본문 바로가기

AI & 머신러닝 coding skill

[SEQUENCES, TIME SERIES AND PREDICTION] Mini batch gradient descent

by 호빵님 2020. 11. 11.

DNN에서 time serise data를 traning할 때, batch data가 많으면, traing에 오랜 시간이 걸린다.

그래서, batch중에 선정한 mini batch를 이용해서 빠르게 gradient descent를 진행하는 것이 traing에 더 빠른 효과가 있다.

아래는 관련 유튜브 강의이니 참고하기 바란다.

www.youtube.com/watch?v=4qJaSmvhxi8

저작자표시 비영리 변경금지

'AI & 머신러닝 coding skill' 카테고리의 다른 글

[INTRODUCTION TO TENSORFLOW FOR ARTIFICIAL INTELLIGENCE, MACHINE LEARNING, AND DEEP LEARNING] Implement a Deep Neural Network to recognize handwritten digits (0)	2020.11.16
[INTRODUCTION TO TENSORFLOW FOR ARTIFICIAL INTELLIGENCE, MACHINE LEARNING, AND DEEP LEARNING] Housing Prices (0)	2020.11.16
[SEQUENCES, TIME SERIES AND PREDICTION] Real data Sunspots 예측 (0)	2020.11.12
[SEQUENCES, TIME SERIES AND PREDICTION] multi layer LSTM for time series (0)	2020.11.11
[SEQUENCES, TIME SERIES AND PREDICTION] RNN, LSTM for time series (0)	2020.11.11
[SEQUENCES, TIME SERIES AND PREDICTION] Deep neural network for time series 예측 (0)	2020.11.11
[SEQUENCES, TIME SERIES AND PREDICTION] Preparing features and labels (0)	2020.11.11
[SEQUENCES, TIME SERIES AND PREDICTION] Sequences and Prediction (0)	2020.11.11

티스토리툴바