lxf

> 如果是的话，那下图中三个圈起来的模块的实现分别是对应哪些代码？ >  > 其中的ffm我知道了是[feat_fuse = self.ffm(feat_res8, feat_cp8)](https://github.com/MichaelFan01/STDC-Seg/blob/master/models/model_stages.py#L282)； > 那请问其他两个呢？ 1.stage5下面的灰色框对应./models/model_stages.py ```python 139 avg = F.avg_pool2d(feat32, feat32.size()[2:]) 140 141 avg = self.conv_avg(avg) 142 avg_up = F.interpolate(avg, (H32, W32), mode='nearest')...

> 另外，请问框起来的变量feat_out_sp2，feat_out_sp4，feat_out_sp8，feat_out_sp16， feat_out， feat_out16， feat_out32都分别是做什么的（大哭，看不懂啊）； > 如果我只想return输出得到最终的output图，请问是哪个变量？是feat_out， feat_out16或者feat_out32三个其中一个都可以吗？还是[feat_out这个变量而已](https://github.com/MichaelFan01/STDC-Seg/blob/master/evaluation.py#L49)，其他变量只是为了多角度约束损失函数的？ >  feat_out_sp2, sp4, sp8, sp16 这四个是backbone从stage1,2,3,4出来的特征图 至于feat_out16, 32是context path里经过ARM结构的两个特征图,用来计算辅助loss的(细节可以去看BISeNet的代码实现),这两个特征图不能直接用 

I meet the same error.I suppose it may has some connection with G2O versions. Have you solved this problem already? PS:my ROS distribution is indigo+unbuntu 14.04

I changed that file as above, and then try "rosmake -s g2o", but the CMakeLists.txt was overwrite by initial version. How can I do