Issues
- 2
低频分量
#30 opened by xiaokedouaa - 0
- 0
关于特定频率分量的选择
#47 opened by liu-anan - 0
- 0
一维的GAP是否可以被视为一维DCT的特例呢
#44 opened by da-orange-zi - 3
如何进行FcaNet-TS的实验
#42 opened by Soso-developer - 2
channel groups?
#26 opened by dreamhua82 - 4
频率分量的确定
#40 opened by yueshuheng - 4
关于7X7频域的问题
#43 opened by Soso-developer - 2
如何理解和解释,固定的DCT比可学习的方式更好?
#41 opened by Frank-Star-fn - 3
get_dct_weights() cannot be found
#35 opened by Yung-zi - 2
- 0
这个不是和SeNet差不多吗?为啥不可以做成即插即用的注意力模块
#39 opened by zhulewen1998 - 2
是否能提供一下对比实验中ResNet50的结果权重呢?
#36 opened by WYHZQ - 2
关于模型精度
#38 opened by 123456789asdfjkl - 5
dct_h and dct_w
#37 opened by myasser63 - 1
model.init
#33 opened by yuan243212790 - 1
想请问一下代码中bot是怎么选取的代表什么意思
#34 opened by Liutingjin - 1
SE-NET基础上加FCA
#28 opened by dfsjdfdv - 1
请问YOLOX可以使用吗?看到是ResNet?
#32 opened by RoeweYaoo - 1
selecting frequency components
#31 opened by InukKang - 0
- 8
- 2
通道切分
#22 opened - 2
- 2
请问一下这个模块在分割的时候如何设置dct_h和dct_w?
#24 opened by nnoyuwan - 3
- 2
Supplementary file
#20 opened by weigq - 4
- 4
Regarding the setting of c2wh = dict([(64,56), (128,28), (256,14) ,(512,7)]) in the fcanet.py file.
#17 opened by Max-Well-Wang - 2
关于通道分组
#18 opened by Asthestarsfalll - 6
有关self.dct_h和self.dct_w的设置?
#10 opened by XFR1998 - 6
关于math.sqrt(POS)
#15 opened by Asthestarsfalll - 0
is there anyone can operate this code on his own machine? does this code work?
#16 opened by yutinyang - 2
get_freq_indices?
#8 opened by qiuqiu-lwf - 4
关于split成n份的目的
#14 opened by HHCorp - 2
跑您的模型的时候遇到的一些问题
#13 opened by LihuiNb - 2
- 14
About the performance on cifar10 or cifar100.
#9 opened by NNNNAI - 1
代码好像和论文里说的不太一样也?
#11 opened by XFR1998 - 5
2d dct FLOPs computing method
#7 opened by TianhaoFu - 2
- 3
get_freq_indices
#4 opened by Slayerxxx - 1
- 4
Official code for FCANet?
#1 opened by Senwang98