Hello, I am doing this:
weights_init(F1)
if args.net == "resnet32":
G_final = nn.Sequential(G,F1.fc1)
F1.fc1 = nn.Identity()
G = deepcopy(G_final)
What I want to do is append the initial layer of F1
to G
. I am initialising the weights of F1
in a particular manner using the weights_init
function.
So, my question is when I do these steps, do the weights which have been specially initialised in F1
also get transferred to the net G_final
network?
Thanks,
Megh