hello, guys. I only have access to a well-trained model and have forgot the exact settings in the Network strucutre.
I used to train a model with a layer like this:
def forward(self, x): x_a = self.linear_a(x) x_b = self.linear_b(x) x_c = self.linear_c(x) s_a = self.score_a(x) s_b = self.score_b(x) s_c = self.score_c(x) s_a = torch.sigmoid(s_a) s_b = torch.sigmoid(s_b) s_c = torch.sigmoid(s_c) f_a = x_a f_b = torch.cat([(s_a+0.5)*x_a, (s_b+0.5)*x_b], 1) f_c = torch.cat([(s_a+0.5)*x_a, (s_b+0.5)*x_b, (s_c)*x_c], 1) return [f_a, f_b, f_c, f_c]
You can see that I use some constant number here like "f_b = torch.cat([(s_a+0.5)*x_a, (s_b+0.5)*x_b], 1)"and I have forgot the exact value of these numbers. But I have stored a model and I am wondering how can I print the numbers within the equation “f_b = torch.cat([(s_a+0.5)*x_a, (s_b+0.5)*x_b], 1),1)”