Lazy Initialization

AttributeError: module ‘d2l.torch’ has no attribute ‘add_to_class’

my d2l version is 0.17.5. I am using Google colab.


I’m very curious to learn more about problem #3… How do we use parameter tying to fix the issue of different - sized inputs?

sos, I have the same problem.

In Colab these lines will load d2l correctly

!curl -o
!unzip > /dev/null
!ln -s /content/pytorch/d2l /content/d2l
from d2l import torch as d2l