Automate your OpenSIPS config tests - OpenSIPS Summit 2024

Automate your OpenSIPS config tests - OpenSIPS Summit 2024

Automate your OpenSIPS config tests - OpenSIPS Summit 2024

Weeding your micro service landscape.pdf

Weeding your micro service landscape.pdf

Weeding your micro service landscape.pdf

Abortion Pill Prices Turfloop ](+27832195400*)[ 🏥 Women's Abortion Clinic in ...

Abortion Pill Prices Turfloop ](+27832195400*)[ 🏥 Women's Abortion Clinic in ...

Abortion Pill Prices Turfloop ](+27832195400*)[ 🏥 Women's Abortion Clinic in ...

Medical / Health Care (+971588192166) Mifepristone and Misoprostol tablets 200mg

Wired_2.0_CREATE YOUR ULTIMATE LEARNING ENVIRONMENT_JCON_16052024

Wired_2.0_CREATE YOUR ULTIMATE LEARNING ENVIRONMENT_JCON_16052024

Wired_2.0_CREATE YOUR ULTIMATE LEARNING ENVIRONMENT_JCON_16052024

Navigation in flutter – how to add stack, tab, and drawer navigators to your ...

Navigation in flutter – how to add stack, tab, and drawer navigators to your ...

Navigation in flutter – how to add stack, tab, and drawer navigators to your ...

Encryption Recap: A Refresher on Key Concepts

Encryption Recap: A Refresher on Key Concepts

Encryption Recap: A Refresher on Key Concepts

Software Engineering - Part 1 which describes the following topics: Introduction: The evolving role of software, The changing nature of software, Software engineering, A Process Framework, Process Patterns, Process Assessment, Personal and Team Process Models, Process Technology, Product and Process. Process Models: Prescriptive models, Waterfall model, Incremental process models, Evolutionary process models, Specialized process models. Requirements Engineering: Requirements Engineering Task, Initiating the Requirement Engineering process, Eliciting Requirements, developing use cases, Building the analysis model, Negotiating Requirements, Validating Requirements, Software Requirement Document.

Software Engineering - Introduction + Process Models + Requirements Engineering

Software Engineering - Introduction + Process Models + Requirements Engineering

Software Engineering - Introduction + Process Models + Requirements Engineering

QUICCK 100% in Gauteng/Western Cape *[☎️+27737758557]]@ @# SOWETO@+27737758557 ABORTION PILLS FOR SALE IN STELLENBOSCH CLINIC/WESTERN CAPE cape … CAPE TOWN CLINIC/WESTERN CAPE BELLVILLE LANGA PAROW cape town bellville cape town bellville langa parow khayelitsha gugulethu elies river kraaifontein good wood capetown DR.SINDY +27737758557// abortion clinics// pills WOMEN ABORTION CLINICS PAINFREE,SAFE,QUICCK 100% in western cape +27737758557 Sindy @ABORTION PILLS FOR SALE IN STELLENBOSCH CLINIC/WESTERN CAPE cape … CAPE TOWN CLINIC/WESTERN CAPE BELLVILLE LANGA PAROW cape town bellville cape town bellville langa parow khayelitsha gugulethu elies river kraaifontein good wood capetown bellville.maitland.belhar. aarl, Wellington, Betty’s Bay, Atlantis, Blue Down, Brackenfell, Crossroads, Eerste Rivier, Elsie’s River, Fish Hoek, Goodwood, Gordon’s Bay, Gugulethu,DR EVE +27737758557 // abortion clinics// pills WOMEN ABORTION CLINICS PAINFREE,SAFE,QUICCK 100% in western cape +27737758557 JAULANI@ABORTION PILLS FOR SALE IN STELLENBOSCH CLINIC/WESTERN CAPE cape … CAPE TOWN CLINIC/WESTERN CAPE BELLVILLE LANGA PAROW cape town bellville cape town bellville langa parow khayelitsha gugulethu elies river kraaifontein good wood capetown bellville.maitland.belhar. aarl, Wellington, Betty’s Bay, Atlantis, Blue Down, Brackenfell, Crossroads, Eerste Rivier, Elsie’s River, Fish Hoek, Goodwood, Gordon’s Bay, Gugulethu, Hout Bay, Khayelitsha, Kraainfontein, Kuils River, Langa, Melkbosstrand, Mfuleni, Milnerton, Mitchell’s Plain, Noordhoek, Nyanga, Parow, Kleinmond, Simon’s town, Somerset west, stellenbosch, Ashton Bonnievale Ceres De Doorns Denneburg Franschhoek Gouda Kayamandi Klapmuts Kylemore Languedo cMcGregor Montagu Op-die-Berg Paarl Pniel Prince Alfred Hamlet Rawsonville Robertson Robertsvlei Rozendal Saron Stellenbosch Touws River Tulbagh Wellington Wemmershoek Wolseley Worcester langa parow khayelitsha gugulethu elies river kraaifontein good wood capetown bellville. Ashton Bonnievale Ceres De Doorns Denneburg Franschhoek Gouda Kayamandi Klapmuts Kylemore Languedoc McGregor Montagu Op-die-Berg Paarl Pniel Prince Alfred Hamlet Rawsonville Robertson Robertsvlei Rozendal Saron Stellenbosch Touws River Tulbagh Wellington Alexandra · Diepsloot · Ennerdale ; Alberton · Bedfordview · Benoni ; Atteridgeville · Bronberg · Mogapeng · Mogodumo · Mogwadi · Mohlake · Mohlarekoma · Mohlatsengwane · Mohlopi · Mohlopi · Mohlakeng · Mohwelere · Mokgoatsane · Mokopane. Makhado · Bela-Bela · Hoedspruit · Haenertsburg · Lephalale (Ellisras) · Phalaborwa · Polokwane (Pietersburg) · Tzaneen.Hoedspruit · Hoedspruit. #1 – Hoedspruit ; Haenertsburg · Haenertsburg. #2 – Haenertsburg ; Tzaneen · Tzaneen. Polokwane, Louis Trichardt, Mahwelereng, Giyani, Limpopo,

Abortion Pills For Sale WhatsApp[[+27737758557]] In Birch Acres, Abortion Pil...

Abortion Pills For Sale WhatsApp[[+27737758557]] In Birch Acres, Abortion Pil...

Abortion Pills For Sale WhatsApp[[+27737758557]] In Birch Acres, Abortion Pil...

Microsoft365_Dev_Security_2024_05_16.pdf

Microsoft365_Dev_Security_2024_05_16.pdf

Microsoft365_Dev_Security_2024_05_16.pdf

Recently uploaded (20)

Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...

Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...

Team Transformation Tactics for Holistic Testing and Quality (NewCrafts Paris...

Community is Just as Important as Code by Andrea Goulet

Community is Just as Important as Code by Andrea Goulet

Community is Just as Important as Code by Andrea Goulet

Anypoint Code Builder - Munich MuleSoft Meetup - 16th May 2024

Anypoint Code Builder - Munich MuleSoft Meetup - 16th May 2024

Anypoint Code Builder - Munich MuleSoft Meetup - 16th May 2024

Workshop: Enabling GenAI Breakthroughs with Knowledge Graphs - GraphSummit Milan

Workshop: Enabling GenAI Breakthroughs with Knowledge Graphs - GraphSummit Milan

Workshop: Enabling GenAI Breakthroughs with Knowledge Graphs - GraphSummit Milan

A Deep Dive into Secure Product Development Frameworks.pdf

A Deep Dive into Secure Product Development Frameworks.pdf

A Deep Dive into Secure Product Development Frameworks.pdf

Rapidoform for Modern Form Building and Insights

Rapidoform for Modern Form Building and Insights

Rapidoform for Modern Form Building and Insights

AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdf

AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdf

AzureNativeQumulo_HPC_Cloud_Native_Benchmarks.pdf

Workshop - Architecting Innovative Graph Applications- GraphSummit Milan

Workshop - Architecting Innovative Graph Applications- GraphSummit Milan

Workshop - Architecting Innovative Graph Applications- GraphSummit Milan

Abortion Pill Prices Jozini ](+27832195400*)[ 🏥 Women's Abortion Clinic in Jo...

Abortion Pill Prices Jozini ](+27832195400*)[ 🏥 Women's Abortion Clinic in Jo...

Abortion Pill Prices Jozini ](+27832195400*)[ 🏥 Women's Abortion Clinic in Jo...

Abortion Clinic In Stanger ](+27832195400*)[ 🏥 Safe Abortion Pills In Stanger...

Abortion Clinic In Stanger ](+27832195400*)[ 🏥 Safe Abortion Pills In Stanger...

Abortion Clinic In Stanger ](+27832195400*)[ 🏥 Safe Abortion Pills In Stanger...

Alluxio Monthly Webinar | Simplify Data Access for AI in Multi-Cloud

Alluxio Monthly Webinar | Simplify Data Access for AI in Multi-Cloud

Alluxio Monthly Webinar | Simplify Data Access for AI in Multi-Cloud

Automate your OpenSIPS config tests - OpenSIPS Summit 2024

Automate your OpenSIPS config tests - OpenSIPS Summit 2024

Automate your OpenSIPS config tests - OpenSIPS Summit 2024

Weeding your micro service landscape.pdf

Weeding your micro service landscape.pdf

Weeding your micro service landscape.pdf

Abortion Pill Prices Turfloop ](+27832195400*)[ 🏥 Women's Abortion Clinic in ...

Abortion Pill Prices Turfloop ](+27832195400*)[ 🏥 Women's Abortion Clinic in ...

Abortion Pill Prices Turfloop ](+27832195400*)[ 🏥 Women's Abortion Clinic in ...

Wired_2.0_CREATE YOUR ULTIMATE LEARNING ENVIRONMENT_JCON_16052024

Wired_2.0_CREATE YOUR ULTIMATE LEARNING ENVIRONMENT_JCON_16052024

Wired_2.0_CREATE YOUR ULTIMATE LEARNING ENVIRONMENT_JCON_16052024

Navigation in flutter – how to add stack, tab, and drawer navigators to your ...

Navigation in flutter – how to add stack, tab, and drawer navigators to your ...

Navigation in flutter – how to add stack, tab, and drawer navigators to your ...

Encryption Recap: A Refresher on Key Concepts

Encryption Recap: A Refresher on Key Concepts

Encryption Recap: A Refresher on Key Concepts

Software Engineering - Introduction + Process Models + Requirements Engineering

Software Engineering - Introduction + Process Models + Requirements Engineering

Software Engineering - Introduction + Process Models + Requirements Engineering

Abortion Pills For Sale WhatsApp[[+27737758557]] In Birch Acres, Abortion Pil...

Abortion Pills For Sale WhatsApp[[+27737758557]] In Birch Acres, Abortion Pil...

Abortion Pills For Sale WhatsApp[[+27737758557]] In Birch Acres, Abortion Pil...

Microsoft365_Dev_Security_2024_05_16.pdf

Microsoft365_Dev_Security_2024_05_16.pdf

Microsoft365_Dev_Security_2024_05_16.pdf

Kruize-HPO-meetup-Aug2022.pdf

1. Choosing performance tunables is not that simple Chandrakala Subramanyam @chandra25ms Kusuma Chalasani @kusumach

2. 2 Agenda ➢ Challenges / Motivation ➢ Introduction to Bayesian Optimization / HPO ➢ HPO Use cases ➢ HPO Demo ➢ Experiment results

3. SRE Optimize App Stack - Challenges 3

4. 4 Choices Are Never This simple!

5. 5 Building an ML Model? Fix the Hyperparameters first! ⚫ Model architecture ⚫ Learning rate ⚫ Number of epochs ⚫ Number of branches in a decision tree ⚫ Number of clusters in a clustering algorithm

6. 6 Fixing Hyperparameters - Options ⚫ Manual Search ⚫ Random Search ⚫ Grid Search ⚫ Automated Hyperparameter Tuning - Bayesian Optimization ⚫ ...

7. Ref: https://distill.pub/2020/bayesian-optimization/ Bayesian Optimization 7

8. HPO Architecture 8

9. 9 HPO - Usecases Native Containers Kubernetes Kruize HPO OS settings tuning Resource Usage Microservices App tuning

10. HPO Demo 10

11. Be careful what you wish for ! 11

12. Kruize HPO vs Default Config [Obj Fn = maximize(1.25*Throughput / 1.5*Response_time / 0.25*Max_response_time), Fixed Resources] Default 14.21 ms 2.39 ms 12

13. Kruize HPO vs Default Config [Obj Fn = maximize(1.25*Throughput / 1.5*Response_time / 0.25*Max_Response_Time), Fixed Resources] 83% better response time 1.3% better throughput 13

14. Interested in HPO? ⚫ HPO - https://github.com/kruize/hpo ⚫ Kruize Demos - https://github.com/kruize/kruize-demos Call for collaboration ! Kruize Slack kruize@redhat.com 14