Combining LogisticHazard models about pycox HOT 1 OPEN

Hi, I think I should be able to help you if you could provide me with some more details, as I don't fully understand what you want.
Do you want three networks each with the same output size $m$ and then you want to average these three? In that case the following should work for you. Let's say you have defined three nets: net1, net2, net3 and three covariate vectors x1, x2, x3. You can average the output of the nets by

class NetTriple(nn.Module):
    def __init__(self, net1, net2, net3):
        super().__init__()
        self.net1 = net1
        self.net2 = net2
        self.net3 = net3
    
    def forward(self, x1, x2, x3):
        out1 = self.net1(x1)
        out2 = self.net2(x2)
        out3 = self.net3(x3)
        return (out1 + out2 + out3) / 3

And fit it with

net = NetTriple(net1, net2, net3)
model = LogisticHazard(net)
log = model.fit([x1, x2, x3], y_, 256, 10)

It might be cleaner to make the nets inside the NetTriple, but that is up to you.

What this what you were looking for, or have I missunderstood?

from pycox.