README - ruby 설치시 버전 예시 주석 추가

README.md

@@ -29,8 +29,8 @@ workspace$ sudo gem install jekyll-seo-tag
 Note that gem may require ruby version update.
 ```bash
 workspace$ \curl -sSL https://get.rvm.io | bash -s stable
-workspace$ rvm install ruby-2.x.x
-workspace$ rvm use ruby-2.x.x --default
+workspace$ rvm install ruby-2.x.x # rvm install ruby-2.6.3
+workspace$ rvm use ruby-2.x.x --default # rvm use ruby-2.6.3 --default
 ```
 
 #### 4. Clone Repository

_posts/2018-06-28-motnewletter-week18.md

@@ -0,0 +1,96 @@
+---
+layout: post
+title: MoT Weekly News Letters (18th weeks)
+description: "18th Weeks, MoT summary"
+author: jwkang
+date: 2018-06-30
+tags: [MoTLabs Study]
+comments: true
+subdir: motnewletter-week18
+thumbnail: mot_logo.jpg
+share: true
+---
+
+
+## MOT News
+**There is Nothing new, but we are moving for a leap!**
+
+
+
+## MOT Study
+**1. 거북목 프로젝트 iOS 데모 (발표자: 곽도영)**
+- [[mot-ios-tensorflow github repo]](https://github.com/motlabs/mot-ios-tensorflow)
+- 해결한 이슈 및 구현된 기능
+
+```
+    - 입력 포맷의 타입 이슈 해결(unit8 ☞ float32)
+    - PoseEstimationForMobile을 ML Kit으로 구현(이슈해결)
+    - PoseEstimationForMobile을 Core ML으로 구현(Core ML 모델을 저장소 오너에게 받음)
+    - 실시간 추론(PoseEstimationForMobile + 카메라)
+```
+- 새로 생긴 이슈
+
+```
+    - 이 모델(PoseEstimationForMobile)은 상반신만 찍은 사진(우리가 사용할 데이터)은 거의 인식을 못함
+```
+- 다음 할 일
+
+```
+    - 전처리시 리사이징 할때 인터폴레이션 방법 조사(CoreGraphic)
+    - 밴치마크 기준 잡아서 평가해보기
+    - 후처리 성능 측정, 개선방향 조사
+    - point들을 점으로 연결해보기
+```
+
+**2. 거북목 프로젝트 Android 데모 (발표자: 신정아)**
+- [[mot-android-tensorflow github repo]](https://github.com/motlabs/mot-android-tensorflow)
+- Single pose estimation demo 앱 빌드
+- fps 측정 api 구현
+
+**3. To TFRecord (발표자: 이준호)**
+- [[코드링크]](https://github.com/motlabs/dont-be-turtle/blob/feature/fb_dataio/tfmodules/tfrecord_converter.py)
+
+
+**4. Single-shot detection TF API 소개 (발표자: 이성진)**
+- [[발표자료 링크]](https://drive.google.com/drive/folders/1TrdtYw9Hzo3ux80gjqXhkwnRb4lQOs6s)
+
+
+**5. 거북목 프로젝트 진행사항 (발표자 강재욱)**
+- [Hourglass 모델 아키텍쳐 설명](https://docs.google.com/document/d/1cOC9CZJv02WEz-v17FCQ-Up19kYP1ZtoadvpC_nLPPM/edit#heading=h.lv5patfw1lss)
+- [tflite Copatibility 가이드](https://docs.google.com/document/d/1cOC9CZJv02WEz-v17FCQ-Up19kYP1ZtoadvpC_nLPPM/edit#heading=h.lv5patfw1lss)
+- [tf slim api 가이드](https://docs.google.com/document/d/19ZiExIc-vdbGrauuRUPKDI_o3sbVTeYMifc_hzvULGA/edit)
+- 데이터 셋 수집 현황
+- [모바일 앱 요구사항 정리](https://docs.google.com/document/d/1eYztqAInoD-6CQTsen6_SV0xVEYO6FRgkO9Zvbsr8DY/edit#)
+
+
+
+## Lab Notices
+- 자율주행 랩 방문! (Welcome!)
+- 강재욱 양서연 김택민 구글 제주캠로 참석 (7/1~)
+    - 제주도에 오실 분들 일정 확인!
+- 다음주 부터 스카이프 미팅 (제주캠프 진행 상황 생중계)
+- 8월 중순에 새로운 프로젝트에 대한 아이디어 공모 중! 
+
+
+## Issues
+- 성능측정앱과 데모앱은 다른 프로젝트로 빌드하기 [(문서참)](https://docs.google.com/document/d/1eYztqAInoD-6CQTsen6_SV0xVEYO6FRgkO9Zvbsr8DY/edit#)
+- 팀원들의 지속적인 서포트 필요
+    - trainer.py / data_loader.py
+    - heatmap_generater.py
+    - ios / android
+
+{% include image.html subdir=page.subdir name='motweek18.jpg' caption='열심히! 연구하는 MoT' %}
+
+
+## Last Weekly Reports in MoT
+- [16-17주차 리포트 blog](https://motlabs.github.io/2018-06-22/motnewletter-week17/)
+- [14-15주차 리포트 blog](https://motlabs.github.io/2018-06-09/motnewletter-week15/)
+- [13 주차 리포트 gslide]( https://goo.gl/SPKZUp )
+- [9-12주차 리포트 gslide](https://goo.gl/SY6gHF)
+- [7-8주차 리포트 gslide]( https://goo.gl/CYwZM2 )
+- [6주차 리포트 gslide](https://goo.gl/GYPzzp)
+- [5주차 리포트 gslide]( https://goo.gl/tBny5m )
+- [4주차 리포트 gslide]( https://goo.gl/4JnQoc )
+- [3주차 리포트 gslide]( https://goo.gl/JGSztP )
+- [2주차 리포트 gslide]( https://goo.gl/T7o9tf )
+- [1주차 리포트 gslide]( https://goo.gl/nNqMYg )

_posts/2018-06-30-tfpattern-week4.md

@@ -0,0 +1,81 @@
+---
+layout: post
+title: "TF Pattern Design (Week4)"
+description: "Week4 Study Summary"
+author: jwkang
+date: 2018-06-30
+tags: [TF Pattern study]
+comments: true
+subdir: tfpattern-week4
+thumbnail: tensorflow_logo.png
+share: true
+---
+
+> Reviewed and edited by Jaewook Kang
+
+
+## 발표1: Recurrent Neural Network 이론 첫걸음!
+
+**Speaker : 양서연**
+
+### Slide
+
+<style>
+.responsive-wrap iframe{ max-width: 100%;}
+</style>
+<div class="responsive-wrap">
+<!-- this is the embed code provided by Google -->
+<iframe src="https://docs.google.com/presentation/d/e/2PACX-1vTlHpaHyVWuglvdwrxSEqNm8luAem6LEl6SliY2ZcE7dV_XZeRAk1vtP913ogWqAhNvnV-sWC6uB2Te/embed?start=false&loop=false&delayms=3000" frameborder="0" width="720" height="380" allowfullscreen="true" mozallowfullscreen="true" webkitallowfullscreen="true"></iframe>
+</div>
+
+
+### 요약
+---------------------------------------
+- RNN
+- Bi-directional RNN
+- Sequence to Sequence
+- LSTM
+- Attention
+
+
+
+## 발표2: To quickly implementing RNN
+
+**Speaker : 김보섭**
+
+### Slide
+[발표자료링크](https://drive.google.com/drive/folders/1ZDMJwMRID6KqEFzTPcALC1W0y0wfg-B1)
+
+
+### 요약
+---------------------------------------
+- TBU
+
+
+
+## 발표3: Recurrent Neural Network 구현 첫걸음!
+
+**Speaker : 류지은**
+
+### Slide
+
+<style>
+.responsive-wrap iframe{ max-width: 100%;}
+</style>
+<div class="responsive-wrap">
+<!-- this is the embed code provided by Google -->
+<iframe src="https://docs.google.com/presentation/d/e/2PACX-1vS6iUPjZoze-pfS6oU7581bBF4QqNtdrElhfPiNfp0PL74wCfF2QuSjEqhXtN3Y4-rrj1yRESHpq07b/embed?start=false&loop=false&delayms=3000" frameborder="0" width="720" height="380" allowfullscreen="true" mozallowfullscreen="true" webkitallowfullscreen="true"></iframe>
+</div>
+
+### 요약
+---------------------------------------
+- tf.scan
+- tf.nn.dynamic
+- tf.contrib.rnn.BasicRNNCell
+- tf.contrib.rnn.MultiRNNCell
+- tf.contrib.rnn.BasicLSTMCell
+
+
+
+## Issues
+- 이슈 없음

_posts/2018-06-30-tfpattern-week5.md

@@ -0,0 +1,70 @@
+---
+layout: post
+title: "TF Pattern Design (Week5)"
+description: "Week5 Study Summary"
+author: jwkang
+date: 2018-06-30
+tags: [TF Pattern study]
+comments: true
+subdir: tfpattern-week5
+thumbnail: tensorflow_logo.png
+share: true
+---
+
+> Reviewed and edited by Doyoung Kwak
+
+
+## 발표1: Hyperparameter & Model evaluation
+
+**Speaker : 백윤범**
+
+### Slide
+<style>
+.responsive-wrap iframe{ max-width: 100%;}
+</style>
+<div class="responsive-wrap">
+<!-- this is the embed code provided by Google -->
+<iframe src="https://docs.google.com/presentation/d/e/2PACX-1vRkH2uDd8fKQke6eLY0N_mDxgrpatqWLLpCFDxoYIR7g4hCXQGyMUK4IhZj3WTEFFXaLrwnuXiCSyAx/embed?start=true&loop=true&delayms=3000" frameborder="0" width="720" height="380" allowfullscreen="true" mozallowfullscreen="true" webkitallowfullscreen="true"></iframe>
+</div>
+
+### 요약
+---------------------------------------
+- Hyperparameter 란
+- Hyperparameter Optimization
+- Bayesian Optimization
+- Model 평가, 성능추정
+- Model selection
+- K fold cross validation
+
+
+
+
+## 발표2: 분산 Tensorflow + 구조화 구현
+
+**Speaker : 곽도영**
+
+### Slide
+<style>
+.responsive-wrap iframe{ max-width: 100%;}
+</style>
+<div class="responsive-wrap">
+<!-- this is the embed code provided by Google -->
+<iframe src="https://docs.google.com/presentation/d/e/2PACX-1vTW7DQiGj-57t5TAnmzINIDZhtqkqsT88s3GBRhDp9j1vnK69xgKJa_j553eo3z3Wm4egdxQxCYspPb/embed?start=false&loop=false&delayms=3000" frameborder="0" width="720" height="380" allowfullscreen="true" mozallowfullscreen="true" webkitallowfullscreen="true"></iframe>
+</div>
+
+
+### 요약
+---------------------------------------
+- Flags 메커니즘 소개
+- Ch9. 분산 텐서플로
+- 부록A.1.1 모델 구조화
+- ~~부록A.1.2 사용자 정의 손실 함수~~
+- 참고 프로젝트
+  - [https://github.com/carpedm20/multi-speaker-tacotron-tensorflow](https://github.com/carpedm20/multi-speaker-tacotron-tensorflow)
+  - [https://github.com/Hezi-Resheff/Oreilly-Learning-TensorFlow](https://github.com/Hezi-Resheff/Oreilly-Learning-TensorFlow)
+
+
+
+
+## Issues
+- `@property` 같은 데코레이터 문법(파이썬) 용도가 뭔지?