Skip to content

Instantly share code, notes, and snippets.

View graykode's full-sized avatar
๐Ÿ’ญ
not degenning rn, focused instead

Tae Hwan Jung graykode

๐Ÿ’ญ
not degenning rn, focused instead
View GitHub Profile
@junha1
junha1 / rust_study.md
Last active October 27, 2025 06:16
Rust ๊ณต๋ถ€ํ•˜๋Š”๋ฒ•

Rust ๊ณต๋ถ€ํ•˜๋Š”๋ฒ•

2022๋…„ 9์›” 5์ผ ์–‘์ค€ํ•˜

์™œ Rust๋ฅผ ๋ฐฐ์›Œ์•ผ ํ•˜๋Š”๊ฐ€?

  • 2021๋…„ 4๋ถ„๊ธฐ ๊ธฐ์ค€ Github ์ ์œ ์œจ 15์œ„ ์–ธ์–ด
  • C++ ์•”์‚ด์ž ํฌ์ง€์…˜ (๊ฐœ์ธ์ ์œผ๋กœ 10๋…„์•ˆ์— C++ ๋„˜์–ด์„ค ๋“ฏ)
  • ๋ธ”๋ก์ฒด์ธ, ์ž„๋ฒ ๋””๋“œ, ์‹œ์Šคํ…œํ”„๋กœ๊ทธ๋ž˜๋ฐ, ์„œ๋ฒ„, ๋ถ„์‚ฐ์ฒ˜๋ฆฌ, WASM ๋“ฑ์—์„œ ํ™œ๋ฐœํžˆ ์‚ฌ์šฉ๋˜๋Š” ๊ณ ์„ฑ๋Šฅ ์–ธ์–ด
  • ๋ชจ๋˜ํ•˜๊ณ  ๊น”๋”ํ•œ ์–ธ์–ด ๋””์ž์ธ, ํ›Œ๋ฅญํ•œ ๊ฐœ๋ฐœํˆด๊ณผ ํŒจํ‚ค์ง€ ๋งค๋‹ˆ์ €
  • ์ˆ˜์ค€๋†’์€ ์‚ฌ์šฉ์ž๋“ค๊ณผ ์ปค๋ฎค๋‹ˆํ‹ฐ, ๋…๋ณด์ ์ธ UX๋ฅผ ๊ธฐ๋ฐ˜์œผ๋กœ ํ•œ ์ฒ ์˜น์„ฑ ๊ฐ™์€ ํŒฌ๋ค๋“ค
@luncliff
luncliff / cmake-tutorial.md
Last active November 22, 2025 08:43
CMake ํ• ๋•Œ ์ชผ์˜ค์˜ค๊ธˆ ๋„์›€์ด ๋˜๋Š” ๋ฌธ์„œ

CMake๋ฅผ ์™œ ์“ฐ๋Š”๊ฑฐ์ฃ ?
์ข‹์€ ํˆด์€ Visual Studio ๋ฟ์ž…๋‹ˆ๋‹ค. ๊ทธ ์ด์™ธ์—๋Š” ์ „๋ถ€ ์‚ฌ๋„(้‚ช้“)์ž…๋‹ˆ๋‹ค ์‚ฌ๋„! - ์ž‘์„ฑ์ž

์ฃผ์˜

  • ์ด ๋ฌธ์„œ๋Š” CMake๋ฅผ ์ฃผ๊ด€์ ์œผ๋กœ ์„œ์ˆ ํ•ฉ๋‹ˆ๋‹ค
  • ์ด ๋ฌธ์„œ๋ฅผ ํ†ตํ•ด CMake๋ฅผ ์‹œ์ž‘ํ•˜๊ธฐ์—” ์ ํ•ฉํ•˜์ง€ ์•Š์Šต๋‹ˆ๋‹ค
    https://cgold.readthedocs.io/en/latest/ 3.1 ์ฑ•ํ„ฐ๊นŒ์ง€ ๋”ฐ๋ผํ•ด๋ณธ ์ดํ›„ ๊ธฐ๋ณธ์‚ฌํ•ญ๋“ค์„ ์†์„ฑ์œผ๋กœ ์ตํžˆ๋Š” ๊ฒƒ์„ ๋•๊ธฐ์œ„ํ•œ ๋ณด์กฐ์ž๋ฃŒ๋กœ์จ ์ž‘์„ฑ๋˜์—ˆ์Šต๋‹ˆ๋‹ค
@thomwolf
thomwolf / datadistributedparallel.py
Last active December 13, 2022 19:15
Using DistributedDataParallel
from torch.utils.data.distributed import DistributedSampler
from torch.utils.data import DataLoader
# Each process runs on 1 GPU device specified by the local_rank argument.
parser = argparse.ArgumentParser()
parser.add_argument("--local_rank", type=int)
args = parser.parse_args()
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.distributed.init_process_group(backend='nccl')
@thomwolf
thomwolf / parallel.py
Last active August 8, 2023 15:50
Data Parallelism in PyTorch for modules and losses
##+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
## Created by: Hang Zhang, Rutgers University, Email: [email protected]
## Modified by Thomas Wolf, HuggingFace Inc., Email: [email protected]
## Copyright (c) 2017-2018
##
## This source code is licensed under the MIT-style license found in the
## LICENSE file in the root directory of this source tree
##+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
"""Encoding Data Parallel"""
@atyachin
atyachin / Android-Emulator-on-AWS-EC2.txt
Last active May 20, 2025 10:07
Installing and running Android Emulator on Amazon AWS EC2 (Ubuntu 16.04 / m5.xlarge)
Update (2022): https://gist.github.com/atyachin/2f7c6054c4cd6945397165a23623987d
Steps for installing the Android Emulator from EC2 console:
-----------------------------------------------------------
sudo apt install default-jdk
wget https://dl.google.com/android/repository/sdk-tools-linux-4333796.zip
unzip sdk-tools-linux-4333796.zip -d android-sdk
sudo mv android-sdk /opt/
export ANDROID_SDK_ROOT=/opt/android-sdk
@W4ngatang
W4ngatang / download_glue_data.py
Last active October 21, 2025 02:22
Script for downloading data of the GLUE benchmark (gluebenchmark.com)
''' Script for downloading all GLUE data.
Note: for legal reasons, we are unable to host MRPC.
You can either use the version hosted by the SentEval team, which is already tokenized,
or you can download the original data from (https://download.microsoft.com/download/D/4/6/D46FF87A-F6B9-4252-AA8B-3604ED519838/MSRParaphraseCorpus.msi) and extract the data from it manually.
For Windows users, you can run the .msi file. For Mac and Linux users, consider an external library such as 'cabextract' (see below for an example).
You should then rename and place specific files in a folder (see below for an example).
mkdir MRPC
cabextract MSRParaphraseCorpus.msi -d MRPC
@seopbo
seopbo / R-CNN.md
Last active August 25, 2020 08:17
R-CNN

R-CNN : Rich feature hierarchies for accurate object detection and semantic segmentation

๋ณธ ๋…ผ๋ฌธ (R-CNN)์€ Object detction์— Convolutional Neural Network๋ฅผ feature extractor๋กœ ์‚ฌ์šฉํ•œ ๋…ผ๋ฌธ ์œผ๋กœ ์ดํ›„ Fast R-CNN, Faster R-CNN ๋“ฑ ์—ฌ๋Ÿฌ๊ฐ€์ง€ ๋…ผ๋ฌธ์˜ ๊ธฐ๋ฐ˜์ด ๋˜๋Š” ๋…ผ๋ฌธ์ž…๋‹ˆ๋‹ค. ์ €์ˆ ๋˜์–ด ๊ณต๊ฐœ๋œ ์ง€ ์˜ค๋ž˜๋œ ๋…ผ๋ฌธ์ด ๋งŒํผ ์—ฌ๋Ÿฌ report๊ฐ€ ์กด์žฌํ•˜๋ฉฐ, ๋ณธ ํฌ์ŠคํŠธ๋Š” Rich feature hierarchies for accurate object detection and semantic segmentation Tech report (v5)์— ๊ธฐ์ดˆํ•˜์—ฌ ์ž‘์„ฑํ•˜์˜€์œผ๋ฉฐ, ์ค‘์š”ํ•œ idea๋งŒ ๋‹ค๋ฃจ๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ์ƒ์„ธํ•œ ๋‚ด์šฉ์€ ๋…ผ๋ฌธ์„ ๋ณด์‹œ๋ฉด ์ข‹์„ ๋“ฏ ํ•ฉ๋‹ˆ๋‹ค.


Abstract

๋ณธ ๋…ผ๋ฌธ์—์„œ๋Š” mAP (mean Avereage Precision) ๋ฅผ ๊ธฐ์ค€์œผ๋กœ VOC 2012์˜ best result์™€ ๋น„๊ตํ•˜์—ฌ, 30% ์ด์ƒ์˜ ์„ฑ๋Šฅ ํ–ฅ์ƒ์„ (mAP : 53.3%) ์ด๋ฃจ์—ˆ๋‹ค๊ณ  ๋งํ•˜๋ฉฐ, ๊ทธ ๊ธฐ๋ฐ˜์ด ๋˜๋Š” ์•„์ด๋””์–ด๋Š” ์•„๋ž˜์˜ ๋‘ ๊ฐ€์ง€์ž…๋‹ˆ๋‹ค.

  • One can apply high-capacity convolutional neural networks (CNNs) to bottom-up region proposal in order to localize and segment objects.
  • When labeled traning data is scarce, supervised pre-training for an auxiliary task. followed by domain-specific fine-tuning, yields a significant performance boost.
@haje01
haje01 / ๋Œ€ํ™”ํ˜• ์ฑ—๋ด‡ ์„ค๊ณ„์˜ ๊ณผ์ œ.md
Last active June 15, 2022 09:33
๋Œ€ํ™”ํ˜• ์ฑ—๋ด‡ ์„ค๊ณ„์˜ ๊ณผ์ œ

์ตœ๊ทผ ์ธ๊ณต์ง€๋Šฅ์„ ํ™œ์šฉํ•œ ์ฑ—๋ด‡์— ๋Œ€ํ•œ ๊ด€์‹ฌ์ด ๋†’์•„์ง€๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ์ฑ—๋ด‡ ์„ค๊ณ„์— ๊ด€ํ•œ ์ข‹์€ ๊ธ€์ด ์žˆ์–ด ๋ฒˆ์—ญ์„ ํ•ด๋ณด์•˜์Šต๋‹ˆ๋‹ค. ์ด ๊ธ€์€ IBM DeveloperWorks์— ๊ธฐ์žฌ๋œ Michael Yuan์˜ ๊ธ€์„ ๋ฒˆ์—ญํ•œ ๊ฒƒ์œผ๋กœ ์˜์—ญ์ด ์žˆ์Šต๋‹ˆ๋‹ค. - ๊น€์ •์ฃผ([email protected])


๋Œ€ํ™”ํ˜• ์ฑ—๋ด‡ ์„ค๊ณ„์˜ ๊ณผ์ œ

์‚ฌ์šฉ์ž๋Š” ์ฑ—๋ด‡์ด ๋งค์šฐ ๊ฐ„๋‹จํ•˜๊ณ  ์ตœ์†Œํ•œ์˜ ์š”๊ตฌ๋งŒ ํ•˜๊ธฐ์— ์ข‹์•„ํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๊ฒƒ์€ ๋Œ€ํ™”์‹ ๋ฌธ์ž ๋ฉ”์‹œ์ง€์ฒ˜๋Ÿผ ๊ฐ„๋‹จํ•ด์งˆ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๋˜ํ•œ, ์‚ฌ์šฉ์ž๋Š” ์ž์‹ ์ด ์„ ํ˜ธํ•˜๋Š” ๋ฉ”์‹œ์ง€ ์•ฑ์— ๊ณ„์† ๋จธ๋ฌผ๊ธฐ๋ฅผ ์„ ํ˜ธํ•ฉ๋‹ˆ๋‹ค. ์•ฑ, ์›น URL, ๋ฉ”๋‰ด, ๋ฒ„ํŠผ, ๊ด‘๊ณ , ํฌ๋กฌ ๋ฐ ๊ธฐํƒ€ ์š”์†Œ๋ฅผ ํƒ์ƒ‰ํ•˜์ง€ ์•Š๊ณ  ๋ฐ”๋กœ ๋ชฉํ‘œ๋ฅผ ๋‹ฌ์„ฑํ•˜๊ณ ์ž ํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์ด ๋‹จ์ˆœ์„ฑ์€ ํฐ ์„ค๊ณ„ ๊ณผ์ œ๋„ ์ œ์‹œํ•ฉ๋‹ˆ๋‹ค. ์ฑ—๋ด‡์€ ์‚ฌ์šฉ์ž์˜ ๋ง์„ ์ •ํ™•ํ•˜๊ฒŒ ์ดํ•ดํ•˜๊ณ  ์ ์ ˆํžˆ ํ–‰๋™ํ•ด์•ผ ํ•ฉ๋‹ˆ๋‹ค. ์ด๊ฒƒ์€ ์˜ค๋Š˜๋‚  ์ตœ๊ณ ์˜ ์ž์—ฐ์–ด AI (์ธ๊ณต ์ง€๋Šฅ)์—๊ฒŒ๋„ ๋งค์šฐ ์–ด๋ ค์šด ๊ณผ์ œ์ž…๋‹ˆ๋‹ค.

ํ˜„์žฌ ์ƒํƒœ์˜ AI์—์„œ๋Š”, ๋Œ€ํ™”์‹ ๋ฌธ์ž ๋ฉ”์‹œ์ง€ ๋˜๋Š” ๋Œ€ํ™”์‹(Conversational) UI, ์ฆ‰ CUI๋Š” (์•ˆํƒ€๊น๊ฒŒ๋„) ๊ฑฐ์˜ ํ•ญ์ƒ ์ž˜ ์„ค๊ณ„๋œ ๊ทธ๋ž˜ํ”ฝ UI(GUI)๋ณด๋‹ค ์—ด๋“ฑํ•ฉ๋‹ˆ๋‹ค. GUI์™€ ๋น„๊ตํ•˜์—ฌ CUI๋Š” ์ดˆ๊ธฐ ๋‹จ๊ณ„์— ์žˆ์Šต๋‹ˆ๋‹ค. ์ปค๋ฎค๋‹ˆํ‹ฐ๋กœ์„œ ์šฐ๋ฆฌ๋Š” ์—ฌ์ „ํžˆ CUI์˜ ๋””์ž์ธ ํŒจํ„ด๊ณผ ์šฐ์ˆ˜ ์‚ฌ๋ก€๋ฅผ ๋ชจ์ƒ‰ํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ์ด ํŠœํ† ๋ฆฌ์–ผ์—์„œ๋Š” ์ฑ—๋ด‡์ด ์™œ ์‹คํŒจํ•˜๊ณ  ์„ฑ๊ณตํ•  ์ˆ˜ ์žˆ๋Š”์ง€ ์„ค๋ช…ํ•ฉ๋‹ˆ๋‹ค.

@pulkitsinghal
pulkitsinghal / shortcuts.md
Last active May 14, 2023 12:07
mac: WebStorm to Visual Studio (VSCode) key mappings

key mappings, keybindings, keyboard shortcuts

https://github.com/Microsoft/vscode-tips-and-tricks

Command: workbench.action.quickOpen

WebStorm: cmd+o VSCode: cmd+e

Command: collapse/fold all code blocks

WebStorm: cmd shift - VSCode: fold level 1 cmd+K cmd+1

@bfoz
bfoz / fixup_committer_date.sh
Created September 7, 2010 19:24
Force GIT_COMMITTER_DATE = GIT_AUTHOR_DATE
git filter-branch --env-filter 'export GIT_COMMITTER_DATE="$GIT_AUTHOR_DATE"'