New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
用GhostModule替换Conv2d,loss降的很慢? #81
Comments
试试 另外,EfficientNet里PWConv换成GhostModule,DWConv就不要再换成GhostModule了 |
是的,我只把PWConv替换成GhostModule。按您给的参数还是不行诶,和Conv2d差别很大 |
网络首尾层的Conv不要换,影响较大。另外,log能发来看看吗? |
首尾的1x1卷积比较耗时吧,我想用GhostModule替换掉。我想要牺牲精度提高EfficientNet的效率,如果不替换首尾,那该替换哪呢? |
这是替换之前的:
替换之后的:
替换前后我都跑了500个epoch,替换之前效果挺好的,替换之后的loss几乎就降不下来了。 |
替换后的backbone也得先imagenet预训练,然后再在detection上finetune吧 |
哦哦,好 那么这个将原来的1x1卷积替换为GhostModule应该是可行的吧 |
我觉得可行,不会掉点那么多的 |
我直接将efficientnet里面的MBConvBlock中的Conv2d替换为GhostModule:
Conv2d(in_channels=inp, out_channels=oup, kernel_size=1, bias=False)
替换为
GhostModule(inp, oup)
,其他参数不变,为什么损失比以前收敛的更慢了,一直降不下来?请问需要修改其他什么参数吗?
The text was updated successfully, but these errors were encountered: