Q&A

What I learned

Docker파기

Docker 안에서 Jupyter 사용하기

Docker

서버 간 파일 이동

가상 환경 venv

linux 긴명령어 → alias

Git 저장소 생성

Github 블로그 생성

github 코웍

gitignore

서버 연결 안될때

slurm

클로드 데스크탑으로 테스트

도커 빌드캐시삭제하기

WHAT I STUDIED

Ensemble

BottleNeck

Training vs Inference(추론)

NMS (비최댓값 억제)

offset

Receptive field, dilated conv

upsampling downsampling

Semantic Segmentation

DNN, FC layer

pooling indices

One-Hot Encoding

Dropout, BarchNorm

SIFT, HOG features

Train / Validation / Test set 차이

활성화 함수 ( activation function)

Optimizer

permute()

Incremental Learning

Meta learning

inductive learning vs transductive learning

Averaging Weights Leads to Wider Optima and Better Generalization

Linear probing

tensorboard url에서 열기

sliding attention window

Softmax(+ temperature scaling)

Instance-wise Contrastive loss & Temporal Contrastive loss

cross attention

Probability Theory

Machine Learning

Auto encoder

Gan_augmentation

Local minima & Optimal minima

Using too large learning rate

NAS(neural architecture search)

grid search

Model Soups

Pruning (가지치기)

Quantization (실수형 변수를 정수형으로)

Huffman encoding(파일 압축)

offset

pseudo labeled

RandAugment

Stochastic Depth

Transfer Learning

outlier detection

Pytorch initializer 정리

Domain Generalization/ Adaptation

Model soups

flops, flop regime

CUDA CUDNN version

Local Attention

lightweight model, Model compression

generative pretraining (linear probing)