samedi 11 janvier 2020

How to manipulate only weights of neural network when biases are intact using Pytorch?

In the following for loop I get name's and parameter's of my neural network. Depending on the bias flag I will do some operations on the weight's and bias's using **block**. However, when bias is False, I just do operations on weight's but if bias flag is True, I will do it for both weight's and bias's.

for name, param in model.named_parameters():
    if bias = Fasle and 'weight' in name:
        **block**
    elif bias = True and 'weight' or 'bias' in name:
        **block**

Question: since **block** is the same for if and elif, I want to avoid writing it and have it once. Can you help me to figure this out?

Aucun commentaire:

Enregistrer un commentaire