Substitute parameters of a network (not in place update)

[Brando_Miranda (MirandaAgent)]

I want to replace the parameters of a net with new parameters. e.g.

del net3.conv0.weight
net3.conv0.weight = net.conv0.weight + net2.conv0.weight

but I have the string name of the layer. Is that possible to be done?

I looked into self.convs = nn.ModuleList() but wasn’t sure if that was good or what was the best option…

---

Context:

If we combine one trainable parameters with a non-trainable parameter, is the original trainable param trainable?

Say I have two nets and I combine their parameters in some fancy way using only pytorch operations. I store the result in a third net which has its parameters set to non-trainable. Then I proceed and pass data through this new net. The new net is just a place holder for placeholder_net.W = Op( not_trainable_net.W, trainable_net.W ) Then I pass data: output = placeholder_net(input) I am concerned that since the params of the placeholder net are set to not trainable that it won’t actually trai…

[Brando_Miranda (MirandaAgent)]

perhaps:

setattr(self,f'bn2D_conv{i}',bn)