2016 ilsvrc
2019.08.31 21:20
Trimps-Soushen used the pretrained models from Inception-v3, Inception-v4, Inception-ResNet-v2, Pre-Activation ResNet-200, and Wide ResNet (WRN-68–2) for classification, and found out Top-10 difficult categories as above. Diverse results are obtained, which means there is no models being dominant for all categories. Each of the models are strong at classifying some categories, but also weak at classifying some categories. The diversity of models can be used for improving the accuracy. During training, Trimps-Soushen just performed multi-scale augmentation & large mini batch size. During testing, multi-scale + flip are used with dense fusion.
댓글 0
번호 | 제목 | 글쓴이 | 날짜 | 조회 수 |
---|---|---|---|---|
1819 | mAP 계산법 | WHRIA | 2019.02.24 | 45575 |
1818 | epitope spreading | han | 2006.05.09 | 41021 |
1817 | encfs | WHRIA | 2020.05.17 | 35582 |
1816 | Heinrich Law (1:29:300 Law) | WHRIA | 2007.08.12 | 24996 |
1815 | 일본 주소 [2] | WHRIA | 2008.06.28 | 24028 |
1814 | 사주팔자 프로그램 | 한승석 | 2003.02.16 | 20330 |
1813 | simple adblock | WHRIA | 2012.04.13 | 18111 |
1812 | 도란사민 | WHRIA | 2011.04.19 | 18093 |
1811 | 탤런트 이윤지씨와 함께 | WHRIA | 2010.02.04 | 17043 |
1810 | penicillin | han | 2003.12.10 | 15712 |
1809 | SSH tunnel | WHRIA | 2007.10.01 | 15085 |
1808 | geexbox [1] | han | 2006.12.01 | 14610 |
1807 | 아이피부과 개원 | WHRIA | 2010.01.18 | 14438 |
1806 | 세무회계 | WHRIA | 2010.01.31 | 14401 |
1805 | 증명사진 | WHRIA | 2010.03.14 | 14330 |