Google Dev Summit Extended Seoul - TensorFlow: Tensorboard & Keras

•

34 likes•6,972 views

Taegyun Jeon

Engineering

Taegyun Jeon
TensorFlow Dev Summit Extended Seoul / 2017.02.22
R&D Center, Satrec Initiative
TensorFlow:
TensorBoard & Keras
GDG Meetup in February

Contents
 Integrating Keras & TensorFlow: The Keras Workflow,
Expanded
▫ Speaker: Francois Chollet
 Hands-on TensorBoard
▫ Speaker: Dandelion Mané
▫ Code & Slide: https://goo.gl/San2uR
 All contents are provided from TensorFlow Dev Summit 2017
(https://events.withgoogle.com/tensorflow-dev-summit/)
Page 2[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras

Keras
 An API spec for building deep learning models across many
platforms
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 3

Keras의 인기 비결?
 간편함
 간결함
 TF의 잦은 API 변경
 연구자와 개발자의
의사소통 도구
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 4

tf.keras
 Keras는 결국 TensorFlow의 품으로!
 TensorFlow layers = Keras layers
 Keras Model
▫ Sequence model and functional model API
 TensorFlow 기능과 통합 가능
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 5
https://www.tensorflow.org/api_docs/python/tf/contrib/learn/Experiment

Keras 사용자들에게 어떤 일들이?
 TensorFlow와 Keras사이에서 고민하지 않으셔도 됩니다.
 TF와 Keras의 장단점을 파악해서 섞어서 쓰세요.
 Distributed Training, Cloud ML, Hyperparameter setting,
TF-Serving
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 6
https://blog.keras.io/keras-as-a-simplified-interface-to-tensorflow-tutorial.html

Keras Example: Video QA
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 7

Keras Example: Video QA
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 8

Keras Example: Video QA
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 9

Keras Example: Video QA
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 10

Keras Example: Video QA
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 11

Keras Example: Video QA
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 12

[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 13

Keras Example: Video QA
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 14

Keras Example: Visual VQA
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 15

More Examples
 Keras Blog
▫ https://blog.keras.io/
 Keras API (keras.applications)
▫ https://keras.io/applications/
 Keras Example Directories
▫ https://github.com/fchollet/keras/tree/master/examples
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 16

Keras: Summary
 TF 사용자: 손쉬운 high-level api 사용 가능
 Keras 사용자: 모델만 keras로 만들고 이후 운영은 TF로 가능
 버전 변경 계획
▫ tf.contrib.keras (TF 1.1 / 올해 3월 중)
▫ tf.keras (TF 1.2)
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 17

TensorBoard
 TensorBoard를 이용해서 얻는 장점
▫ Debug: 내부를 알기 힘든 딥러닝 모델을 눈으로 확인
▫ Hyperparameter Tuning
▫ Visualize inference results
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 18

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 19

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 20

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 21

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 22

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 23

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 24

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 25

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 26
엄청 복잡해보이네요

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 27

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 28

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 29

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 30

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 31
이제 레이어별로
깔끔해졌습니다.

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 32
이전과 비교해봐도
눈에 잘 들어옵니다

TensorBoard: MNIST Example
 tf.summary.scalar
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 33
 tf.summary.image
 tf.summary.audio  tf.summary.histogram

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 34

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 35

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 36

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 37
정확도가 엉망이군요!

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 38
Cross-entropy도 뭔가
이상하죠?

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 39
Convolution layer가
변하지 않았음을
알 수 있습니다.

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 40
Histogram도 역시 마
찬가지군요

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 41

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 42

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 43
이제 제대로
학습이 되는군요

TensorBoard: MNIST Example
 Hyperparameter search
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 44

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 45
다양한 run을 기록

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 46
Hyperparameter를 찾을땐
epoch을 조금만 설정해서
초기 성능만 보고
ckpt 관리 및 추가학습

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 47

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 48

TensorBoard: MNIST Example
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 49

TensorBoard
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 50

TensorBoard: Summary
 Debug
▫ name_scope 정의와 tensor의 name 붙이기를 생활화
▫ Summary 기능을 활용
• Scalar, Image, Audio, Histogram
 Hyperparameter search
▫ 다양한 run을 짧게 돌려보는 것을 권장
▫ 파라미터 뿐만 아니라, 모델 형태에도 적용 가능
 Embedding visualization
▫ 표현하기 힘든 내용은 가시화!
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 51

Q & A
Any Questions?
[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 52

This slide was presented at the SK Telecom T Developer Forum. It contains the brief evaluation results of the query execution performance of Tajo on Swift. I conducted two kinds of experiments; The first experiment was to compare the performance of Tajo with on another distributed storage, i.e., HDFS. And the second experiment was the scalability test of Swift. Interestingly, the scan performance on Swift is slower more than two times than that on HDFS. In addition, the task scheduling time on Swift is much greater than that on HDFS, which means the query initialization cost is very high.

Lightning Talk: MongoDB Migration Strategies

MongoDB

SchemEX - Creating the Yellow Pages for the Linked Open Data Cloud

Ansgar Scherp

HBaseCon 2015: Solving HBase Performance Problems with Apache HTrace

HBaseCon

HTrace is a new Apache incubator project which makes it much easier to diagnose and detect performance problems in HBase. It provides a unified view of the performance of requests, following them from their origin in the HBase client, through the HBase region servers, and finally into HDFS. System administrators can use a central web interface to query and view aggregate performance information for the whole cluster. This talk will cover the motivations for creating HTrace, its design, and some examples of how HTrace can help diagnose real-world HBase problems.

End-to-end Big Data Projects with Python - StampedeCon Big Data Conference 2017

StampedeCon

This talk will go over how to build an end-to-end data processing system in Python, from data ingest, to data analytics, to machine learning, to user presentation. Developments in old and new tools have made this particularly possible today. The talk in particular will talk about Airflow for process workflows, PySpark for data processing, Python data science libraries for machine learning and advanced analytics, and building agile microservices in Python. System architects, software engineers, data scientists, and business leaders can all benefit from attending the talk. They should learn how to build more agile data processing systems and take away some ideas on how their data systems could be simpler and more powerful.

MongoDB-Migration-Strategies

andyjwoodard

RubiX

Shubham Tagra

Big data in the cloud

Ben Sullins

Query optimization in Apache Tajo

Jihoon Son

Deep dive into deeplearn.js

Kai Sasaki

Mathias Brandewinder, Software Engineer & Data Scientist, Clear Lines Consult...

MLconf

Scripts that Scale with F# and mbrace.io: Nothing beats interactive scripting for productive data exploration and rapid prototyping: grab data, run code, and iterate based on feedback. However, that story starts to break down once you need to process large datasets or expensive computations. Your local machine becomes the bottleneck, and your are left with a slow and unresponsive environment. In this talk, we will demonstrate on live examples how you can have your cake and eat it, too, using mbrace.io, a free, open-source engine for scalable cloud programming. Using a simple programming model, you can keep working from your favorite scripting environment, and execute code interactively against a cluster on the Azure cloud. We will discuss the relevance of F# and mbrace in a data science and machine learning context, from parallelizing code and data processing in a functional style, to leveraging F# type providers to consume data or even run R packages.

Xephon K A Time series database with multiple backends

University of California, Santa Cruz

As simple as Apache Spark

Data Science Warsaw

Lecture5 Expert Systems And Artificial IntelligenceKodok Ngorex

Introduction to CLIPS Expert System

Motaz Saad

What's hot

Introduction to Apache Tajo: Future of Data Warehouse

Jihoon Son

Google's Dremel

Maria Stylianou

Hadoop Architecture in Depth

Syed Hadoop

Improve Presto Architectural Decisions with Shadow Cache

Alluxio, Inc.

Toronto meetup 20190917

Bill Liu

Strata NYC 2015: Sketching Big Data with Spark: randomized algorithms for lar...

Databricks

Data Science Stack with MongoDB and RStudio

Winston Chen

Apache Tajo on Swift: Bringing SQL to the OpenStack World

Jihoon Son

Lightning Talk: MongoDB Migration Strategies

MongoDB

SchemEX - Creating the Yellow Pages for the Linked Open Data Cloud

Ansgar Scherp

HBaseCon 2015: Solving HBase Performance Problems with Apache HTrace

HBaseCon

End-to-end Big Data Projects with Python - StampedeCon Big Data Conference 2017

StampedeCon

MongoDB-Migration-Strategies

andyjwoodard

RubiX

Shubham Tagra

Big data in the cloud

Ben Sullins

Query optimization in Apache Tajo

Jihoon Son

Deep dive into deeplearn.js

Kai Sasaki

Mathias Brandewinder, Software Engineer & Data Scientist, Clear Lines Consult...

MLconf

Xephon K A Time series database with multiple backends

University of California, Santa Cruz

As simple as Apache Spark

Data Science Warsaw

What's hot (20)

Introduction to Apache Tajo: Future of Data Warehouse

Google's Dremel

Hadoop Architecture in Depth

Improve Presto Architectural Decisions with Shadow Cache

Toronto meetup 20190917

Strata NYC 2015: Sketching Big Data with Spark: randomized algorithms for lar...

Data Science Stack with MongoDB and RStudio

Apache Tajo on Swift: Bringing SQL to the OpenStack World

Lightning Talk: MongoDB Migration Strategies

SchemEX - Creating the Yellow Pages for the Linked Open Data Cloud

HBaseCon 2015: Solving HBase Performance Problems with Apache HTrace

End-to-end Big Data Projects with Python - StampedeCon Big Data Conference 2017

MongoDB-Migration-Strategies

RubiX

Big data in the cloud

Query optimization in Apache Tajo

Deep dive into deeplearn.js

Mathias Brandewinder, Software Engineer & Data Scientist, Clear Lines Consult...

Xephon K A Time series database with multiple backends

As simple as Apache Spark

Viewers also liked

Lecture5 Expert Systems And Artificial IntelligenceKodok Ngorex

Introduction to CLIPS Expert System

Motaz Saad

Expert Systemsosmancikk

Expert Systems

sadeenedian08

Electricity price forecasting with Recurrent Neural Networks

Taegyun Jeon

텐서플로우 설치도 했고 튜토리얼도 봤고 기초 예제도 짜봤다면 TensorFlow KR Meetup 2016

Taehoon Kim

Viewers also liked (6)

Lecture5 Expert Systems And Artificial Intelligence

Introduction to CLIPS Expert System

Expert Systems

Electricity price forecasting with Recurrent Neural Networks

텐서플로우 설치도 했고 튜토리얼도 봤고 기초 예제도 짜봤다면 TensorFlow KR Meetup 2016

Similar to Google Dev Summit Extended Seoul - TensorFlow: Tensorboard & Keras

TensorFlow Tutorial | Deep Learning Using TensorFlow | TensorFlow Tutorial Py...

Edureka!

This Edureka TensorFlow Tutorial (Blog: https://goo.gl/HTE7uB) will help you in understanding various important basics of TensorFlow. It also includes a use-case in which we will create a model that will differentiate between a rock and a mine using TensorFlow. Below are the topics covered in this tutorial: 1. What are Tensors? 2. What is TensorFlow? 3. TensorFlow Code-basics 4. Graph Visualization 5. TensorFlow Data structures 6. Use-Case Naval Mine Identifier (NMI)

TonY: Native support of TensorFlow on Hadoop

Anthony Hsu

Anthony Hsu, Jonathan Hung, and Keqiu Hu offer an overview of TensorFlow on YARN (TonY), a framework to natively run TensorFlow on Hadoop. TonY enables running TensorFlow distributed training as a new type of Hadoop application. Its native Hadoop connector, together with other features, aims to run TensorFlow jobs as reliably and flexibly as other applications on Hadoop. Video: https://youtu.be/sIfnsU-5jHM

Container orchestration: the cold war - Giulio De Donato - Codemotion Rome 2017

Codemotion

Why LaTeX

Jaspreet Sarao

Callbacks part2

Alireza AkhavanPour

Tensorflow in production with AWS Lambda

Fabian Dubois

A TensorFlow ppt for the gdsc event used in my college.

MUVVALALAKSHMIKOUSIK

High Performance Distributed TensorFlow with GPUs - NYC Workshop - July 9 2017

Chris Fregly

http://pipeline.io Title PipelineAI Distributed Spark ML + Tensorflow AI + GPU Workshop *A GPU-based cloud instance will be provided to each attendee as part of this event Highlights We will each build an end-to-end, continuous Tensorflow AI model training and deployment pipeline on our own GPU-based cloud instance. At the end, we will combine our cloud instances to create the LARGEST Distributed Tensorflow AI Training and Serving Cluster in the WORLD! Pre-requisites Just a modern browser, internet connection, and a good night's sleep! We'll provide the rest. Agenda Spark ML TensorFlow AI Storing and Serving Models with HDFS Trade-offs of CPU vs. *GPU, Scale Up vs. Scale Out CUDA + cuDNN GPU Development Overview TensorFlow Model Checkpointing, Saving, Exporting, and Importing Distributed TensorFlow AI Model Training (Distributed Tensorflow) TensorFlow's Accelerated Linear Algebra Framework (XLA) TensorFlow's Just-in-Time (JIT) Compiler, Ahead of Time (AOT) Compiler Centralized Logging and Visualizing of Distributed TensorFlow Training (Tensorboard) Distributed Tensorflow AI Model Serving/Predicting (TensorFlow Serving) Centralized Logging and Metrics Collection (Prometheus, Grafana) Continuous TensorFlow AI Model Deployment (TensorFlow, Airflow) Hybrid Cross-Cloud and On-Premise Deployments (Kubernetes) High-Performance and Fault-Tolerant Micro-services (NetflixOSS) Bio Chris Fregly is Founder and Research Engineer at PipelineIO, a Streaming Machine Learning and Artificial Intelligence Startup based in San Francisco. He is also an Apache Spark Contributor, a Netflix Open Source Committer, founder of the Global Advanced Spark and TensorFlow Meetup, author of the O’Reilly Training and Video Series titled, "High Performance TensorFlow in Production." Previously, Chris was a Distributed Systems Engineer at Netflix, a Data Solutions Engineer at Databricks, and a Founding Member and Principal Engineer at the IBM Spark Technology Center in San Francisco. Github Repo https://github.com/fluxcapacitor/pipeline Video https://youtu.be/oNf3I1fVmg8

The Flow of TensorFlow

Jeongkyu Shin

이 발표에서는 TensorFlow의 지난 1년을 간단하게 돌아보고, TensorFlow의 차기 로드맵에 따라 개발 및 도입될 예정인 여러 기능들을 소개합니다. 또한 2017년 및 2018년의 머신러닝 프레임워크 개발 트렌드와 방향에 대한 이야기도 함께 합니다. In this talk, I look back the TensorFlow development over the past year. Then discusses the overall development direction of machine learning frameworks, with an introduction to features that will be added to TensorFlow later on.

Terraform AWS modules and some best-practices - May 2019

Anton Babenko

GitHub as a Landing Page

Pronovix

TechSEO Boost 2018: Implementing Hreflang on Legacy Tech Stacks Using Service...

Catalyst

One of the challenges faced at enterprise SEO level is often the legacy platforms and tech stacks that you inherit. Finding a cost-effective way of implementing international SEO best practice is often a barrier to internationalisation. Edge technology is creating new opportunities to optimise websites independently of the inherited technological barriers. In this session, SALT.agency’s Dan Taylor will explore their findings from implementing Hreflang using cutting edge technology to remove these barriers.

DevOps-Redefining your IT Strategy-28thJan15

Edureka!

OpenStack Trove Update - Juno, Kilo and Beyond

OpenStack_Online

Overview of TensorFlow For Natural Language Processing

ananth

Docker + Tenserflow + GOlang - Golang singapore Meetup

sangam biradar

Terraform 0.12 + Terragrunt

Anton Babenko

TensorFlowOnSpark: Scalable TensorFlow Learning on Spark Clusters

DataWorks Summit

In recent releases, TensorFlow has been enhanced for distributed learning and HDFS access. Outside of the Google cloud, however, users still needed a dedicated cluster for TensorFlow applications. There are several community projects wiring TensorFlow onto Apache Spark clusters. Unfortunately, they are limited to support synchronous distributed learning only, and don’t allow TensorFlow servers to communicate with each other directly. In this talk, we will introduce a new framework, TensorFlowOnSpark, for scalable TensorFlow learning, which will be open sourced in Q1 2017. This new framework enables easy experimentation for algorithm designs, and supports scalable training & inferencing on Spark clusters. It supports all TensorFlow functionalities including synchronous & asynchronous learning, model & data parallelism, and TensorBoard. It provides architectural flexibility for data ingestion to TensorFlow and network protocols for server-to-server communication. With a few lines of code changes, an existing TensorFlow algorithm can be transformed into a scalable application.

Basics of LaTeX

Jaspreet Sarao

High Performance TensorFlow in Production -- Sydney ML / AI Train Workshop @ ...

Chris Fregly

http://pipeline.ai Title PipelineAI Distributed Spark ML + Tensorflow AI + GPU Workshop *A GPU-based cloud instance will be provided to each attendee as part of this event Highlights We will each build an end-to-end, continuous Tensorflow AI model training and deployment pipeline on our own GPU-based cloud instance. At the end, we will combine our cloud instances to create the LARGEST Distributed Tensorflow AI Training and Serving Cluster in the WORLD! Agenda Spark ML Tensorflow AI Storing and Serving Models with HDFS Trade-offs of CPU vs. *GPU, Scale Up vs. Scale Out CUDA + cuDNN GPU Development Overview Tensorflow Model Checkpointing, Saving, Exporting, and Importing Distributed Tensorflow AI Model Training (Distributed Tensorflow) Centralized Logging and Visualizing of Distributed Tensorflow Training (Tensorboard) Distributed Tensorflow AI Model Serving/Predicting (Tensorflow Serving) Centralized Logging and Metrics Collection (Prometheus, Grafana) Continuous Tensorflow AI Model Deployment (Tensorflow, Airflow) Hybrid Cross-Cloud and On-Premise Deployments (Kubernetes) High-Performance and Fault-Tolerant Microsservices using Request Batching and Circuit Breakers (NetflixOSS) Github Repo https://github.com/fluxcapacitor/pipeline

Similar to Google Dev Summit Extended Seoul - TensorFlow: Tensorboard & Keras (20)

TensorFlow Tutorial | Deep Learning Using TensorFlow | TensorFlow Tutorial Py...

TonY: Native support of TensorFlow on Hadoop

Container orchestration: the cold war - Giulio De Donato - Codemotion Rome 2017

Why LaTeX

Callbacks part2

Tensorflow in production with AWS Lambda

A TensorFlow ppt for the gdsc event used in my college.

High Performance Distributed TensorFlow with GPUs - NYC Workshop - July 9 2017

The Flow of TensorFlow

Terraform AWS modules and some best-practices - May 2019

GitHub as a Landing Page

TechSEO Boost 2018: Implementing Hreflang on Legacy Tech Stacks Using Service...

DevOps-Redefining your IT Strategy-28thJan15

OpenStack Trove Update - Juno, Kilo and Beyond

Overview of TensorFlow For Natural Language Processing

Docker + Tenserflow + GOlang - Golang singapore Meetup

Terraform 0.12 + Terragrunt

TensorFlowOnSpark: Scalable TensorFlow Learning on Spark Clusters

Basics of LaTeX

High Performance TensorFlow in Production -- Sydney ML / AI Train Workshop @ ...

More from Taegyun Jeon

TensorFlow-KR 3rd meetup - Lightning Talk for SI Analytics

Taegyun Jeon

TensorFlow Dev Summit 2018 Extended: TensorFlow Eager Execution

Taegyun Jeon

[OSGeo-KR Tech Workshop] Deep Learning for Single Image Super-Resolution

Taegyun Jeon

[PR12] PR-063: Peephole predicting network performance before training

Taegyun Jeon

GDG DevFest Xiamen 2017

Taegyun Jeon

[PR12] PR-050: Convolutional LSTM Network: A Machine Learning Approach for Pr...

Taegyun Jeon

GDG DevFest Seoul 2017: Codelab - Time Series Analysis for Kaggle using Tenso...

Taegyun Jeon

[PR12] PR-036 Learning to Remember Rare Events

Taegyun Jeon

[대전AI포럼] 위성영상 분석 기술 개발 현황 소개

Taegyun Jeon

[PR12] PR-026: Notes for CVPR Machine Learning Sessions

Taegyun Jeon

PR-026: Notes for CVPR Machine Learning Session Paper 1: Borrowing Treasures From the Wealthy: Deep Transfer Learning Through Selective Joint Fine-Tuning, https://arxiv.org/abs/1702.08690 Paper 2: The More You Know: Using Knowledge Graphs for Image Classification, https://arxiv.org/abs/1612.04844 Paper 3: On Compressing Deep Models by Low Rank and Sparse Decomposition, http://openaccess.thecvf.com/content_cvpr_2017/papers/Yu_On_Compressing_Deep_CVPR_2017_paper.pdf

[PR12] You Only Look Once (YOLO): Unified Real-Time Object Detection

Taegyun Jeon

[PR12] image super resolution using deep convolutional networks

Taegyun Jeon

TensorFlow KR 2nd Meetup - Lightening talk (Satrec Initiative)

Taegyun Jeon

인공지능: 변화와 능력개발

Taegyun Jeon

More from Taegyun Jeon (14)

TensorFlow-KR 3rd meetup - Lightning Talk for SI Analytics

TensorFlow Dev Summit 2018 Extended: TensorFlow Eager Execution

[OSGeo-KR Tech Workshop] Deep Learning for Single Image Super-Resolution

[PR12] PR-063: Peephole predicting network performance before training

GDG DevFest Xiamen 2017

[PR12] PR-050: Convolutional LSTM Network: A Machine Learning Approach for Pr...

GDG DevFest Seoul 2017: Codelab - Time Series Analysis for Kaggle using Tenso...

[PR12] PR-036 Learning to Remember Rare Events

[대전AI포럼] 위성영상 분석 기술 개발 현황 소개

[PR12] PR-026: Notes for CVPR Machine Learning Sessions

[PR12] You Only Look Once (YOLO): Unified Real-Time Object Detection

[PR12] image super resolution using deep convolutional networks

TensorFlow KR 2nd Meetup - Lightening talk (Satrec Initiative)

인공지능: 변화와 능력개발

Recently uploaded

English lab ppt no titlespecENG PPTt.pdf

BrazilAccount1

Final project report on grocery store management system..pdf

Kamal Acharya

In today’s fast-changing business environment, it’s extremely important to be able to respond to client needs in the most effective and timely manner. If your customers wish to see your business online and have instant access to your products or services. Online Grocery Store is an e-commerce website, which retails various grocery products. This project allows viewing various products available enables registered users to purchase desired products instantly using Paytm, UPI payment processor (Instant Pay) and also can place order by using Cash on Delivery (Pay Later) option. This project provides an easy access to Administrators and Managers to view orders placed using Pay Later and Instant Pay options. In order to develop an e-commerce website, a number of Technologies must be studied and understood. These include multi-tiered architecture, server and client-side scripting techniques, implementation technologies, programming language (such as PHP, HTML, CSS, JavaScript) and MySQL relational databases. This is a project with the objective to develop a basic website where a consumer is provided with a shopping cart website and also to know about the technologies used to develop such a website. This document will discuss each of the underlying technologies to create and implement an e- commerce website.

Student information management system project report ii.pdf

Kamal Acharya

power quality voltage fluctuation UNIT - I.pptx

ViniHema

Investor-Presentation-Q1FY2024 investor presentation document.pptx

AmarGB2

CME397 Surface Engineering- Professional Elective

karthi keyan

H.Seo, ICLR 2024, MLILAB, KAIST AI.pdf

MLILAB

Hybrid optimization of pumped hydro system and solar- Engr. Abdul-Azeez.pdf

fxintegritypublishin

Advancements in technology unveil a myriad of electrical and electronic breakthroughs geared towards efficiently harnessing limited resources to meet human energy demands. The optimization of hybrid solar PV panels and pumped hydro energy supply systems plays a pivotal role in utilizing natural resources effectively. This initiative not only benefits humanity but also fosters environmental sustainability. The study investigated the design optimization of these hybrid systems, focusing on understanding solar radiation patterns, identifying geographical influences on solar radiation, formulating a mathematical model for system optimization, and determining the optimal configuration of PV panels and pumped hydro storage. Through a comparative analysis approach and eight weeks of data collection, the study addressed key research questions related to solar radiation patterns and optimal system design. The findings highlighted regions with heightened solar radiation levels, showcasing substantial potential for power generation and emphasizing the system's efficiency. Optimizing system design significantly boosted power generation, promoted renewable energy utilization, and enhanced energy storage capacity. The study underscored the benefits of optimizing hybrid solar PV panels and pumped hydro energy supply systems for sustainable energy usage. Optimizing the design of solar PV panels and pumped hydro energy supply systems as examined across diverse climatic conditions in a developing country, not only enhances power generation but also improves the integration of renewable energy sources and boosts energy storage capacities, particularly beneficial for less economically prosperous regions. Additionally, the study provides valuable insights for advancing energy research in economically viable areas. Recommendations included conducting site-specific assessments, utilizing advanced modeling tools, implementing regular maintenance protocols, and enhancing communication among system components.

Hierarchical Digital Twin of a Naval Power System

Kerry Sado

A hierarchical digital twin of a Naval DC power system has been developed and experimentally verified. Similar to other state-of-the-art digital twins, this technology creates a digital replica of the physical system executed in real-time or faster, which can modify hardware controls. However, its advantage stems from distributing computational efforts by utilizing a hierarchical structure composed of lower-level digital twin blocks and a higher-level system digital twin. Each digital twin block is associated with a physical subsystem of the hardware and communicates with a singular system digital twin, which creates a system-level response. By extracting information from each level of the hierarchy, power system controls of the hardware were reconfigured autonomously. This hierarchical digital twin development offers several advantages over other digital twins, particularly in the field of naval power systems. The hierarchical structure allows for greater computational efficiency and scalability while the ability to autonomously reconfigure hardware controls offers increased flexibility and responsiveness. The hierarchical decomposition and models utilized were well aligned with the physical twin, as indicated by the maximum deviations between the developed digital twin hierarchy and the hardware.

在线办理(ANU毕业证书)澳洲国立大学毕业证录取通知书一模一样

obonagu

学校原件一模一样【微信：741003700 】《(ANU毕业证书)澳洲国立大学毕业证》【微信：741003700 】学位证，留信认证（真实可查，永久存档）原件一模一样纸张工艺/offer、雅思、外壳等材料/诚信可靠,可直接看成品样本，帮您解决无法毕业带来的各种难题！外壳，原版制作，诚信可靠，可直接看成品样本。行业标杆！精益求精，诚心合作，真诚制作！多年品质 ,按需精细制作，24小时接单,全套进口原装设备。十五年致力于帮助留学生解决难题，包您满意。本公司拥有海外各大学样板无数，能完美还原。 1:1完美还原海外各大学毕业材料上的工艺：水印，阴影底纹，钢印LOGO烫金烫银，LOGO烫金烫银复合重叠。文字图案浮雕、激光镭射、紫外荧光、温感、复印防伪等防伪工艺。材料咨询办理、认证咨询办理请加学历顾问Q/微741003700 【主营项目】一.毕业证【q微741003700】成绩单、使馆认证、教育部认证、雅思托福成绩单、学生卡等！二.真实使馆公证(即留学回国人员证明,不成功不收费) 三.真实教育部学历学位认证（教育部存档！教育部留服网站永久可查）四.办理各国各大学文凭(一对一专业服务,可全程监控跟踪进度) 如果您处于以下几种情况： ◇在校期间，因各种原因未能顺利毕业……拿不到官方毕业证【q/微741003700】 ◇面对父母的压力，希望尽快拿到； ◇不清楚认证流程以及材料该如何准备； ◇回国时间很长，忘记办理； ◇回国马上就要找工作，办给用人单位看； ◇企事业单位必须要求办理的 ◇需要报考公务员、购买免税车、落转户口 ◇申请留学生创业基金留信网认证的作用: 1:该专业认证可证明留学生真实身份 2:同时对留学生所学专业登记给予评定 3:国家专业人才认证中心颁发入库证书 4:这个认证书并且可以归档倒地方 5:凡事获得留信网入网的信息将会逐步更新到个人身份内，将在公安局网内查询个人身份证信息后，同步读取人才网入库信息 6:个人职称评审加20分 7:个人信誉贷款加10分 8:在国家人才网主办的国家网络招聘大会中纳入资料，供国家高端企业选择人才

ASME IX(9) 2007 Full Version .pdf

AhmedHussein950959

AP LAB PPT.pdf ap lab ppt no title specific

BrazilAccount1

ethical hacking-mobile hacking methods.ppt

Jayaprasanna4

Planning Of Procurement o different goods and services

JoytuBarua2

block diagram and signal flow graph representation

Divya Somashekar

WATER CRISIS and its solutions-pptx 1234

AafreenAbuthahir2

AKS UNIVERSITY Satna Final Year Project By OM Hardaha.pdf

SamSarthak3

Nuclear Power Economics and Structuring 2024

Massimo Talia

一比一原版(UofT毕业证)多伦多大学毕业证成绩单如何办理

ydteq

UofT毕业证原版定制【微信：176555708】【多伦多大学毕业证成绩单-学位证】【微信：176555708】（留信学历认证永久存档查询）采用学校原版纸张、特殊工艺完全按照原版一比一制作（包括：隐形水印，阴影底纹，钢印LOGO烫金烫银，LOGO烫金烫银复合重叠，文字图案浮雕，激光镭射，紫外荧光，温感，复印防伪）行业标杆！精益求精，诚心合作，真诚制作！多年品质 ,按需精细制作，24小时接单,全套进口原装设备，十五年致力于帮助留学生解决难题，业务范围有加拿大、英国、澳洲、韩国、美国、新加坡，新西兰等学历材料，包您满意。 ◆◆◆◆◆ — — — — — — — — 【留学教育】留学归国服务中心 — — — — — -◆◆◆◆◆ 【主营项目】一.毕业证【微信：176555708】成绩单、使馆认证、教育部认证、雅思托福成绩单、学生卡等！二.真实使馆公证(即留学回国人员证明,不成功不收费) 三.真实教育部学历学位认证（教育部存档！教育部留服网站永久可查）四.办理各国各大学文凭(一对一专业服务,可全程监控跟踪进度) 如果您处于以下几种情况： ◇在校期间，因各种原因未能顺利毕业……拿不到官方毕业证【微信：176555708】 ◇面对父母的压力，希望尽快拿到； ◇不清楚认证流程以及材料该如何准备； ◇回国时间很长，忘记办理； ◇回国马上就要找工作，办给用人单位看； ◇企事业单位必须要求办理的 ◇需要报考公务员、购买免税车、落转户口 ◇申请留学生创业基金留信网认证的作用: 1:该专业认证可证明留学生真实身份 2:同时对留学生所学专业登记给予评定 3:国家专业人才认证中心颁发入库证书 4:这个认证书并且可以归档倒地方 5:凡事获得留信网入网的信息将会逐步更新到个人身份内，将在公安局网内查询个人身份证信息后，同步读取人才网入库信息 6:个人职称评审加20分 7:个人信誉贷款加10分→ 【关于价格问题（保证一手价格）我们所定的价格是非常合理的，而且我们现在做得单子大多数都是代理和回头客户介绍的所以一般现在有新的单子我给客户的都是第一手的代理价格，因为我想坦诚对待大家不想跟大家在价格方面浪费时间对于老客户或者被老客户介绍过来的朋友，我们都会适当给一些优惠。 8:在国家人才网主办的国家网络招聘大会中纳入资料，供国家高端企业选择人才选择实体注册公司办理，更放心，更安全！我们的承诺：可来公司面谈，可签订合同，会陪同客户一起到教育部认证窗口递交认证材料，客户在教育部官方认证查询网站查询到认证通过结果后付款，不成功不收费！学历顾问：微信：176555708

Architectural Portfolio Sean Lockwood

seandesed

Recently uploaded (20)

English lab ppt no titlespecENG PPTt.pdf

Final project report on grocery store management system..pdf

Student information management system project report ii.pdf

power quality voltage fluctuation UNIT - I.pptx

Investor-Presentation-Q1FY2024 investor presentation document.pptx

CME397 Surface Engineering- Professional Elective

H.Seo, ICLR 2024, MLILAB, KAIST AI.pdf

Hybrid optimization of pumped hydro system and solar- Engr. Abdul-Azeez.pdf

Hierarchical Digital Twin of a Naval Power System

在线办理(ANU毕业证书)澳洲国立大学毕业证录取通知书一模一样

ASME IX(9) 2007 Full Version .pdf

AP LAB PPT.pdf ap lab ppt no title specific

ethical hacking-mobile hacking methods.ppt

Planning Of Procurement o different goods and services

block diagram and signal flow graph representation

WATER CRISIS and its solutions-pptx 1234

AKS UNIVERSITY Satna Final Year Project By OM Hardaha.pdf

Nuclear Power Economics and Structuring 2024

一比一原版(UofT毕业证)多伦多大学毕业证成绩单如何办理

Architectural Portfolio Sean Lockwood

Google Dev Summit Extended Seoul - TensorFlow: Tensorboard & Keras

1. Taegyun Jeon TensorFlow Dev Summit Extended Seoul / 2017.02.22 R&D Center, Satrec Initiative TensorFlow: TensorBoard & Keras GDG Meetup in February

2. Contents  Integrating Keras & TensorFlow: The Keras Workflow, Expanded ▫ Speaker: Francois Chollet  Hands-on TensorBoard ▫ Speaker: Dandelion Mané ▫ Code & Slide: https://goo.gl/San2uR  All contents are provided from TensorFlow Dev Summit 2017 (https://events.withgoogle.com/tensorflow-dev-summit/) Page 2[TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras

3. Keras  An API spec for building deep learning models across many platforms [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 3

4. Keras의 인기 비결?  간편함  간결함  TF의 잦은 API 변경  연구자와 개발자의 의사소통 도구 [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 4

5. tf.keras  Keras는 결국 TensorFlow의 품으로!  TensorFlow layers = Keras layers  Keras Model ▫ Sequence model and functional model API  TensorFlow 기능과 통합 가능 [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 5 https://www.tensorflow.org/api_docs/python/tf/contrib/learn/Experiment

6. Keras 사용자들에게 어떤 일들이?  TensorFlow와 Keras사이에서 고민하지 않으셔도 됩니다.  TF와 Keras의 장단점을 파악해서 섞어서 쓰세요.  Distributed Training, Cloud ML, Hyperparameter setting, TF-Serving [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 6 https://blog.keras.io/keras-as-a-simplified-interface-to-tensorflow-tutorial.html

7. Keras Example: Video QA [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 7

8. Keras Example: Video QA [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 8

9. Keras Example: Video QA [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 9

10. Keras Example: Video QA [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 10

11. Keras Example: Video QA [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 11

12. Keras Example: Video QA [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 12

13. [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 13

14. Keras Example: Video QA [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 14

15. Keras Example: Visual VQA [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 15

16. More Examples  Keras Blog ▫ https://blog.keras.io/  Keras API (keras.applications) ▫ https://keras.io/applications/  Keras Example Directories ▫ https://github.com/fchollet/keras/tree/master/examples [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 16

17. Keras: Summary  TF 사용자: 손쉬운 high-level api 사용 가능  Keras 사용자: 모델만 keras로 만들고 이후 운영은 TF로 가능  버전 변경 계획 ▫ tf.contrib.keras (TF 1.1 / 올해 3월 중) ▫ tf.keras (TF 1.2) [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 17

18. TensorBoard  TensorBoard를 이용해서 얻는 장점 ▫ Debug: 내부를 알기 힘든 딥러닝 모델을 눈으로 확인 ▫ Hyperparameter Tuning ▫ Visualize inference results [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 18

19. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 19

20. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 20

21. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 21

22. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 22

23. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 23

24. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 24

25. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 25

26. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 26 엄청 복잡해보이네요

27. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 27

28. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 28

29. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 29

30. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 30

31. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 31 이제 레이어별로 깔끔해졌습니다.

32. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 32 이전과 비교해봐도 눈에 잘 들어옵니다

33. TensorBoard: MNIST Example  tf.summary.scalar [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 33  tf.summary.image  tf.summary.audio  tf.summary.histogram

34. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 34

35. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 35

36. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 36

37. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 37 정확도가 엉망이군요!

38. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 38 Cross-entropy도 뭔가 이상하죠?

39. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 39 Convolution layer가 변하지 않았음을 알 수 있습니다.

40. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 40 Histogram도 역시 마 찬가지군요

41. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 41

42. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 42

43. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 43 이제 제대로 학습이 되는군요

44. TensorBoard: MNIST Example  Hyperparameter search [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 44

45. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 45 다양한 run을 기록

46. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 46 Hyperparameter를 찾을땐 epoch을 조금만 설정해서 초기 성능만 보고 ckpt 관리 및 추가학습

47. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 47

48. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 48

49. TensorBoard: MNIST Example [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 49

50. TensorBoard [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 50

51. TensorBoard: Summary  Debug ▫ name_scope 정의와 tensor의 name 붙이기를 생활화 ▫ Summary 기능을 활용 • Scalar, Image, Audio, Histogram  Hyperparameter search ▫ 다양한 run을 짧게 돌려보는 것을 권장 ▫ 파라미터 뿐만 아니라, 모델 형태에도 적용 가능  Embedding visualization ▫ 표현하기 힘든 내용은 가시화! [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 51

52. Q & A Any Questions? [TensorFlow Dev Summit Extended] TensorFlow: TensorBoard & Keras Page 52

Google Dev Summit Extended Seoul - TensorFlow: Tensorboard & Keras

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (6)

Similar to Google Dev Summit Extended Seoul - TensorFlow: Tensorboard & Keras

Similar to Google Dev Summit Extended Seoul - TensorFlow: Tensorboard & Keras (20)

More from Taegyun Jeon

More from Taegyun Jeon (14)

Recently uploaded

Recently uploaded (20)

Google Dev Summit Extended Seoul - TensorFlow: Tensorboard & Keras