Posts by Category

DevCourse2

Git Stash

less than 1 minute read

git stash

Git Credential

less than 1 minute read

git credential둜 μ•”ν˜Έν‚€ 맀번 μž…λ ₯μ•ˆν•΄λ„ μˆ¨κΈ΄μƒνƒœμ—μ„œ git clone κ°€λŠ₯ν•˜κ²Œ ν•  수 μžˆλ‹€.

K-Fashion preprocessing

less than 1 minute read

K-Fashion Dataset Preprocessing To-do list json: coco format으둜 json λ§Œλ“€μ–΄ μ€˜μ•Ό 함 mv: data root μž¬μ„€μ • rename: νŒŒμΌμ΄λ¦„μ— 곡백과 κ΄„ν˜Έ 등이 쑴재

Faster RCNN

less than 1 minute read

Appendix Reference https://medium.datadriveninvestor.com/review-on-fast-rcnn-202c9eadd23b 2021 Lecture 15 Detection and Segmentation http://cs231n.stanfor...

NLP - Transformer and BERT

10 minute read

ML λͺ¨λΈ 이해λ₯Ό μœ„ν•œ 팁 λ³΅μž‘ν•œ ML λͺ¨λΈμ„ μ΄ν•΄ν•˜κΈ° μœ„ν•΄μ„œ λ‹€μŒκ³Ό 같은 μˆœμ„œλ₯Ό λ”°λΌκ°ˆ 것 이 λͺ¨λΈμ΄ ν’€λ €κ³  ν•˜λŠ” λ¬Έμ œλŠ” 무엇인가 μΆ”λ‘  단계λ₯Ό λ¨Όμ € 이해. β€œν•™μŠ΅β€λœ λͺ¨λΈμ΄ μ–΄λ–€ β€œμž…λ ₯”을 λ°›μ•„μ–΄ μ–΄λ–€ β€œμΆœλ ₯”을 κ³„μ‚°ν•˜λŠ”μ§€ νŒŒμ•…ν•  것 ...

Apply K-fashion 3rd solution to deepfashion2 dataset

1 minute read

3λ“± μ†”λ£¨μ…˜μ—μ„œ DetectoRS λ₯Ό μ“΄ 이유?(μΆ”μΈ‘) instance-segmentationμ—μ„œ κ·Έ λ‹Ήμ‹œ SOTA μ˜€κΈ° λ•Œλ¬Έμ— https://paperswithcode.com/sota/instance-segmentation-on-coco

NLP - Word Embedding

8 minute read

Word Embedding λ‹¨μ–΄μ˜ 의미λ₯Ό μ–΄λ–»κ²Œ λ‚˜νƒ€λ‚Ό 것인가 쒋은 ν‘œν˜„ 방식: λ‹¨μ–΄κ°„μ˜ 관계λ₯Ό 잘 ν‘œν˜„ν•  수 μžˆμ–΄μ•Ό 함

NLP - Document Classification

6 minute read

λ¬Έμ„œ λΆ„λ₯˜ (Text Classification) λ¬Έμ„œ λΆ„λ₯˜λž€ μž…λ ₯으둜 λ°›μ•„, ν…μŠ€νŠΈκ°€ μ–΄λ–€ μ’…λ₯˜μ˜ 범주에 μ†ν•˜λŠ”μ§€λ₯Ό κ΅¬λΆ„ν•˜λŠ” μž‘μ—… λ‹€μ–‘ν•œ λ¬Έμ„œ λΆ„λ₯˜ 문제 λ¬Έμ„œμ˜ λ²”μ£Ό, 주제 λΆ„λ₯˜ 이메일 슀팸 λΆ„λ₯˜ 감성 λΆ„λ₯˜ μ–Έμ–΄ λΆ„λ₯˜ ...

Docker Build

less than 1 minute read

Build Docker Image & run, start, stop and exit container by using shell command $ bash docker/build.sh

NLP - Language Model

5 minute read

μ–Έμ–΄λͺ¨λΈ λ‹€μŒ λ¬Έμž₯ λ‹€μŒμ— μ΄μ–΄μ§ˆ λ‹¨μ–΄λŠ”? Please turn your homework in or the ? λ‹€μŒ 두 λ¬Έμž₯ 쀑 λ‚˜νƒ€λ‚  ν™•λ₯ μ΄ 더 높은 것은? ...

NLP - text preprocessing

5 minute read

Intro μžμ—°μ–΄μ˜ 의미λ₯Ό μ»΄ν“¨ν„°λ‘œ λΆ„μ„ν•΄μ„œ νŠΉμ •μž‘μ—…μ„ μœ„ν•΄ μ‚¬μš©ν•  수 μžˆλ„λ‘ ν•˜λŠ” 것 μ‘μš©λΆ„μ•Ό κΈ°κ³„λ²ˆμ—­ 감성뢄석 λ¬Έμ„œλΆ„λ₯˜ μ§ˆμ˜μ‘λ‹΅μ‹œμŠ€ν…œ 챗봇 언어생성 μŒμ„±μΈμ‹ μΆ”μ²œμ‹œμŠ€ν…œ ...

Big data - ML Pipelineκ³Ό Tuning μ†Œκ°œμ™€ μ‹€μŠ΅

4 minute read

Spark MLlib λͺ¨λΈ νŠœλ‹ μ†Œκ°œ Spark MLlib λͺ¨λΈ νŠœλ‹ 졜적의 ν•˜μ΄νΌ νŒŒλΌλ―Έν„° 선택 졜적의 λͺ¨λΈ ν˜Ήμ€ λͺ¨λΈμ˜ νŒŒλΌλ―Έν„°λ₯Ό μ°ΎλŠ” 것이 μ•„μ£Ό μ€‘μš” ν•˜λ‚˜μ”© ν…ŒμŠ€νŠΈν•΄λ³΄λŠ” 것 vs. λ‹€μˆ˜λ₯Ό λ™μ‹œ ν…ŒμŠ€νŠΈ ν•˜λŠ” 것 λͺ¨λΈ μ„ νƒμ˜ μ€‘μš”ν•œ ...

07-22 Live Session

less than 1 minute read

Spark μ΄λž€ Pandas의 μŠ€ν…Œλ‘œμ΄λ“œ 버전 + SQL + Scikit-learn + Streaming

Big data - SparkSQL을 μ΄μš©ν•œ 데이터 뢄석

10 minute read

컀리어 이야기 남과 λΉ„κ΅ν•˜μ§€ 말고 μ•žμœΌλ‘œ 20-30년을 보기 ν•˜λ‚˜λ₯Ό ν•˜κΈ°λ‘œ ν–ˆμœΌλ©΄ 적어도 6κ°œμ›”μ€ 파고 λ“€κΈ° λ„ˆλ¬΄ 빨리 ν¬κΈ°ν•˜μ§€ μ•ŠκΈ° λ­”κ°€ 잘 μ•ˆλ˜λ©΄ μ„œλ‘λ₯΄κΈ° λ³΄λ‹€λŠ” 였히렀 천천히 κ°€κΈ° 곡뢀λ₯Ό μœ„ν•œ 곡뢀λ₯Ό ν•˜κΈ° λ³΄λ‹€λŠ” 일을 ...

Big data - Spark

6 minute read

빅데이터 κΈ°μˆ μ΄λž€ Spark μ†Œκ°œ νŒλ‹€μŠ€μ™€ 비ꡐ Spark μ‹€μŠ΅

Big data

6 minute read

데이터 νŒ€μ˜ μ—­ν•  데이터 νŒ€μ˜ λ―Έμ…˜ μ‹ λ’°ν•  수 μžˆλŠ” 데이터λ₯Ό λ°”νƒ•μœΌλ‘œ λΆ€κ°€κ°€μΉ˜ 생성

Project Fashion mmfashion

1 minute read

Get started Recommender Environment conda conda conda create -n fashion python=3.8 install pytorch c...

07-01 Live Session

3 minute read

Live Session 컀리어 μ–΄λ–€ κΈ°μˆ μ„ μ•„λŠλƒκ°€ μ•„λ‹Œ κ²°κ³Όλ₯Ό λ‚΄λŠ”λ° 초점

SQL Analysis - νŠΈλžœμž­μ…˜κ³Ό 기타 κ³ κΈ‰ SQL 문법

10 minute read

μˆ™μ œ1 리뷰 μ‚¬μš©μžλ³„λ‘œ 처음 채널과 λ§ˆμ§€λ§‰ 채널 μ•Œμ•„λ‚΄κΈ° ROW_NUMBER vs. FIRST_VALUE/LAST_VALUE μ‚¬μš©μž 251번의 μ‹œκ°„μˆœμœΌλ‘œ 봀을 λ•Œ 첫 번째 채널과 λ§ˆμ§€λ§‰ 채널은 무엇인가? μ•„λž˜ 쿼리λ₯Ό μ‹€ν–‰ν•΄μ„œ 처음과 λ§ˆμ§€λ§‰ 채널을 보면 ...

SQL Analysis - JOIN & 리뷰

10 minute read

JOIN SQL 쑰인은 두 개 ν˜Ήμ€ κ·Έ μ΄μƒμ˜ ν…Œμ΄λΈ”λ“€μ„ 곡톡 ν•„λ“œλ₯Ό 가지고 MERGE ν•˜λŠ”λ° μ‚¬μš©λ¨. μ΄λŠ” μŠ€νƒ€ μŠ€ν‚€λ§ˆλ‘œ κ΅¬μ„±λœ ν…Œμ΄λΈ”λ“€λ‘œ λΆ„μ‚°λ˜μ–΄ 있던 정보λ₯Ό ν†΅ν•©ν•˜λŠ” 데 μ‚¬μš©. JOIN의 κ²°κ³ΌλŠ” 방식에 상관없이 μ–‘μͺ½μ˜ ν•„λ“œλ₯Ό λͺ¨λ‘ 가진 μƒˆλ‘œμš΄ ν…Œμ΄λΈ”μ„ λ§Œλ“€μ–΄λ‚΄κ²Œ 됨. μ‘°...

SQL Analysis - GROUP BY & CTAS

4 minute read

GROUP BY와 AGGREGATE ν•¨μˆ˜ ν…Œμ΄λΈ”μ˜ λ ˆμ½”λ“œλ₯Ό κ·Έλ£Ήν•‘ν•˜μ—¬ κ·Έλ£Ήλ³„λ‘œ λ‹€μ–‘ν•œ 정보λ₯Ό 계산 μ΄λŠ” 두 λ‹¨κ³„λ‘œ 이루어짐 그룹핑을 ν•  ν•„λ“œλ₯Ό κ²°μ • (ν•˜λ‚˜ μ΄μƒμ˜ ν•„λ“œκ°€ 될 수 있음) GROUP BY둜 지정 (ν•„λ“œ 이름...

SQL Analysis - SELECT

5 minute read

Redshift 둠치 데λͺ¨ 예제 ν…Œμ΄λΈ” μ†Œκ°œ μ‚¬μš©μž ID: μ›Ήμ„œλΉ„μŠ€μ—μ„œ λ“±λ‘λœ μ‚¬μš©μžλ§ˆλ‹€ λΆ€μ—¬ν•˜λŠ” μœ μΌν•œ ID μ„Έμ…˜ ID: μ„Έμ…˜λ§ˆλ‹€ λΆ€μ—¬λ˜λŠ” ID μ„Έμ…˜: μ‚¬μš©μžμ˜ 방문을 논리적인 λ‹¨μœ„λ‘œ λ‚˜λˆˆ 것 μ‚¬μš©μžκ°€ λ°©λ¬Έ ν›„ 30λΆ„...

SQL Analysis - SQL & DB

4 minute read

SQL의 μ€‘μš”μ„± 데이터 κ΄€λ ¨ 3개의 직ꡰ 데이터 μ—”μ§€λ‹ˆμ–΄ Python, Java/Scalar SQL, DB ETL/ELT(airflow, DBT) Spark, Hadoop 데이터 뢄석가 ...

NN basics - RNN

5 minute read

Sequential Data ν‘œν˜„ λ²‘ν„°μ˜ 벑터 \(\boldsymbol{x} = (\boldsymbol{x}^{(1)}, \boldsymbol{x}^{(2)},\ldots ,\boldsymbol{x}^{(T)})^{T}\). 사전 ...

NN basics - DL Optimization

9 minute read

Objective Function Objective Funtion Vs. Loss Function(alias: Cost Function) β€œThe function we want to minimize or maximize is called the objective function, ...

cs231n Assignment

10 minute read

1. 예제 λ”°λΌν•˜κ³ , 개인 Github μ •λ¦¬ν•˜κΈ° CNN

07-01 Live Session

1 minute read

GAP (Global Average Pooling) import torch.nn.functional as F x = F.adaptive_avg_pool2d(x, (1, 1))

NN basics - DL Basics

2 minute read

Computer Vision’s Difficulties κ΄€μ μ˜ λ³€ν™”λ‘œ 인해 같은 λŒ€μƒμ΄λΌλ„ 픽셀값이 닀름. λ°°κ²½κ³Ό ν”Όμ‚¬μ²΄μ˜ ꡬ뢄이 어렀움. μ‘°λͺ…에 λ”°λ₯Έ λ³€ν™” λŒ€μƒμ˜ κΈ°ν˜•μ  ν˜•νƒœ 가렀진 ν˜•νƒœ 같은 ν΄λž˜μŠ€μ•ˆμ—μ„œμ˜ λ³€ν™”κ°€ 큼

NN basics - MLP

5 minute read

νΌμ…‰νŠΈλ‘  νΌμ…‰νŠΈλ‘ μ˜ λ™μž‘ 내적에 μ˜ν•œ μ„ ν˜•κ³Ό ν™œμ„±ν•¨μˆ˜μ— μ˜ν•œ λΉ„μ„ ν˜•μ— μ˜ν•΄ 의미λ₯Ό ꡬ뢄. κ²°κ΅­μ—” κ³΅κ°„μ˜ 뢄할을 λ§Œλ“ λ‹€. μ„ ν˜•λŒ€μˆ˜μ—μ„œ wλΌλŠ” κ°€μ€‘μΉ˜μ— input을 λ‚΄μ ν•˜κ³  w에 projectionν–ˆμ„λ•Œ, κ·Έ 값이 0이면 μ§κ΅ν•˜λŠ” 것이고, 0을 κΈ°μ€€μœΌλ‘œ λΉ„μ„ ν˜•μ  μš”μ†Œλ‘œ +1, -1둜 ...

Live Session 0622

less than 1 minute read

Underprediction \(t_{i}\): target label(i-th sample) \(y_{i}\): model’s prediction

Assignment - KNN

3 minute read

This assignment basically comes from cs231n. https://cs231n.github.io/assignments2021/assignment1/#q1-k-nearest-neighbor-classifier

ML basics - Probability Distributions 1

6 minute read

밀도좔정(Density Estimation) \(N\)개의 관찰데이터(observations) \(\mathbf{x}_1,\ldots\mathbf{x}_N\)κ°€ μ£Όμ–΄μ‘Œμ„ λ•Œ λΆ„ν¬ν•¨μˆ˜ \(p(\mathbf{x})\)λ₯Ό μ°ΎλŠ” 것

Live Session 06/15

less than 1 minute read

Live Session with κ°•μ°½μ„± λ°•μ‚¬λ‹˜

ML basics - Linear Algebra

24 minute read

μ™œ μ„ ν˜•λŒ€μˆ˜λ₯Ό μ•Œμ•„μ•Ό ν•˜λŠ”κ°€?

ML basics - MSE & MAE

4 minute read

Dataset   Restaurant Location Cuisines AverageCost MinimumOrder Rating Votes Reviews De...

Django Chart Visualization

4 minute read

1. Add chart.js and jquery (layouts.html) This is currently located at head, but later I’ve to move those scripts down to the end of the body tag. <!-- ...

Django HTML Extension with Issue & Solution

less than 1 minute read

Django VScode Extension While I work with django-html files, everytime I had to type curly brackets and percentage sign. I found most popular django html ext...

Django Static with Debug Mode

less than 1 minute read

Django Static & Debug Mode Keep in mind. If Debug = False, even though you correctly impletmented your static files, Django do not load any of your stati...

Flask API with Swagger

less than 1 minute read

Flask run as debug mode $ export FLASK_ENV=development $ flask run if not set as dev mode, need to keep restart flask like: $ flask run

Git Large File push Error and Solution

less than 1 minute read

Plotting 과제 pushν•˜λ‹€κ°€ 100MB λ„˜λŠ” νŒŒμΌμ„ pushν•˜λ €κ³  ν•˜λ‹ˆ errorκ°€ 났닀. 100MB 이상 파일 올리렀면 Git LFS(Large File Storage)λ₯Ό μ‚¬μš©ν•΄μ•Ό ν•œλ‹€λŠ”λ° ꡳ이 μ˜¬λ¦¬μ§€ μ•Šμ•„λ„ λœλ‹€λ©΄, μ—λŸ¬λ§Œ ν•΄κ²°ν•˜κ³  λ‹€μ‹œ PUSH ν•˜κ³  μ‹Άλ‹€λ©΄ μ•„λž˜ 링크λ₯Ό...

N Queen

1 minute read

Backtracking λŒ€ν‘œ 문제 Backtracking λŒ€ν‘œμ  문제라고 ν•œλ‹€.

SVD

less than 1 minute read

In Linear Algebra Course, SVD is one of matrix decomposition methods. (such as LU, QR)

Back to top ↑

MathJax

Project Fashion mmfashion

1 minute read

Get started Recommender Environment conda conda conda create -n fashion python=3.8 install pytorch c...

SQL Analysis - νŠΈλžœμž­μ…˜κ³Ό 기타 κ³ κΈ‰ SQL 문법

10 minute read

μˆ™μ œ1 리뷰 μ‚¬μš©μžλ³„λ‘œ 처음 채널과 λ§ˆμ§€λ§‰ 채널 μ•Œμ•„λ‚΄κΈ° ROW_NUMBER vs. FIRST_VALUE/LAST_VALUE μ‚¬μš©μž 251번의 μ‹œκ°„μˆœμœΌλ‘œ 봀을 λ•Œ 첫 번째 채널과 λ§ˆμ§€λ§‰ 채널은 무엇인가? μ•„λž˜ 쿼리λ₯Ό μ‹€ν–‰ν•΄μ„œ 처음과 λ§ˆμ§€λ§‰ 채널을 보면 ...

SQL Analysis - JOIN & 리뷰

10 minute read

JOIN SQL 쑰인은 두 개 ν˜Ήμ€ κ·Έ μ΄μƒμ˜ ν…Œμ΄λΈ”λ“€μ„ 곡톡 ν•„λ“œλ₯Ό 가지고 MERGE ν•˜λŠ”λ° μ‚¬μš©λ¨. μ΄λŠ” μŠ€νƒ€ μŠ€ν‚€λ§ˆλ‘œ κ΅¬μ„±λœ ν…Œμ΄λΈ”λ“€λ‘œ λΆ„μ‚°λ˜μ–΄ 있던 정보λ₯Ό ν†΅ν•©ν•˜λŠ” 데 μ‚¬μš©. JOIN의 κ²°κ³ΌλŠ” 방식에 상관없이 μ–‘μͺ½μ˜ ν•„λ“œλ₯Ό λͺ¨λ‘ 가진 μƒˆλ‘œμš΄ ν…Œμ΄λΈ”μ„ λ§Œλ“€μ–΄λ‚΄κ²Œ 됨. μ‘°...

SQL Analysis - GROUP BY & CTAS

4 minute read

GROUP BY와 AGGREGATE ν•¨μˆ˜ ν…Œμ΄λΈ”μ˜ λ ˆμ½”λ“œλ₯Ό κ·Έλ£Ήν•‘ν•˜μ—¬ κ·Έλ£Ήλ³„λ‘œ λ‹€μ–‘ν•œ 정보λ₯Ό 계산 μ΄λŠ” 두 λ‹¨κ³„λ‘œ 이루어짐 그룹핑을 ν•  ν•„λ“œλ₯Ό κ²°μ • (ν•˜λ‚˜ μ΄μƒμ˜ ν•„λ“œκ°€ 될 수 있음) GROUP BY둜 지정 (ν•„λ“œ 이름...

SQL Analysis - SELECT

5 minute read

Redshift 둠치 데λͺ¨ 예제 ν…Œμ΄λΈ” μ†Œκ°œ μ‚¬μš©μž ID: μ›Ήμ„œλΉ„μŠ€μ—μ„œ λ“±λ‘λœ μ‚¬μš©μžλ§ˆλ‹€ λΆ€μ—¬ν•˜λŠ” μœ μΌν•œ ID μ„Έμ…˜ ID: μ„Έμ…˜λ§ˆλ‹€ λΆ€μ—¬λ˜λŠ” ID μ„Έμ…˜: μ‚¬μš©μžμ˜ 방문을 논리적인 λ‹¨μœ„λ‘œ λ‚˜λˆˆ 것 μ‚¬μš©μžκ°€ λ°©λ¬Έ ν›„ 30λΆ„...

SQL Analysis - SQL & DB

4 minute read

SQL의 μ€‘μš”μ„± 데이터 κ΄€λ ¨ 3개의 직ꡰ 데이터 μ—”μ§€λ‹ˆμ–΄ Python, Java/Scalar SQL, DB ETL/ELT(airflow, DBT) Spark, Hadoop 데이터 뢄석가 ...

NN basics - RNN

5 minute read

Sequential Data ν‘œν˜„ λ²‘ν„°μ˜ 벑터 \(\boldsymbol{x} = (\boldsymbol{x}^{(1)}, \boldsymbol{x}^{(2)},\ldots ,\boldsymbol{x}^{(T)})^{T}\). 사전 ...

NN basics - DL Optimization

9 minute read

Objective Function Objective Funtion Vs. Loss Function(alias: Cost Function) β€œThe function we want to minimize or maximize is called the objective function, ...

cs231n Assignment

10 minute read

1. 예제 λ”°λΌν•˜κ³ , 개인 Github μ •λ¦¬ν•˜κΈ° CNN

NN basics - DL Basics

2 minute read

Computer Vision’s Difficulties κ΄€μ μ˜ λ³€ν™”λ‘œ 인해 같은 λŒ€μƒμ΄λΌλ„ 픽셀값이 닀름. λ°°κ²½κ³Ό ν”Όμ‚¬μ²΄μ˜ ꡬ뢄이 어렀움. μ‘°λͺ…에 λ”°λ₯Έ λ³€ν™” λŒ€μƒμ˜ κΈ°ν˜•μ  ν˜•νƒœ 가렀진 ν˜•νƒœ 같은 ν΄λž˜μŠ€μ•ˆμ—μ„œμ˜ λ³€ν™”κ°€ 큼

NN basics - MLP

5 minute read

νΌμ…‰νŠΈλ‘  νΌμ…‰νŠΈλ‘ μ˜ λ™μž‘ 내적에 μ˜ν•œ μ„ ν˜•κ³Ό ν™œμ„±ν•¨μˆ˜μ— μ˜ν•œ λΉ„μ„ ν˜•μ— μ˜ν•΄ 의미λ₯Ό ꡬ뢄. κ²°κ΅­μ—” κ³΅κ°„μ˜ 뢄할을 λ§Œλ“ λ‹€. μ„ ν˜•λŒ€μˆ˜μ—μ„œ wλΌλŠ” κ°€μ€‘μΉ˜μ— input을 λ‚΄μ ν•˜κ³  w에 projectionν–ˆμ„λ•Œ, κ·Έ 값이 0이면 μ§κ΅ν•˜λŠ” 것이고, 0을 κΈ°μ€€μœΌλ‘œ λΉ„μ„ ν˜•μ  μš”μ†Œλ‘œ +1, -1둜 ...

Assignment - KNN

3 minute read

This assignment basically comes from cs231n. https://cs231n.github.io/assignments2021/assignment1/#q1-k-nearest-neighbor-classifier

ML basics - Probability Distributions 1

6 minute read

밀도좔정(Density Estimation) \(N\)개의 관찰데이터(observations) \(\mathbf{x}_1,\ldots\mathbf{x}_N\)κ°€ μ£Όμ–΄μ‘Œμ„ λ•Œ λΆ„ν¬ν•¨μˆ˜ \(p(\mathbf{x})\)λ₯Ό μ°ΎλŠ” 것

ML basics - Linear Algebra

24 minute read

μ™œ μ„ ν˜•λŒ€μˆ˜λ₯Ό μ•Œμ•„μ•Ό ν•˜λŠ”κ°€?

ML basics - MSE & MAE

4 minute read

Dataset   Restaurant Location Cuisines AverageCost MinimumOrder Rating Votes Reviews De...

Aligning Equations in mathjax

less than 1 minute read

Align Equations in mathjax format \(\begin{align} \sqrt{37} & = \sqrt{\frac{73^2-1}{12^2}} \\ & = \sqrt{\frac{73^2}{12^2}\cdot\frac{73^2-1}{73^2}} \...

Back to top ↑

Docker

Docker Registry

less than 1 minute read

https://www.44bits.io/ko/post/running-docker-registry-and-using-s3-storage

Docker Redis

less than 1 minute read

Redis with Docker get redis image from docker hub Docker image pull redis redis-cli와 같이 ꡬ동해야 됨. 두 μ»¨ν…Œμ΄λ„ˆλ₯Ό μ—°κ²°...

Bitbucket Pipeline runners

less than 1 minute read

create pipeline write bitbucket-pipelines.yml commit pipelines.yml file run docker at local (test runners in local env) ...

Docker Container push

less than 1 minute read

Dockerhub에 Push κΉŒμ§€ ν•˜λŠ” 경우 private repositoryλŠ” ν•œκ°œκΉŒμ§€λ§Œ κ°€λŠ₯ν•˜κ³  더 μ“°λ €λ©΄ 유료 ν”Œλžœμ— κ°€μž…ν•΄μ•Ό ν•œλ‹€.

Docker Jupyter lab access denied

less than 1 minute read

ν™˜κ²½ localμ—μ„œ server에 μžˆλŠ” docker containerμ—μ„œ jupyter lab으둜 접속

Docker Container memory shortage

less than 1 minute read

Error Message: RuntimeError: DataLoader worker (pid 4301) is killed by signal: Bus error. It is possible that dataloader's workers are out of shared memory. ...

Apply K-fashion 3rd solution to deepfashion2 dataset

1 minute read

3λ“± μ†”λ£¨μ…˜μ—μ„œ DetectoRS λ₯Ό μ“΄ 이유?(μΆ”μΈ‘) instance-segmentationμ—μ„œ κ·Έ λ‹Ήμ‹œ SOTA μ˜€κΈ° λ•Œλ¬Έμ— https://paperswithcode.com/sota/instance-segmentation-on-coco

Docker gpu and jupyter lab

less than 1 minute read

λ„μ»€μ—μ„œ Container ν¬νŠΈμ™€ Host 포트의 κ°œλ…

Docker Build

less than 1 minute read

Build Docker Image & run, start, stop and exit container by using shell command $ bash docker/build.sh

Docker install on ubuntu

5 minute read

Setup the docker via Repository 1. Update the apt package index and install packages to allow apt to use a repository over HTTPS: $ sudo apt-get update $ sud...

Back to top ↑

DevCourse2_ML_Basics

ML basics - Probability Distributions 1

6 minute read

밀도좔정(Density Estimation) \(N\)개의 관찰데이터(observations) \(\mathbf{x}_1,\ldots\mathbf{x}_N\)κ°€ μ£Όμ–΄μ‘Œμ„ λ•Œ λΆ„ν¬ν•¨μˆ˜ \(p(\mathbf{x})\)λ₯Ό μ°ΎλŠ” 것

ML basics - Linear Algebra

24 minute read

μ™œ μ„ ν˜•λŒ€μˆ˜λ₯Ό μ•Œμ•„μ•Ό ν•˜λŠ”κ°€?

ML basics - MSE & MAE

4 minute read

Dataset   Restaurant Location Cuisines AverageCost MinimumOrder Rating Votes Reviews De...

Back to top ↑

PythonAlgorithms

Back to top ↑

MinimalMistakes

Markdown Named Anchor

less than 1 minute read

```markdown go to header 1 go to header 2 go to header 3

Jekyll unmodifided code block

less than 1 minute read

While writing django code snippet, I found that it hides double curly brackets( {{}} ). Since Jekyll uses the Liquid templating language to process templates...

Aligning Equations in mathjax

less than 1 minute read

Align Equations in mathjax format \(\begin{align} \sqrt{37} & = \sqrt{\frac{73^2-1}{12^2}} \\ & = \sqrt{\frac{73^2}{12^2}\cdot\frac{73^2-1}{73^2}} \...

였늘의 μ‚½μ§ˆ - insert image in minimal mistakes

less than 1 minute read

아무 생각 없이 이미지λ₯Ό _site/assets/images 에 이미지λ₯Ό μƒˆλ‘œ μ‚½μž…ν•˜λ €κ³  ν•˜λ‹ˆ 자꾸 이미지가 μ‚­μ œλ˜λŠ” κ²ƒμ΄μ˜€λ‹€β€¦ vscode remote connection이 λ¬Έμ œμΈκ°€ μ‹Άμ–΄ server κ°€μ„œ dir에닀 넣어도 보고 connection도 λ‹€μ‹œ ν•΄λ΄€μ§€λ§Œ 계속 μ‚­μ œ 되길래

Back to top ↑

Python

name == main

4 minute read

if name == β€œmain”: To better understand why and how this matters, we need to take a step back to understand how Python initializes scripts and how this inter...

Python GMP

less than 1 minute read

Python GMP

Pytorch IterableDataset

less than 1 minute read

https://inmoonlight.github.io/2021/02/21/PyTorch-IterableDataset/

python concurrent futures

less than 1 minute read

```py import time data = [] start = time.time() from concurrent.futures import ThreadPoolExecutor import concurrent.futures

Python underbar underscore naminig

less than 1 minute read

4-1. 언더바가 μ•žμ— ν•˜λ‚˜ 뢙은 경우 (ex: _variable) 이 κ²½μš°λŠ” λͺ¨λ“ˆ λ‚΄μ—μ„œλ§Œ ν•΄λ‹Ή λ³€μˆ˜/ν•¨μˆ˜λ₯Ό μ‚¬μš©ν•˜κ² λ‹€λŠ” μ˜λ―Έμž…λ‹ˆλ‹€.(private) 특히 ν˜‘μ—…μ„ ν•  λ•Œ λ‹€λ₯Έ νŒ€μ›μ—κ²Œ β€˜μ΄ λ³€μˆ˜/ν•¨μˆ˜λŠ” 이 λͺ¨λ“ˆ λ‚΄λΆ€μ—μ„œλ§Œ μ‚¬μš©ν•  거닀’라고 λͺ…μ‹œμ μœΌλ‘œ 힌트λ₯Ό 쀄 수 μžˆμŠ΅λ‹ˆλ‹€.

urllib parse unquote

less than 1 minute read

urllib.parse.unquote(string, encoding=’utf-8’, errors=’replace’) Replace %xx escapes with their single-character equivalent. The optional encoding and err...

Python Regex

less than 1 minute read

Regex 자주 μ‚¬μš©ν•˜λŠ” 문자 클래슀

Rename files with Python using os.rename

less than 1 minute read

```py import os def changeName(path, cName): i = 1 for filename in os.listdir(path): print(path+filename, β€˜=>’, path+str(cName)+str(i)+’....

Back to top ↑

FashionGAN_Proj

K-Fashion preprocessing

less than 1 minute read

K-Fashion Dataset Preprocessing To-do list json: coco format으둜 json λ§Œλ“€μ–΄ μ€˜μ•Ό 함 mv: data root μž¬μ„€μ • rename: νŒŒμΌμ΄λ¦„μ— 곡백과 κ΄„ν˜Έ 등이 쑴재

Apply K-fashion 3rd solution to deepfashion2 dataset

1 minute read

3λ“± μ†”λ£¨μ…˜μ—μ„œ DetectoRS λ₯Ό μ“΄ 이유?(μΆ”μΈ‘) instance-segmentationμ—μ„œ κ·Έ λ‹Ήμ‹œ SOTA μ˜€κΈ° λ•Œλ¬Έμ— https://paperswithcode.com/sota/instance-segmentation-on-coco

Project Fashion mmfashion

1 minute read

Get started Recommender Environment conda conda conda create -n fashion python=3.8 install pytorch c...

Back to top ↑

Competition

Kaggle - Sartorius Cell Instance Segmentation

less than 1 minute read

Overall Entered around 2 weeks before the competition ends. Get used to implementing Detectron2 with custom data First Instance Segmentation Task

Heart Datathon 2021

less than 1 minute read

Detection Metrics Dice Coefficient \[Dice = \frac{2 |A∩B|}{(|A|+|B|)} = \frac{2 TP}{(2 TP + FP + FN)}\]

NoonBody SemanticSegmentation

1 minute read

Task: Semantic Segmentation of Body labels are come from DensePose. Using these body labels, inference body parts.

Conda Vs. Pip

less than 1 minute read

conda 둜 ν™˜κ²½ λ§Œλ“€μ–΄μ„œ torch install을 conda둜 ν–ˆλ”λ‹ˆ cuda Falseκ°€ λ˜λ²„λ €μ„œ 버전 λ¬Έμ œμΈμ€„ μ•Œμ•˜λ”λ‹ˆ pip install 둜 ν•˜λ‹ˆ 문제 없이 cuda μ‚¬μš© κ°€λŠ₯ν•˜κ²Œ λ˜μ—ˆλ‹€.

OpenSlide Usages

less than 1 minute read

OpenSlide lib μ‚¬μš©λ°©λ²•

Setup Env with conda

less than 1 minute read

λŒ€νšŒλ§ˆλ‹€ νŒ¨ν‚€μ§€ 버전 μ΄μŠˆκ°€ μžˆμ„ 수 있기 λ•Œλ¬Έμ— λŒ€νšŒ μ°Έμ—¬ν•  λ•Œλ§ˆλ‹€ λ‹€λ₯Έ conda ν™˜κ²½μ„ λ§Œλ“€μ–΄μ£Όκ³  μ‹œμž‘ν•˜λŠ”κ²Œ μ’‹λ‹€κ³  μƒκ°ν•œλ‹€.

Setup DL Environment in Competition Container

less than 1 minute read

의료 λŒ€νšŒμ—μ„œ container둜 μ§„ν–‰ν•˜λ‹€λ³΄λ‹ˆ ν•„μš”ν•œ νŒŒμΌλ“€μ„ κ°€μƒν™˜κ²½μ²˜λŸΌ μ„€μΉ˜ν•œ ν›„ μ»¨ν…Œμ΄λ„ˆ μž¬μ‹œμž‘μ„ ν•˜λ©΄ λ‹€ λ‚ μ•„κ°€ 버린닀. 이에 따라 마운트 된 backup directory에 μ„€μΉ˜ 및 μ €μž₯ ν›„ μ‚¬μš©ν•΄μ•Ό ν•˜λŠ”λ°

Back to top ↑

Django

Django Chart Visualization

4 minute read

1. Add chart.js and jquery (layouts.html) This is currently located at head, but later I’ve to move those scripts down to the end of the body tag. <!-- ...

Django HTML Extension with Issue & Solution

less than 1 minute read

Django VScode Extension While I work with django-html files, everytime I had to type curly brackets and percentage sign. I found most popular django html ext...

Django Static with Debug Mode

less than 1 minute read

Django Static & Debug Mode Keep in mind. If Debug = False, even though you correctly impletmented your static files, Django do not load any of your stati...

Back to top ↑

DevCourse2_Django

Django Chart Visualization

4 minute read

1. Add chart.js and jquery (layouts.html) This is currently located at head, but later I’ve to move those scripts down to the end of the body tag. <!-- ...

Django HTML Extension with Issue & Solution

less than 1 minute read

Django VScode Extension While I work with django-html files, everytime I had to type curly brackets and percentage sign. I found most popular django html ext...

Django Static with Debug Mode

less than 1 minute read

Django Static & Debug Mode Keep in mind. If Debug = False, even though you correctly impletmented your static files, Django do not load any of your stati...

Back to top ↑

DL

Project Fashion mmfashion

1 minute read

Get started Recommender Environment conda conda conda create -n fashion python=3.8 install pytorch c...

NN basics - RNN

5 minute read

Sequential Data ν‘œν˜„ λ²‘ν„°μ˜ 벑터 \(\boldsymbol{x} = (\boldsymbol{x}^{(1)}, \boldsymbol{x}^{(2)},\ldots ,\boldsymbol{x}^{(T)})^{T}\). 사전 ...

NN basics - DL Optimization

9 minute read

Objective Function Objective Funtion Vs. Loss Function(alias: Cost Function) β€œThe function we want to minimize or maximize is called the objective function, ...

NN basics - DL Basics

2 minute read

Computer Vision’s Difficulties κ΄€μ μ˜ λ³€ν™”λ‘œ 인해 같은 λŒ€μƒμ΄λΌλ„ 픽셀값이 닀름. λ°°κ²½κ³Ό ν”Όμ‚¬μ²΄μ˜ ꡬ뢄이 어렀움. μ‘°λͺ…에 λ”°λ₯Έ λ³€ν™” λŒ€μƒμ˜ κΈ°ν˜•μ  ν˜•νƒœ 가렀진 ν˜•νƒœ 같은 ν΄λž˜μŠ€μ•ˆμ—μ„œμ˜ λ³€ν™”κ°€ 큼

Back to top ↑

NN

NN basics - RNN

5 minute read

Sequential Data ν‘œν˜„ λ²‘ν„°μ˜ 벑터 \(\boldsymbol{x} = (\boldsymbol{x}^{(1)}, \boldsymbol{x}^{(2)},\ldots ,\boldsymbol{x}^{(T)})^{T}\). 사전 ...

NN basics - DL Optimization

9 minute read

Objective Function Objective Funtion Vs. Loss Function(alias: Cost Function) β€œThe function we want to minimize or maximize is called the objective function, ...

cs231n Assignment

10 minute read

1. 예제 λ”°λΌν•˜κ³ , 개인 Github μ •λ¦¬ν•˜κΈ° CNN

NN basics - DL Basics

2 minute read

Computer Vision’s Difficulties κ΄€μ μ˜ λ³€ν™”λ‘œ 인해 같은 λŒ€μƒμ΄λΌλ„ 픽셀값이 닀름. λ°°κ²½κ³Ό ν”Όμ‚¬μ²΄μ˜ ꡬ뢄이 어렀움. μ‘°λͺ…에 λ”°λ₯Έ λ³€ν™” λŒ€μƒμ˜ κΈ°ν˜•μ  ν˜•νƒœ 가렀진 ν˜•νƒœ 같은 ν΄λž˜μŠ€μ•ˆμ—μ„œμ˜ λ³€ν™”κ°€ 큼

NN basics - MLP

5 minute read

νΌμ…‰νŠΈλ‘  νΌμ…‰νŠΈλ‘ μ˜ λ™μž‘ 내적에 μ˜ν•œ μ„ ν˜•κ³Ό ν™œμ„±ν•¨μˆ˜μ— μ˜ν•œ λΉ„μ„ ν˜•μ— μ˜ν•΄ 의미λ₯Ό ꡬ뢄. κ²°κ΅­μ—” κ³΅κ°„μ˜ 뢄할을 λ§Œλ“ λ‹€. μ„ ν˜•λŒ€μˆ˜μ—μ„œ wλΌλŠ” κ°€μ€‘μΉ˜μ— input을 λ‚΄μ ν•˜κ³  w에 projectionν–ˆμ„λ•Œ, κ·Έ 값이 0이면 μ§κ΅ν•˜λŠ” 것이고, 0을 κΈ°μ€€μœΌλ‘œ λΉ„μ„ ν˜•μ  μš”μ†Œλ‘œ +1, -1둜 ...

Back to top ↑

DevCourse2_NLP

Faster RCNN

less than 1 minute read

Appendix Reference https://medium.datadriveninvestor.com/review-on-fast-rcnn-202c9eadd23b 2021 Lecture 15 Detection and Segmentation http://cs231n.stanfor...

NLP - Transformer and BERT

10 minute read

ML λͺ¨λΈ 이해λ₯Ό μœ„ν•œ 팁 λ³΅μž‘ν•œ ML λͺ¨λΈμ„ μ΄ν•΄ν•˜κΈ° μœ„ν•΄μ„œ λ‹€μŒκ³Ό 같은 μˆœμ„œλ₯Ό λ”°λΌκ°ˆ 것 이 λͺ¨λΈμ΄ ν’€λ €κ³  ν•˜λŠ” λ¬Έμ œλŠ” 무엇인가 μΆ”λ‘  단계λ₯Ό λ¨Όμ € 이해. β€œν•™μŠ΅β€λœ λͺ¨λΈμ΄ μ–΄λ–€ β€œμž…λ ₯”을 λ°›μ•„μ–΄ μ–΄λ–€ β€œμΆœλ ₯”을 κ³„μ‚°ν•˜λŠ”μ§€ νŒŒμ•…ν•  것 ...

NLP - Word Embedding

8 minute read

Word Embedding λ‹¨μ–΄μ˜ 의미λ₯Ό μ–΄λ–»κ²Œ λ‚˜νƒ€λ‚Ό 것인가 쒋은 ν‘œν˜„ 방식: λ‹¨μ–΄κ°„μ˜ 관계λ₯Ό 잘 ν‘œν˜„ν•  수 μžˆμ–΄μ•Ό 함

NLP - Document Classification

6 minute read

λ¬Έμ„œ λΆ„λ₯˜ (Text Classification) λ¬Έμ„œ λΆ„λ₯˜λž€ μž…λ ₯으둜 λ°›μ•„, ν…μŠ€νŠΈκ°€ μ–΄λ–€ μ’…λ₯˜μ˜ 범주에 μ†ν•˜λŠ”μ§€λ₯Ό κ΅¬λΆ„ν•˜λŠ” μž‘μ—… λ‹€μ–‘ν•œ λ¬Έμ„œ λΆ„λ₯˜ 문제 λ¬Έμ„œμ˜ λ²”μ£Ό, 주제 λΆ„λ₯˜ 이메일 슀팸 λΆ„λ₯˜ 감성 λΆ„λ₯˜ μ–Έμ–΄ λΆ„λ₯˜ ...

NLP - Language Model

5 minute read

μ–Έμ–΄λͺ¨λΈ λ‹€μŒ λ¬Έμž₯ λ‹€μŒμ— μ΄μ–΄μ§ˆ λ‹¨μ–΄λŠ”? Please turn your homework in or the ? λ‹€μŒ 두 λ¬Έμž₯ 쀑 λ‚˜νƒ€λ‚  ν™•λ₯ μ΄ 더 높은 것은? ...

NLP - text preprocessing

5 minute read

Intro μžμ—°μ–΄μ˜ 의미λ₯Ό μ»΄ν“¨ν„°λ‘œ λΆ„μ„ν•΄μ„œ νŠΉμ •μž‘μ—…μ„ μœ„ν•΄ μ‚¬μš©ν•  수 μžˆλ„λ‘ ν•˜λŠ” 것 μ‘μš©λΆ„μ•Ό κΈ°κ³„λ²ˆμ—­ 감성뢄석 λ¬Έμ„œλΆ„λ₯˜ μ§ˆμ˜μ‘λ‹΅μ‹œμŠ€ν…œ 챗봇 언어생성 μŒμ„±μΈμ‹ μΆ”μ²œμ‹œμŠ€ν…œ ...

Back to top ↑

Misc

First-class citizen

less than 1 minute read

programming In programming language design, a first-class citizen (also type, object, entity, or value) in a given programming language is an entity which su...

vscode jupyter kernel

less than 1 minute read

python -m ipykernel install --user --name AI_hub_pre --display-name AI_hub_kernel

Linux tar

less than 1 minute read

extract tar

Download Gdrive file via wget or gdown

less than 1 minute read

Wget 링크 생성(링크가 μžˆλŠ” λͺ¨λ“  μ‚¬μš©μžμ—κ²Œ 곡개) wget μΏ ν‚€ μ„€μ •μœΌλ‘œ λ‹€μš΄λ‘œλ“œ wget --load-cookies ~/cookies.txt "https://docs.google.com/uc?export=download&confirm=$(wget -...

Back to top ↑

coursera

Back to top ↑

google

Back to top ↑

GoogleDataAnalytics

Back to top ↑

Segmentation

Kaggle - Sartorius Cell Instance Segmentation

less than 1 minute read

Overall Entered around 2 weeks before the competition ends. Get used to implementing Detectron2 with custom data First Instance Segmentation Task

Heart Datathon 2021

less than 1 minute read

Detection Metrics Dice Coefficient \[Dice = \frac{2 |A∩B|}{(|A|+|B|)} = \frac{2 TP}{(2 TP + FP + FN)}\]

NoonBody SemanticSegmentation

1 minute read

Task: Semantic Segmentation of Body labels are come from DensePose. Using these body labels, inference body parts.

Conda Vs. Pip

less than 1 minute read

conda 둜 ν™˜κ²½ λ§Œλ“€μ–΄μ„œ torch install을 conda둜 ν–ˆλ”λ‹ˆ cuda Falseκ°€ λ˜λ²„λ €μ„œ 버전 λ¬Έμ œμΈμ€„ μ•Œμ•˜λ”λ‹ˆ pip install 둜 ν•˜λ‹ˆ 문제 없이 cuda μ‚¬μš© κ°€λŠ₯ν•˜κ²Œ λ˜μ—ˆλ‹€.

OpenSlide Usages

less than 1 minute read

OpenSlide lib μ‚¬μš©λ°©λ²•

Setup Env with conda

less than 1 minute read

λŒ€νšŒλ§ˆλ‹€ νŒ¨ν‚€μ§€ 버전 μ΄μŠˆκ°€ μžˆμ„ 수 있기 λ•Œλ¬Έμ— λŒ€νšŒ μ°Έμ—¬ν•  λ•Œλ§ˆλ‹€ λ‹€λ₯Έ conda ν™˜κ²½μ„ λ§Œλ“€μ–΄μ£Όκ³  μ‹œμž‘ν•˜λŠ”κ²Œ μ’‹λ‹€κ³  μƒκ°ν•œλ‹€.

Setup DL Environment in Competition Container

less than 1 minute read

의료 λŒ€νšŒμ—μ„œ container둜 μ§„ν–‰ν•˜λ‹€λ³΄λ‹ˆ ν•„μš”ν•œ νŒŒμΌλ“€μ„ κ°€μƒν™˜κ²½μ²˜λŸΌ μ„€μΉ˜ν•œ ν›„ μ»¨ν…Œμ΄λ„ˆ μž¬μ‹œμž‘μ„ ν•˜λ©΄ λ‹€ λ‚ μ•„κ°€ 버린닀. 이에 따라 마운트 된 backup directory에 μ„€μΉ˜ 및 μ €μž₯ ν›„ μ‚¬μš©ν•΄μ•Ό ν•˜λŠ”λ°

Back to top ↑

LiveSession

07-22 Live Session

less than 1 minute read

Spark μ΄λž€ Pandas의 μŠ€ν…Œλ‘œμ΄λ“œ 버전 + SQL + Scikit-learn + Streaming

07-01 Live Session

3 minute read

Live Session 컀리어 μ–΄λ–€ κΈ°μˆ μ„ μ•„λŠλƒκ°€ μ•„λ‹Œ κ²°κ³Όλ₯Ό λ‚΄λŠ”λ° 초점

07-01 Live Session

1 minute read

GAP (Global Average Pooling) import torch.nn.functional as F x = F.adaptive_avg_pool2d(x, (1, 1))

Live Session 0622

less than 1 minute read

Underprediction \(t_{i}\): target label(i-th sample) \(y_{i}\): model’s prediction

Live Session 06/15

less than 1 minute read

Live Session with κ°•μ°½μ„± λ°•μ‚¬λ‹˜

Back to top ↑

DevCourse2_NN_Basics

NN basics - RNN

5 minute read

Sequential Data ν‘œν˜„ λ²‘ν„°μ˜ 벑터 \(\boldsymbol{x} = (\boldsymbol{x}^{(1)}, \boldsymbol{x}^{(2)},\ldots ,\boldsymbol{x}^{(T)})^{T}\). 사전 ...

NN basics - DL Optimization

9 minute read

Objective Function Objective Funtion Vs. Loss Function(alias: Cost Function) β€œThe function we want to minimize or maximize is called the objective function, ...

NN basics - DL Basics

2 minute read

Computer Vision’s Difficulties κ΄€μ μ˜ λ³€ν™”λ‘œ 인해 같은 λŒ€μƒμ΄λΌλ„ 픽셀값이 닀름. λ°°κ²½κ³Ό ν”Όμ‚¬μ²΄μ˜ ꡬ뢄이 어렀움. μ‘°λͺ…에 λ”°λ₯Έ λ³€ν™” λŒ€μƒμ˜ κΈ°ν˜•μ  ν˜•νƒœ 가렀진 ν˜•νƒœ 같은 ν΄λž˜μŠ€μ•ˆμ—μ„œμ˜ λ³€ν™”κ°€ 큼

NN basics - MLP

5 minute read

νΌμ…‰νŠΈλ‘  νΌμ…‰νŠΈλ‘ μ˜ λ™μž‘ 내적에 μ˜ν•œ μ„ ν˜•κ³Ό ν™œμ„±ν•¨μˆ˜μ— μ˜ν•œ λΉ„μ„ ν˜•μ— μ˜ν•΄ 의미λ₯Ό ꡬ뢄. κ²°κ΅­μ—” κ³΅κ°„μ˜ 뢄할을 λ§Œλ“ λ‹€. μ„ ν˜•λŒ€μˆ˜μ—μ„œ wλΌλŠ” κ°€μ€‘μΉ˜μ— input을 λ‚΄μ ν•˜κ³  w에 projectionν–ˆμ„λ•Œ, κ·Έ 값이 0이면 μ§κ΅ν•˜λŠ” 것이고, 0을 κΈ°μ€€μœΌλ‘œ λΉ„μ„ ν˜•μ  μš”μ†Œλ‘œ +1, -1둜 ...

Back to top ↑

ObjectDetection

Faster RCNN

less than 1 minute read

Appendix Reference https://medium.datadriveninvestor.com/review-on-fast-rcnn-202c9eadd23b 2021 Lecture 15 Detection and Segmentation http://cs231n.stanfor...

Back to top ↑

NLP

NLP - Transformer and BERT

10 minute read

ML λͺ¨λΈ 이해λ₯Ό μœ„ν•œ 팁 λ³΅μž‘ν•œ ML λͺ¨λΈμ„ μ΄ν•΄ν•˜κΈ° μœ„ν•΄μ„œ λ‹€μŒκ³Ό 같은 μˆœμ„œλ₯Ό λ”°λΌκ°ˆ 것 이 λͺ¨λΈμ΄ ν’€λ €κ³  ν•˜λŠ” λ¬Έμ œλŠ” 무엇인가 μΆ”λ‘  단계λ₯Ό λ¨Όμ € 이해. β€œν•™μŠ΅β€λœ λͺ¨λΈμ΄ μ–΄λ–€ β€œμž…λ ₯”을 λ°›μ•„μ–΄ μ–΄λ–€ β€œμΆœλ ₯”을 κ³„μ‚°ν•˜λŠ”μ§€ νŒŒμ•…ν•  것 ...

NLP - Word Embedding

8 minute read

Word Embedding λ‹¨μ–΄μ˜ 의미λ₯Ό μ–΄λ–»κ²Œ λ‚˜νƒ€λ‚Ό 것인가 쒋은 ν‘œν˜„ 방식: λ‹¨μ–΄κ°„μ˜ 관계λ₯Ό 잘 ν‘œν˜„ν•  수 μžˆμ–΄μ•Ό 함

NLP - Document Classification

6 minute read

λ¬Έμ„œ λΆ„λ₯˜ (Text Classification) λ¬Έμ„œ λΆ„λ₯˜λž€ μž…λ ₯으둜 λ°›μ•„, ν…μŠ€νŠΈκ°€ μ–΄λ–€ μ’…λ₯˜μ˜ 범주에 μ†ν•˜λŠ”μ§€λ₯Ό κ΅¬λΆ„ν•˜λŠ” μž‘μ—… λ‹€μ–‘ν•œ λ¬Έμ„œ λΆ„λ₯˜ 문제 λ¬Έμ„œμ˜ λ²”μ£Ό, 주제 λΆ„λ₯˜ 이메일 슀팸 λΆ„λ₯˜ 감성 λΆ„λ₯˜ μ–Έμ–΄ λΆ„λ₯˜ ...

NLP - Language Model

5 minute read

μ–Έμ–΄λͺ¨λΈ λ‹€μŒ λ¬Έμž₯ λ‹€μŒμ— μ΄μ–΄μ§ˆ λ‹¨μ–΄λŠ”? Please turn your homework in or the ? λ‹€μŒ 두 λ¬Έμž₯ 쀑 λ‚˜νƒ€λ‚  ν™•λ₯ μ΄ 더 높은 것은? ...

NLP - text preprocessing

5 minute read

Intro μžμ—°μ–΄μ˜ 의미λ₯Ό μ»΄ν“¨ν„°λ‘œ λΆ„μ„ν•΄μ„œ νŠΉμ •μž‘μ—…μ„ μœ„ν•΄ μ‚¬μš©ν•  수 μžˆλ„λ‘ ν•˜λŠ” 것 μ‘μš©λΆ„μ•Ό κΈ°κ³„λ²ˆμ—­ 감성뢄석 λ¬Έμ„œλΆ„λ₯˜ μ§ˆμ˜μ‘λ‹΅μ‹œμŠ€ν…œ 챗봇 언어생성 μŒμ„±μΈμ‹ μΆ”μ²œμ‹œμŠ€ν…œ ...

Back to top ↑

Algorithm

Heap Sort Implementation

2 minute read

μ½”λ“œ κ΅¬ν˜„ 원칙 μˆ˜λ„μ½”λ“œλ³΄κ³  κ΅¬ν˜„ μ‹€νŒ¨μ‹œ, μ½”λ“œλ³΄κ³  μ½”λ©˜νŠΈ 달고 μž‘μ„±ν•΄λ³΄κΈ° μž¬κ·€λ‘œ κ΅¬ν˜„ μ™„μ„± iterative version κ΅¬ν˜„ν•˜κΈ°

Quick Sort Implementation

1 minute read

Pseudo Code 기쀀점(pivot)을 μ •ν•œλ‹€. (4가지 방법 쑴재) Always pick first element as pivot. Always pick last element as pivot (implemented below) ...

Bubble Sort Implementation

less than 1 minute read

Pseudo Code adjacent elements 끼리 λΉ„κ΅ν•œλ‹€. μ–΄λ–»κ²Œ sortedκ°€ λ˜μ—ˆλŠ”μ§€ νŒλ‹¨ν•˜λŠ”κ°€(μ’…λ£Œμ‘°κ±΄) 2-1. 첫 passλ₯Ό ν†΅κ³Όν•˜λ©΄ λ§ˆμ§€λ§‰ μ›μ†ŒλŠ” 무쑰건 κ°€μž₯ 큰 μ›μ†Œκ°€ 됨 2-2. λ”°λΌμ„œ pass ν•  λ•Œλ§ˆλ‹€ λ’€μ—μ„œλΆ€ν„° 정렬이 λ˜λŠ” μ•Œκ³ λ¦¬μ¦˜ 2-...

N Queen

1 minute read

Backtracking λŒ€ν‘œ 문제 Backtracking λŒ€ν‘œμ  문제라고 ν•œλ‹€.

Back to top ↑

Git

Git Reset Modes: soft, mixed, hard

less than 1 minute read

commit을 resetν•΄μ„œ 이전 commit μ§€μ μœΌλ‘œ λŒμ•„κ°ˆ ν•„μš”κ°€ μžˆλŠ” κ²½μš°κ°€ μ‘΄μž¬ν•œλ‹€.

Git checkout submodule

less than 1 minute read

λ‹€λ₯Έ branch둜 checkout ν• λ•Œ submodule 이 없을 경우 submodule init을 ν•΄μ£Όκ³  update –recursiveλ₯Ό ν•΄μ€˜μ•Ό 함.

Git checkout from commit

less than 1 minute read

νŠΉμ • commit으둜 λΆ€ν„° μƒˆλ‘œμš΄ 브랜치λ₯Ό checkout ν•˜λŠ” cmd

Git Credential

less than 1 minute read

git credential둜 μ•”ν˜Έν‚€ 맀번 μž…λ ₯μ•ˆν•΄λ„ μˆ¨κΈ΄μƒνƒœμ—μ„œ git clone κ°€λŠ₯ν•˜κ²Œ ν•  수 μžˆλ‹€.

Back to top ↑

SQL

SQL Analysis - νŠΈλžœμž­μ…˜κ³Ό 기타 κ³ κΈ‰ SQL 문법

10 minute read

μˆ™μ œ1 리뷰 μ‚¬μš©μžλ³„λ‘œ 처음 채널과 λ§ˆμ§€λ§‰ 채널 μ•Œμ•„λ‚΄κΈ° ROW_NUMBER vs. FIRST_VALUE/LAST_VALUE μ‚¬μš©μž 251번의 μ‹œκ°„μˆœμœΌλ‘œ 봀을 λ•Œ 첫 번째 채널과 λ§ˆμ§€λ§‰ 채널은 무엇인가? μ•„λž˜ 쿼리λ₯Ό μ‹€ν–‰ν•΄μ„œ 처음과 λ§ˆμ§€λ§‰ 채널을 보면 ...

SQL Analysis - JOIN & 리뷰

10 minute read

JOIN SQL 쑰인은 두 개 ν˜Ήμ€ κ·Έ μ΄μƒμ˜ ν…Œμ΄λΈ”λ“€μ„ 곡톡 ν•„λ“œλ₯Ό 가지고 MERGE ν•˜λŠ”λ° μ‚¬μš©λ¨. μ΄λŠ” μŠ€νƒ€ μŠ€ν‚€λ§ˆλ‘œ κ΅¬μ„±λœ ν…Œμ΄λΈ”λ“€λ‘œ λΆ„μ‚°λ˜μ–΄ 있던 정보λ₯Ό ν†΅ν•©ν•˜λŠ” 데 μ‚¬μš©. JOIN의 κ²°κ³ΌλŠ” 방식에 상관없이 μ–‘μͺ½μ˜ ν•„λ“œλ₯Ό λͺ¨λ‘ 가진 μƒˆλ‘œμš΄ ν…Œμ΄λΈ”μ„ λ§Œλ“€μ–΄λ‚΄κ²Œ 됨. μ‘°...

SQL Analysis - GROUP BY & CTAS

4 minute read

GROUP BY와 AGGREGATE ν•¨μˆ˜ ν…Œμ΄λΈ”μ˜ λ ˆμ½”λ“œλ₯Ό κ·Έλ£Ήν•‘ν•˜μ—¬ κ·Έλ£Ήλ³„λ‘œ λ‹€μ–‘ν•œ 정보λ₯Ό 계산 μ΄λŠ” 두 λ‹¨κ³„λ‘œ 이루어짐 그룹핑을 ν•  ν•„λ“œλ₯Ό κ²°μ • (ν•˜λ‚˜ μ΄μƒμ˜ ν•„λ“œκ°€ 될 수 있음) GROUP BY둜 지정 (ν•„λ“œ 이름...

SQL Analysis - SELECT

5 minute read

Redshift 둠치 데λͺ¨ 예제 ν…Œμ΄λΈ” μ†Œκ°œ μ‚¬μš©μž ID: μ›Ήμ„œλΉ„μŠ€μ—μ„œ λ“±λ‘λœ μ‚¬μš©μžλ§ˆλ‹€ λΆ€μ—¬ν•˜λŠ” μœ μΌν•œ ID μ„Έμ…˜ ID: μ„Έμ…˜λ§ˆλ‹€ λΆ€μ—¬λ˜λŠ” ID μ„Έμ…˜: μ‚¬μš©μžμ˜ 방문을 논리적인 λ‹¨μœ„λ‘œ λ‚˜λˆˆ 것 μ‚¬μš©μžκ°€ λ°©λ¬Έ ν›„ 30λΆ„...

SQL Analysis - SQL & DB

4 minute read

SQL의 μ€‘μš”μ„± 데이터 κ΄€λ ¨ 3개의 직ꡰ 데이터 μ—”μ§€λ‹ˆμ–΄ Python, Java/Scalar SQL, DB ETL/ELT(airflow, DBT) Spark, Hadoop 데이터 뢄석가 ...

Back to top ↑

DevCourse2_SQL

SQL Analysis - νŠΈλžœμž­μ…˜κ³Ό 기타 κ³ κΈ‰ SQL 문법

10 minute read

μˆ™μ œ1 리뷰 μ‚¬μš©μžλ³„λ‘œ 처음 채널과 λ§ˆμ§€λ§‰ 채널 μ•Œμ•„λ‚΄κΈ° ROW_NUMBER vs. FIRST_VALUE/LAST_VALUE μ‚¬μš©μž 251번의 μ‹œκ°„μˆœμœΌλ‘œ 봀을 λ•Œ 첫 번째 채널과 λ§ˆμ§€λ§‰ 채널은 무엇인가? μ•„λž˜ 쿼리λ₯Ό μ‹€ν–‰ν•΄μ„œ 처음과 λ§ˆμ§€λ§‰ 채널을 보면 ...

SQL Analysis - JOIN & 리뷰

10 minute read

JOIN SQL 쑰인은 두 개 ν˜Ήμ€ κ·Έ μ΄μƒμ˜ ν…Œμ΄λΈ”λ“€μ„ 곡톡 ν•„λ“œλ₯Ό 가지고 MERGE ν•˜λŠ”λ° μ‚¬μš©λ¨. μ΄λŠ” μŠ€νƒ€ μŠ€ν‚€λ§ˆλ‘œ κ΅¬μ„±λœ ν…Œμ΄λΈ”λ“€λ‘œ λΆ„μ‚°λ˜μ–΄ 있던 정보λ₯Ό ν†΅ν•©ν•˜λŠ” 데 μ‚¬μš©. JOIN의 κ²°κ³ΌλŠ” 방식에 상관없이 μ–‘μͺ½μ˜ ν•„λ“œλ₯Ό λͺ¨λ‘ 가진 μƒˆλ‘œμš΄ ν…Œμ΄λΈ”μ„ λ§Œλ“€μ–΄λ‚΄κ²Œ 됨. μ‘°...

SQL Analysis - GROUP BY & CTAS

4 minute read

GROUP BY와 AGGREGATE ν•¨μˆ˜ ν…Œμ΄λΈ”μ˜ λ ˆμ½”λ“œλ₯Ό κ·Έλ£Ήν•‘ν•˜μ—¬ κ·Έλ£Ήλ³„λ‘œ λ‹€μ–‘ν•œ 정보λ₯Ό 계산 μ΄λŠ” 두 λ‹¨κ³„λ‘œ 이루어짐 그룹핑을 ν•  ν•„λ“œλ₯Ό κ²°μ • (ν•˜λ‚˜ μ΄μƒμ˜ ν•„λ“œκ°€ 될 수 있음) GROUP BY둜 지정 (ν•„λ“œ 이름...

SQL Analysis - SELECT

5 minute read

Redshift 둠치 데λͺ¨ 예제 ν…Œμ΄λΈ” μ†Œκ°œ μ‚¬μš©μž ID: μ›Ήμ„œλΉ„μŠ€μ—μ„œ λ“±λ‘λœ μ‚¬μš©μžλ§ˆλ‹€ λΆ€μ—¬ν•˜λŠ” μœ μΌν•œ ID μ„Έμ…˜ ID: μ„Έμ…˜λ§ˆλ‹€ λΆ€μ—¬λ˜λŠ” ID μ„Έμ…˜: μ‚¬μš©μžμ˜ 방문을 논리적인 λ‹¨μœ„λ‘œ λ‚˜λˆˆ 것 μ‚¬μš©μžκ°€ λ°©λ¬Έ ν›„ 30λΆ„...

SQL Analysis - SQL & DB

4 minute read

SQL의 μ€‘μš”μ„± 데이터 κ΄€λ ¨ 3개의 직ꡰ 데이터 μ—”μ§€λ‹ˆμ–΄ Python, Java/Scalar SQL, DB ETL/ELT(airflow, DBT) Spark, Hadoop 데이터 뢄석가 ...

Back to top ↑

Spark

Big data - ML Pipelineκ³Ό Tuning μ†Œκ°œμ™€ μ‹€μŠ΅

4 minute read

Spark MLlib λͺ¨λΈ νŠœλ‹ μ†Œκ°œ Spark MLlib λͺ¨λΈ νŠœλ‹ 졜적의 ν•˜μ΄νΌ νŒŒλΌλ―Έν„° 선택 졜적의 λͺ¨λΈ ν˜Ήμ€ λͺ¨λΈμ˜ νŒŒλΌλ―Έν„°λ₯Ό μ°ΎλŠ” 것이 μ•„μ£Ό μ€‘μš” ν•˜λ‚˜μ”© ν…ŒμŠ€νŠΈν•΄λ³΄λŠ” 것 vs. λ‹€μˆ˜λ₯Ό λ™μ‹œ ν…ŒμŠ€νŠΈ ν•˜λŠ” 것 λͺ¨λΈ μ„ νƒμ˜ μ€‘μš”ν•œ ...

Big data - SparkSQL을 μ΄μš©ν•œ 데이터 뢄석

10 minute read

컀리어 이야기 남과 λΉ„κ΅ν•˜μ§€ 말고 μ•žμœΌλ‘œ 20-30년을 보기 ν•˜λ‚˜λ₯Ό ν•˜κΈ°λ‘œ ν–ˆμœΌλ©΄ 적어도 6κ°œμ›”μ€ 파고 λ“€κΈ° λ„ˆλ¬΄ 빨리 ν¬κΈ°ν•˜μ§€ μ•ŠκΈ° λ­”κ°€ 잘 μ•ˆλ˜λ©΄ μ„œλ‘λ₯΄κΈ° λ³΄λ‹€λŠ” 였히렀 천천히 κ°€κΈ° 곡뢀λ₯Ό μœ„ν•œ 곡뢀λ₯Ό ν•˜κΈ° λ³΄λ‹€λŠ” 일을 ...

Big data - Spark

6 minute read

빅데이터 κΈ°μˆ μ΄λž€ Spark μ†Œκ°œ νŒλ‹€μŠ€μ™€ 비ꡐ Spark μ‹€μŠ΅

Big data

6 minute read

데이터 νŒ€μ˜ μ—­ν•  데이터 νŒ€μ˜ λ―Έμ…˜ μ‹ λ’°ν•  수 μžˆλŠ” 데이터λ₯Ό λ°”νƒ•μœΌλ‘œ λΆ€κ°€κ°€μΉ˜ 생성

Back to top ↑

BigData

Big data - ML Pipelineκ³Ό Tuning μ†Œκ°œμ™€ μ‹€μŠ΅

4 minute read

Spark MLlib λͺ¨λΈ νŠœλ‹ μ†Œκ°œ Spark MLlib λͺ¨λΈ νŠœλ‹ 졜적의 ν•˜μ΄νΌ νŒŒλΌλ―Έν„° 선택 졜적의 λͺ¨λΈ ν˜Ήμ€ λͺ¨λΈμ˜ νŒŒλΌλ―Έν„°λ₯Ό μ°ΎλŠ” 것이 μ•„μ£Ό μ€‘μš” ν•˜λ‚˜μ”© ν…ŒμŠ€νŠΈν•΄λ³΄λŠ” 것 vs. λ‹€μˆ˜λ₯Ό λ™μ‹œ ν…ŒμŠ€νŠΈ ν•˜λŠ” 것 λͺ¨λΈ μ„ νƒμ˜ μ€‘μš”ν•œ ...

Big data - SparkSQL을 μ΄μš©ν•œ 데이터 뢄석

10 minute read

컀리어 이야기 남과 λΉ„κ΅ν•˜μ§€ 말고 μ•žμœΌλ‘œ 20-30년을 보기 ν•˜λ‚˜λ₯Ό ν•˜κΈ°λ‘œ ν–ˆμœΌλ©΄ 적어도 6κ°œμ›”μ€ 파고 λ“€κΈ° λ„ˆλ¬΄ 빨리 ν¬κΈ°ν•˜μ§€ μ•ŠκΈ° λ­”κ°€ 잘 μ•ˆλ˜λ©΄ μ„œλ‘λ₯΄κΈ° λ³΄λ‹€λŠ” 였히렀 천천히 κ°€κΈ° 곡뢀λ₯Ό μœ„ν•œ 곡뢀λ₯Ό ν•˜κΈ° λ³΄λ‹€λŠ” 일을 ...

Big data - Spark

6 minute read

빅데이터 κΈ°μˆ μ΄λž€ Spark μ†Œκ°œ νŒλ‹€μŠ€μ™€ 비ꡐ Spark μ‹€μŠ΅

Big data

6 minute read

데이터 νŒ€μ˜ μ—­ν•  데이터 νŒ€μ˜ λ―Έμ…˜ μ‹ λ’°ν•  수 μžˆλŠ” 데이터λ₯Ό λ°”νƒ•μœΌλ‘œ λΆ€κ°€κ°€μΉ˜ 생성

Back to top ↑

DevCourse2_Spark

Big data - ML Pipelineκ³Ό Tuning μ†Œκ°œμ™€ μ‹€μŠ΅

4 minute read

Spark MLlib λͺ¨λΈ νŠœλ‹ μ†Œκ°œ Spark MLlib λͺ¨λΈ νŠœλ‹ 졜적의 ν•˜μ΄νΌ νŒŒλΌλ―Έν„° 선택 졜적의 λͺ¨λΈ ν˜Ήμ€ λͺ¨λΈμ˜ νŒŒλΌλ―Έν„°λ₯Ό μ°ΎλŠ” 것이 μ•„μ£Ό μ€‘μš” ν•˜λ‚˜μ”© ν…ŒμŠ€νŠΈν•΄λ³΄λŠ” 것 vs. λ‹€μˆ˜λ₯Ό λ™μ‹œ ν…ŒμŠ€νŠΈ ν•˜λŠ” 것 λͺ¨λΈ μ„ νƒμ˜ μ€‘μš”ν•œ ...

Big data - SparkSQL을 μ΄μš©ν•œ 데이터 뢄석

10 minute read

컀리어 이야기 남과 λΉ„κ΅ν•˜μ§€ 말고 μ•žμœΌλ‘œ 20-30년을 보기 ν•˜λ‚˜λ₯Ό ν•˜κΈ°λ‘œ ν–ˆμœΌλ©΄ 적어도 6κ°œμ›”μ€ 파고 λ“€κΈ° λ„ˆλ¬΄ 빨리 ν¬κΈ°ν•˜μ§€ μ•ŠκΈ° λ­”κ°€ 잘 μ•ˆλ˜λ©΄ μ„œλ‘λ₯΄κΈ° λ³΄λ‹€λŠ” 였히렀 천천히 κ°€κΈ° 곡뢀λ₯Ό μœ„ν•œ 곡뢀λ₯Ό ν•˜κΈ° λ³΄λ‹€λŠ” 일을 ...

Big data - Spark

6 minute read

빅데이터 κΈ°μˆ μ΄λž€ Spark μ†Œκ°œ νŒλ‹€μŠ€μ™€ 비ꡐ Spark μ‹€μŠ΅

Big data

6 minute read

데이터 νŒ€μ˜ μ—­ν•  데이터 νŒ€μ˜ λ―Έμ…˜ μ‹ λ’°ν•  수 μžˆλŠ” 데이터λ₯Ό λ°”νƒ•μœΌλ‘œ λΆ€κ°€κ°€μΉ˜ 생성

Back to top ↑

Sort

Heap Sort Implementation

2 minute read

μ½”λ“œ κ΅¬ν˜„ 원칙 μˆ˜λ„μ½”λ“œλ³΄κ³  κ΅¬ν˜„ μ‹€νŒ¨μ‹œ, μ½”λ“œλ³΄κ³  μ½”λ©˜νŠΈ 달고 μž‘μ„±ν•΄λ³΄κΈ° μž¬κ·€λ‘œ κ΅¬ν˜„ μ™„μ„± iterative version κ΅¬ν˜„ν•˜κΈ°

Quick Sort Implementation

1 minute read

Pseudo Code 기쀀점(pivot)을 μ •ν•œλ‹€. (4가지 방법 쑴재) Always pick first element as pivot. Always pick last element as pivot (implemented below) ...

Bubble Sort Implementation

less than 1 minute read

Pseudo Code adjacent elements 끼리 λΉ„κ΅ν•œλ‹€. μ–΄λ–»κ²Œ sortedκ°€ λ˜μ—ˆλŠ”μ§€ νŒλ‹¨ν•˜λŠ”κ°€(μ’…λ£Œμ‘°κ±΄) 2-1. 첫 passλ₯Ό ν†΅κ³Όν•˜λ©΄ λ§ˆμ§€λ§‰ μ›μ†ŒλŠ” 무쑰건 κ°€μž₯ 큰 μ›μ†Œκ°€ 됨 2-2. λ”°λΌμ„œ pass ν•  λ•Œλ§ˆλ‹€ λ’€μ—μ„œλΆ€ν„° 정렬이 λ˜λŠ” μ•Œκ³ λ¦¬μ¦˜ 2-...

Back to top ↑

DevCourse2_DL_Math

SVD

less than 1 minute read

In Linear Algebra Course, SVD is one of matrix decomposition methods. (such as LU, QR)

Back to top ↑

Linux

Linux tar

less than 1 minute read

extract tar

Back to top ↑

CICD

ML Flow Tutorial

less than 1 minute read

MLOps 반볡적인 μž‘μ—…μ„ ν•΄μ†Œν•  수 μžˆλŠ” 방법이면 λ‹€ MLOps 의 μ˜μ—­μœΌλ‘œ κ°„μ£Όν•œλ‹€.

GitLab runner

less than 1 minute read

Git lab runner둜 이주 μ‹œμž‘

Runners and ci-cd

less than 1 minute read

https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners

Back to top ↑

GitLab

GitLab cicd starclass test

1 minute read

make test_run.py to pass with python uniitest 1-1. make cicd variables in settings.

GitLab runner

less than 1 minute read

bitbucket 용으둜 sshλ₯Ό ν•˜λ‚˜λ§Œ μ“°λ‹€κ°€

GitLab runner

less than 1 minute read

Git lab runner둜 이주 μ‹œμž‘

Back to top ↑

VSCode

VSCode Editor tabs to wrap

less than 1 minute read

Go to setting: ctrl + , search: workbench.editor.wrapTabs Turn to True(as checked)

Remote-ssh access in VSCode

1 minute read

Remote-ssh VScode install VSCODE press F1 Find Remote-SSH: Connect current window to Host (or Remote-SSH: Connect to Host) λ§Œμ•½ 졜초 config λ§Œλ“€μ–΄μ•Ό 되면 4-1...

Back to top ↑

Jekyll

Jekyll unmodifided code block

less than 1 minute read

While writing django code snippet, I found that it hides double curly brackets( {{}} ). Since Jekyll uses the Liquid templating language to process templates...

Back to top ↑

HTML

Django Chart Visualization

4 minute read

1. Add chart.js and jquery (layouts.html) This is currently located at head, but later I’ve to move those scripts down to the end of the body tag. <!-- ...

Django HTML Extension with Issue & Solution

less than 1 minute read

Django VScode Extension While I work with django-html files, everytime I had to type curly brackets and percentage sign. I found most popular django html ext...

Back to top ↑

Pathology

Conda Vs. Pip

less than 1 minute read

conda 둜 ν™˜κ²½ λ§Œλ“€μ–΄μ„œ torch install을 conda둜 ν–ˆλ”λ‹ˆ cuda Falseκ°€ λ˜λ²„λ €μ„œ 버전 λ¬Έμ œμΈμ€„ μ•Œμ•˜λ”λ‹ˆ pip install 둜 ν•˜λ‹ˆ 문제 없이 cuda μ‚¬μš© κ°€λŠ₯ν•˜κ²Œ λ˜μ—ˆλ‹€.

OpenSlide Usages

less than 1 minute read

OpenSlide lib μ‚¬μš©λ°©λ²•

Setup DL Environment in Competition Container

less than 1 minute read

의료 λŒ€νšŒμ—μ„œ container둜 μ§„ν–‰ν•˜λ‹€λ³΄λ‹ˆ ν•„μš”ν•œ νŒŒμΌλ“€μ„ κ°€μƒν™˜κ²½μ²˜λŸΌ μ„€μΉ˜ν•œ ν›„ μ»¨ν…Œμ΄λ„ˆ μž¬μ‹œμž‘μ„ ν•˜λ©΄ λ‹€ λ‚ μ•„κ°€ 버린닀. 이에 따라 마운트 된 backup directory에 μ„€μΉ˜ 및 μ €μž₯ ν›„ μ‚¬μš©ν•΄μ•Ό ν•˜λŠ”λ°

Back to top ↑

Flask

Flask API with Swagger

less than 1 minute read

Flask run as debug mode $ export FLASK_ENV=development $ flask run if not set as dev mode, need to keep restart flask like: $ flask run

Back to top ↑

DevCourse2_Flask

Flask API with Swagger

less than 1 minute read

Flask run as debug mode $ export FLASK_ENV=development $ flask run if not set as dev mode, need to keep restart flask like: $ flask run

Back to top ↑

Probability

Back to top ↑

LinearAlgebra

ML basics - Linear Algebra

24 minute read

μ™œ μ„ ν˜•λŒ€μˆ˜λ₯Ό μ•Œμ•„μ•Ό ν•˜λŠ”κ°€?

Back to top ↑

ProbabilityDistributions

ML basics - Probability Distributions 1

6 minute read

밀도좔정(Density Estimation) \(N\)개의 관찰데이터(observations) \(\mathbf{x}_1,\ldots\mathbf{x}_N\)κ°€ μ£Όμ–΄μ‘Œμ„ λ•Œ λΆ„ν¬ν•¨μˆ˜ \(p(\mathbf{x})\)λ₯Ό μ°ΎλŠ” 것

Back to top ↑

DevCourse2_Assignment

Assignment - KNN

3 minute read

This assignment basically comes from cs231n. https://cs231n.github.io/assignments2021/assignment1/#q1-k-nearest-neighbor-classifier

Back to top ↑

Classification

Back to top ↑

DL_basics

Optimizers

1 minute read

Optimizers https://youtu.be/mdKjMPmcWjY

Back to top ↑

SparkMLlib

Big data - ML Pipelineκ³Ό Tuning μ†Œκ°œμ™€ μ‹€μŠ΅

4 minute read

Spark MLlib λͺ¨λΈ νŠœλ‹ μ†Œκ°œ Spark MLlib λͺ¨λΈ νŠœλ‹ 졜적의 ν•˜μ΄νΌ νŒŒλΌλ―Έν„° 선택 졜적의 λͺ¨λΈ ν˜Ήμ€ λͺ¨λΈμ˜ νŒŒλΌλ―Έν„°λ₯Ό μ°ΎλŠ” 것이 μ•„μ£Ό μ€‘μš” ν•˜λ‚˜μ”© ν…ŒμŠ€νŠΈν•΄λ³΄λŠ” 것 vs. λ‹€μˆ˜λ₯Ό λ™μ‹œ ν…ŒμŠ€νŠΈ ν•˜λŠ” 것 λͺ¨λΈ μ„ νƒμ˜ μ€‘μš”ν•œ ...

Back to top ↑

Bitbucket

Bitbucket Pipeline

less than 1 minute read

SSH key trouble shooting –build-arg ssh_prv_key=”$(cat ~/.ssh/id_rsa)” stargan

Bitbucket Pipeline runners

less than 1 minute read

create pipeline write bitbucket-pipelines.yml commit pipelines.yml file run docker at local (test runners in local env) ...

Back to top ↑

jupyter

vscode jupyter kernel

less than 1 minute read

python -m ipykernel install --user --name AI_hub_pre --display-name AI_hub_kernel

Back to top ↑

cicd

GitLab cicd starclass test

1 minute read

make test_run.py to pass with python uniitest 1-1. make cicd variables in settings.

Back to top ↑

Minimal-Mistakes

design scroll bar

less than 1 minute read

https://ansohxxn.github.io/blog/scrollbar/

create category UI

less than 1 minute read

https://ansohxxn.github.io/blog/category/

Back to top ↑

programming

Python GMP

less than 1 minute read

Python GMP

First-class citizen

less than 1 minute read

programming In programming language design, a first-class citizen (also type, object, entity, or value) in a given programming language is an entity which su...

Back to top ↑

NoonBody

NoonBody SemanticSegmentation

1 minute read

Task: Semantic Segmentation of Body labels are come from DensePose. Using these body labels, inference body parts.

Setup Env with conda

less than 1 minute read

λŒ€νšŒλ§ˆλ‹€ νŒ¨ν‚€μ§€ 버전 μ΄μŠˆκ°€ μžˆμ„ 수 있기 λ•Œλ¬Έμ— λŒ€νšŒ μ°Έμ—¬ν•  λ•Œλ§ˆλ‹€ λ‹€λ₯Έ conda ν™˜κ²½μ„ λ§Œλ“€μ–΄μ£Όκ³  μ‹œμž‘ν•˜λŠ”κ²Œ μ’‹λ‹€κ³  μƒκ°ν•œλ‹€.

Back to top ↑

Medical

Heart Datathon 2021

less than 1 minute read

Detection Metrics Dice Coefficient \[Dice = \frac{2 |A∩B|}{(|A|+|B|)} = \frac{2 TP}{(2 TP + FP + FN)}\]

Back to top ↑

conda

Git bash & conda in Windows

less than 1 minute read

Go to conda.sh location such as C:\Users\{your_username}\anaconda3\etc\profile.d Open git bash here by right click as Git ...

Back to top ↑

CodingTest

N Queen

1 minute read

Backtracking λŒ€ν‘œ 문제 Backtracking λŒ€ν‘œμ  문제라고 ν•œλ‹€.

Back to top ↑

remote-ssh

Remote-ssh access in VSCode

1 minute read

Remote-ssh VScode install VSCODE press F1 Find Remote-SSH: Connect current window to Host (or Remote-SSH: Connect to Host) λ§Œμ•½ 졜초 config λ§Œλ“€μ–΄μ•Ό 되면 4-1...

Back to top ↑

markdown

Jekyll unmodifided code block

less than 1 minute read

While writing django code snippet, I found that it hides double curly brackets( {{}} ). Since Jekyll uses the Liquid templating language to process templates...

Back to top ↑

Deploy

Back to top ↑

SCSS

Back to top ↑

Interview

Back to top ↑

ML

Back to top ↑

Pandas

Back to top ↑

DecisionTheory

Back to top ↑

LinearRegression

Back to top ↑

Ankle

Ankle Supports

less than 1 minute read

ν•˜μ΄μ—”λ“œ: λ°”μš°μ–΄νŒŒμΈνŠΈ: 말레였트레인(ν‰μƒμ‹œ), 말레였슀포츠(μš΄λ™μ‹œ) κ²Œλ‰΄νŠΈλ ˆμΈ

Back to top ↑

scss

Back to top ↑

KNN

Assignment - KNN

3 minute read

This assignment basically comes from cs231n. https://cs231n.github.io/assignments2021/assignment1/#q1-k-nearest-neighbor-classifier

Back to top ↑

Regression

Back to top ↑

Markdown

Markdown Named Anchor

less than 1 minute read

```markdown go to header 1 go to header 2 go to header 3

Back to top ↑

Jupyter

Back to top ↑

cs231n

cs231n Assignment

10 minute read

1. 예제 λ”°λΌν•˜κ³ , 개인 Github μ •λ¦¬ν•˜κΈ° CNN

Back to top ↑

Assignment

cs231n Assignment

10 minute read

1. 예제 λ”°λΌν•˜κ³ , 개인 Github μ •λ¦¬ν•˜κΈ° CNN

Back to top ↑

RNN

NN basics - RNN

5 minute read

Sequential Data ν‘œν˜„ λ²‘ν„°μ˜ 벑터 \(\boldsymbol{x} = (\boldsymbol{x}^{(1)}, \boldsymbol{x}^{(2)},\ldots ,\boldsymbol{x}^{(T)})^{T}\). 사전 ...

Back to top ↑

Gdrive

Download Gdrive file via wget or gdown

less than 1 minute read

Wget 링크 생성(링크가 μžˆλŠ” λͺ¨λ“  μ‚¬μš©μžμ—κ²Œ 곡개) wget μΏ ν‚€ μ„€μ •μœΌλ‘œ λ‹€μš΄λ‘œλ“œ wget --load-cookies ~/cookies.txt "https://docs.google.com/uc?export=download&confirm=$(wget -...

Back to top ↑

tmux

Back to top ↑

OS

Rename files with Python using os.rename

less than 1 minute read

```py import os def changeName(path, cName): i = 1 for filename in os.listdir(path): print(path+filename, β€˜=>’, path+str(cName)+str(i)+’....

Back to top ↑

Regex

Python Regex

less than 1 minute read

Regex 자주 μ‚¬μš©ν•˜λŠ” 문자 클래슀

Back to top ↑

CocoAnnotator

Back to top ↑

Backend

Back to top ↑

RabbitMQ

Back to top ↑

Celery

Back to top ↑

Redis

Docker Redis

less than 1 minute read

Redis with Docker get redis image from docker hub Docker image pull redis redis-cli와 같이 ꡬ동해야 됨. 두 μ»¨ν…Œμ΄λ„ˆλ₯Ό μ—°κ²°...

Back to top ↑

submodule

Git checkout from commit

less than 1 minute read

νŠΉμ • commit으둜 λΆ€ν„° μƒˆλ‘œμš΄ 브랜치λ₯Ό checkout ν•˜λŠ” cmd

Back to top ↑

checkout

Git checkout from commit

less than 1 minute read

νŠΉμ • commit으둜 λΆ€ν„° μƒˆλ‘œμš΄ 브랜치λ₯Ό checkout ν•˜λŠ” cmd

Back to top ↑

Coroutines

Back to top ↑

Tasks

Back to top ↑

urllib

urllib parse unquote

less than 1 minute read

urllib.parse.unquote(string, encoding=’utf-8’, errors=’replace’) Replace %xx escapes with their single-character equivalent. The optional encoding and err...

Back to top ↑

decode

urllib parse unquote

less than 1 minute read

urllib.parse.unquote(string, encoding=’utf-8’, errors=’replace’) Replace %xx escapes with their single-character equivalent. The optional encoding and err...

Back to top ↑

Pipeline

Bitbucket Pipeline

less than 1 minute read

SSH key trouble shooting –build-arg ssh_prv_key=”$(cat ~/.ssh/id_rsa)” stargan

Back to top ↑

Pytorch

Pytorch IterableDataset

less than 1 minute read

https://inmoonlight.github.io/2021/02/21/PyTorch-IterableDataset/

Back to top ↑

Gitlab

Back to top ↑

Network

Back to top ↑

GAN

Back to top ↑

Coursera

Back to top ↑

GAN_Specialization

Back to top ↑

Fork

Back to top ↑

system

Back to top ↑

Conda

Conda Vs. Pip

less than 1 minute read

conda 둜 ν™˜κ²½ λ§Œλ“€μ–΄μ„œ torch install을 conda둜 ν–ˆλ”λ‹ˆ cuda Falseκ°€ λ˜λ²„λ €μ„œ 버전 λ¬Έμ œμΈμ€„ μ•Œμ•˜λ”λ‹ˆ pip install 둜 ν•˜λ‹ˆ 문제 없이 cuda μ‚¬μš© κ°€λŠ₯ν•˜κ²Œ λ˜μ—ˆλ‹€.

Back to top ↑

yaml

Back to top ↑

Cell

Kaggle - Sartorius Cell Instance Segmentation

less than 1 minute read

Overall Entered around 2 weeks before the competition ends. Get used to implementing Detectron2 with custom data First Instance Segmentation Task

Back to top ↑

git

Git bash & conda in Windows

less than 1 minute read

Go to conda.sh location such as C:\Users\{your_username}\anaconda3\etc\profile.d Open git bash here by right click as Git ...

Back to top ↑

bash

Git bash & conda in Windows

less than 1 minute read

Go to conda.sh location such as C:\Users\{your_username}\anaconda3\etc\profile.d Open git bash here by right click as Git ...

Back to top ↑

gitlab

Windows gitlab ssh

less than 1 minute read

Execute git bash by windows key command ssh-keygen Enter file in which to save the key (μƒμ„±μœ„μΉ˜ μ •ν•˜λŠ” 파트) 기본경둜둜 ν•  μ˜ˆμ •μ΄λ―€λ‘œ ENTER ...

Back to top ↑

ssh

Windows gitlab ssh

less than 1 minute read

Execute git bash by windows key command ssh-keygen Enter file in which to save the key (μƒμ„±μœ„μΉ˜ μ •ν•˜λŠ” 파트) 기본경둜둜 ν•  μ˜ˆμ •μ΄λ―€λ‘œ ENTER ...

Back to top ↑

windows

Windows gitlab ssh

less than 1 minute read

Execute git bash by windows key command ssh-keygen Enter file in which to save the key (μƒμ„±μœ„μΉ˜ μ •ν•˜λŠ” 파트) 기본경둜둜 ν•  μ˜ˆμ •μ΄λ―€λ‘œ ENTER ...

Back to top ↑

MLFlow

ML Flow Tutorial

less than 1 minute read

MLOps 반볡적인 μž‘μ—…μ„ ν•΄μ†Œν•  수 μžˆλŠ” 방법이면 λ‹€ MLOps 의 μ˜μ—­μœΌλ‘œ κ°„μ£Όν•œλ‹€.

Back to top ↑

random_thoughts

μ‹œμž‘

10 minute read

μ‹œμž‘ν•˜κ²Œ 된 이유 λŒ€μ²΄ 뭘 μ‹œμž‘ν•˜λŠλƒμ— λŒ€ν•œ 닡은. μƒκ°μ˜ 정리 및 μ‚΄λ©΄μ„œ λ“œλŠ” 생각듀 및 μΌν•˜λ©΄μ„œ 느꼈던 것에 λŒ€ν•œ νšŒκ³ μ΄λ‹€. μ™œ μ‹œμž‘ν•˜κ²Œ λ˜μ—ˆλŠ”κ°€. μΌν•œμ§€ 1년이 μ§€λ‚˜κ³  λ‚΄ λ”΄μ—λŠ” μΌν•˜λŠ” ν™˜κ²½λ„ 크게 λ°”λ€Œμ§„ μ•Šμ•˜μ§€λ§Œ, μ’…μ „κΉŒμ§€λŠ” 혼자 μΌν•˜κ³  μ±…μž„μ§€λŠ” ν™˜κ²½μ—μ„œ μ•„λ§ˆλ„ νŒ€μ›λ“€μ„ κΎΈ...

Back to top ↑