M net output

[judgeeeeee]

raw_alpha = self.conv_0(x1d)

and

alpha_r = self.m_net(m_net_input)

# fusion module

# paper : alpha_p = fs + us * alpha_r

alpha_p = fg + unsure * alpha_r

alpha_p may not in [0,1]

is that trouble?

[tsing90]

you are right, my mistake, a sigmoid function should be added to the end of m-net

[LoveU3tHousand2]

And how to do that?

[LoveU3tHousand2]

you are right, my mistake, a sigmoid function should be added to the end of m-net

Hi, 关于这个问题， 您是不是未在公布的代码中修改啊， 因为我这目前mnet的pre_train效果特别奇怪， 我找了两天的原因， 刚刚才看到条issue， 以下是我的'奇怪的结果'

[LoveU3tHousand2]

还希望能回复